kernels.JCudaVec_kernels_float_64_cc30.ptx Maven / Gradle / Ivy
Go to download
Show more of this group Show more artifacts with this name
Show all versions of jcuda-vec Show documentation
Show all versions of jcuda-vec Show documentation
Vector operations for JCuda
The newest version!
//
// Generated by NVIDIA NVVM Compiler
//
// Compiler Build ID: CL-19805474
// Cuda compilation tools, release 7.5, V7.5.16
// Based on LLVM 3.4svn
//
.version 4.3
.target sm_30
.address_size 64
// .globl vec_setf
.const .align 4 .b8 __cudart_i2opi_f[24] = {65, 144, 67, 60, 153, 149, 98, 219, 192, 221, 52, 245, 209, 87, 39, 252, 41, 21, 68, 78, 110, 131, 249, 162};
.visible .entry vec_setf(
.param .u64 vec_setf_param_0,
.param .u64 vec_setf_param_1,
.param .f32 vec_setf_param_2
)
{
.reg .pred %p<2>;
.reg .f32 %f<2>;
.reg .b32 %r<5>;
.reg .b64 %rd<7>;
ld.param.u64 %rd2, [vec_setf_param_0];
ld.param.u64 %rd1, [vec_setf_param_1];
ld.param.f32 %f1, [vec_setf_param_2];
mov.u32 %r2, %tid.x;
mov.u32 %r3, %ntid.x;
mov.u32 %r4, %ctaid.x;
mad.lo.s32 %r1, %r3, %r4, %r2;
cvt.s64.s32 %rd3, %r1;
setp.ge.u64 %p1, %rd3, %rd2;
@%p1 bra BB0_2;
cvta.to.global.u64 %rd4, %rd1;
mul.wide.s32 %rd5, %r1, 4;
add.s64 %rd6, %rd4, %rd5;
st.global.f32 [%rd6], %f1;
BB0_2:
ret;
}
// .globl vec_addf
.visible .entry vec_addf(
.param .u64 vec_addf_param_0,
.param .u64 vec_addf_param_1,
.param .u64 vec_addf_param_2,
.param .u64 vec_addf_param_3
)
{
.reg .pred %p<2>;
.reg .f32 %f<4>;
.reg .b32 %r<5>;
.reg .b64 %rd<13>;
ld.param.u64 %rd4, [vec_addf_param_0];
ld.param.u64 %rd1, [vec_addf_param_1];
ld.param.u64 %rd2, [vec_addf_param_2];
ld.param.u64 %rd3, [vec_addf_param_3];
mov.u32 %r2, %tid.x;
mov.u32 %r3, %ntid.x;
mov.u32 %r4, %ctaid.x;
mad.lo.s32 %r1, %r3, %r4, %r2;
cvt.s64.s32 %rd5, %r1;
setp.ge.u64 %p1, %rd5, %rd4;
@%p1 bra BB1_2;
cvta.to.global.u64 %rd6, %rd2;
mul.wide.s32 %rd7, %r1, 4;
add.s64 %rd8, %rd6, %rd7;
cvta.to.global.u64 %rd9, %rd3;
add.s64 %rd10, %rd9, %rd7;
ld.global.f32 %f1, [%rd10];
ld.global.f32 %f2, [%rd8];
add.f32 %f3, %f2, %f1;
cvta.to.global.u64 %rd11, %rd1;
add.s64 %rd12, %rd11, %rd7;
st.global.f32 [%rd12], %f3;
BB1_2:
ret;
}
// .globl vec_subf
.visible .entry vec_subf(
.param .u64 vec_subf_param_0,
.param .u64 vec_subf_param_1,
.param .u64 vec_subf_param_2,
.param .u64 vec_subf_param_3
)
{
.reg .pred %p<2>;
.reg .f32 %f<4>;
.reg .b32 %r<5>;
.reg .b64 %rd<13>;
ld.param.u64 %rd4, [vec_subf_param_0];
ld.param.u64 %rd1, [vec_subf_param_1];
ld.param.u64 %rd2, [vec_subf_param_2];
ld.param.u64 %rd3, [vec_subf_param_3];
mov.u32 %r2, %tid.x;
mov.u32 %r3, %ntid.x;
mov.u32 %r4, %ctaid.x;
mad.lo.s32 %r1, %r3, %r4, %r2;
cvt.s64.s32 %rd5, %r1;
setp.ge.u64 %p1, %rd5, %rd4;
@%p1 bra BB2_2;
cvta.to.global.u64 %rd6, %rd2;
mul.wide.s32 %rd7, %r1, 4;
add.s64 %rd8, %rd6, %rd7;
cvta.to.global.u64 %rd9, %rd3;
add.s64 %rd10, %rd9, %rd7;
ld.global.f32 %f1, [%rd10];
ld.global.f32 %f2, [%rd8];
sub.f32 %f3, %f2, %f1;
cvta.to.global.u64 %rd11, %rd1;
add.s64 %rd12, %rd11, %rd7;
st.global.f32 [%rd12], %f3;
BB2_2:
ret;
}
// .globl vec_mulf
.visible .entry vec_mulf(
.param .u64 vec_mulf_param_0,
.param .u64 vec_mulf_param_1,
.param .u64 vec_mulf_param_2,
.param .u64 vec_mulf_param_3
)
{
.reg .pred %p<2>;
.reg .f32 %f<4>;
.reg .b32 %r<5>;
.reg .b64 %rd<13>;
ld.param.u64 %rd4, [vec_mulf_param_0];
ld.param.u64 %rd1, [vec_mulf_param_1];
ld.param.u64 %rd2, [vec_mulf_param_2];
ld.param.u64 %rd3, [vec_mulf_param_3];
mov.u32 %r2, %tid.x;
mov.u32 %r3, %ntid.x;
mov.u32 %r4, %ctaid.x;
mad.lo.s32 %r1, %r3, %r4, %r2;
cvt.s64.s32 %rd5, %r1;
setp.ge.u64 %p1, %rd5, %rd4;
@%p1 bra BB3_2;
cvta.to.global.u64 %rd6, %rd2;
mul.wide.s32 %rd7, %r1, 4;
add.s64 %rd8, %rd6, %rd7;
cvta.to.global.u64 %rd9, %rd3;
add.s64 %rd10, %rd9, %rd7;
ld.global.f32 %f1, [%rd10];
ld.global.f32 %f2, [%rd8];
mul.f32 %f3, %f2, %f1;
cvta.to.global.u64 %rd11, %rd1;
add.s64 %rd12, %rd11, %rd7;
st.global.f32 [%rd12], %f3;
BB3_2:
ret;
}
// .globl vec_divf
.visible .entry vec_divf(
.param .u64 vec_divf_param_0,
.param .u64 vec_divf_param_1,
.param .u64 vec_divf_param_2,
.param .u64 vec_divf_param_3
)
{
.reg .pred %p<2>;
.reg .f32 %f<4>;
.reg .b32 %r<5>;
.reg .b64 %rd<13>;
ld.param.u64 %rd5, [vec_divf_param_0];
ld.param.u64 %rd2, [vec_divf_param_1];
ld.param.u64 %rd3, [vec_divf_param_2];
ld.param.u64 %rd4, [vec_divf_param_3];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd5;
@%p1 bra BB4_2;
cvta.to.global.u64 %rd6, %rd3;
shl.b64 %rd7, %rd1, 2;
add.s64 %rd8, %rd6, %rd7;
cvta.to.global.u64 %rd9, %rd4;
add.s64 %rd10, %rd9, %rd7;
ld.global.f32 %f1, [%rd10];
ld.global.f32 %f2, [%rd8];
div.rn.f32 %f3, %f2, %f1;
cvta.to.global.u64 %rd11, %rd2;
add.s64 %rd12, %rd11, %rd7;
st.global.f32 [%rd12], %f3;
BB4_2:
ret;
}
// .globl vec_negatef
.visible .entry vec_negatef(
.param .u64 vec_negatef_param_0,
.param .u64 vec_negatef_param_1,
.param .u64 vec_negatef_param_2
)
{
.reg .pred %p<2>;
.reg .f32 %f<3>;
.reg .b32 %r<5>;
.reg .b64 %rd<10>;
ld.param.u64 %rd3, [vec_negatef_param_0];
ld.param.u64 %rd1, [vec_negatef_param_1];
ld.param.u64 %rd2, [vec_negatef_param_2];
mov.u32 %r2, %tid.x;
mov.u32 %r3, %ntid.x;
mov.u32 %r4, %ctaid.x;
mad.lo.s32 %r1, %r3, %r4, %r2;
cvt.s64.s32 %rd4, %r1;
setp.ge.u64 %p1, %rd4, %rd3;
@%p1 bra BB5_2;
cvta.to.global.u64 %rd5, %rd2;
mul.wide.s32 %rd6, %r1, 4;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f1, [%rd7];
neg.f32 %f2, %f1;
cvta.to.global.u64 %rd8, %rd1;
add.s64 %rd9, %rd8, %rd6;
st.global.f32 [%rd9], %f2;
BB5_2:
ret;
}
// .globl vec_addScalarf
.visible .entry vec_addScalarf(
.param .u64 vec_addScalarf_param_0,
.param .u64 vec_addScalarf_param_1,
.param .u64 vec_addScalarf_param_2,
.param .f32 vec_addScalarf_param_3
)
{
.reg .pred %p<2>;
.reg .f32 %f<4>;
.reg .b32 %r<5>;
.reg .b64 %rd<10>;
ld.param.u64 %rd3, [vec_addScalarf_param_0];
ld.param.u64 %rd1, [vec_addScalarf_param_1];
ld.param.u64 %rd2, [vec_addScalarf_param_2];
ld.param.f32 %f1, [vec_addScalarf_param_3];
mov.u32 %r2, %tid.x;
mov.u32 %r3, %ntid.x;
mov.u32 %r4, %ctaid.x;
mad.lo.s32 %r1, %r3, %r4, %r2;
cvt.s64.s32 %rd4, %r1;
setp.ge.u64 %p1, %rd4, %rd3;
@%p1 bra BB6_2;
cvta.to.global.u64 %rd5, %rd2;
mul.wide.s32 %rd6, %r1, 4;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f2, [%rd7];
add.f32 %f3, %f2, %f1;
cvta.to.global.u64 %rd8, %rd1;
add.s64 %rd9, %rd8, %rd6;
st.global.f32 [%rd9], %f3;
BB6_2:
ret;
}
// .globl vec_subScalarf
.visible .entry vec_subScalarf(
.param .u64 vec_subScalarf_param_0,
.param .u64 vec_subScalarf_param_1,
.param .u64 vec_subScalarf_param_2,
.param .f32 vec_subScalarf_param_3
)
{
.reg .pred %p<2>;
.reg .f32 %f<4>;
.reg .b32 %r<5>;
.reg .b64 %rd<10>;
ld.param.u64 %rd3, [vec_subScalarf_param_0];
ld.param.u64 %rd1, [vec_subScalarf_param_1];
ld.param.u64 %rd2, [vec_subScalarf_param_2];
ld.param.f32 %f1, [vec_subScalarf_param_3];
mov.u32 %r2, %tid.x;
mov.u32 %r3, %ntid.x;
mov.u32 %r4, %ctaid.x;
mad.lo.s32 %r1, %r3, %r4, %r2;
cvt.s64.s32 %rd4, %r1;
setp.ge.u64 %p1, %rd4, %rd3;
@%p1 bra BB7_2;
cvta.to.global.u64 %rd5, %rd2;
mul.wide.s32 %rd6, %r1, 4;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f2, [%rd7];
sub.f32 %f3, %f2, %f1;
cvta.to.global.u64 %rd8, %rd1;
add.s64 %rd9, %rd8, %rd6;
st.global.f32 [%rd9], %f3;
BB7_2:
ret;
}
// .globl vec_mulScalarf
.visible .entry vec_mulScalarf(
.param .u64 vec_mulScalarf_param_0,
.param .u64 vec_mulScalarf_param_1,
.param .u64 vec_mulScalarf_param_2,
.param .f32 vec_mulScalarf_param_3
)
{
.reg .pred %p<2>;
.reg .f32 %f<4>;
.reg .b32 %r<5>;
.reg .b64 %rd<10>;
ld.param.u64 %rd3, [vec_mulScalarf_param_0];
ld.param.u64 %rd1, [vec_mulScalarf_param_1];
ld.param.u64 %rd2, [vec_mulScalarf_param_2];
ld.param.f32 %f1, [vec_mulScalarf_param_3];
mov.u32 %r2, %tid.x;
mov.u32 %r3, %ntid.x;
mov.u32 %r4, %ctaid.x;
mad.lo.s32 %r1, %r3, %r4, %r2;
cvt.s64.s32 %rd4, %r1;
setp.ge.u64 %p1, %rd4, %rd3;
@%p1 bra BB8_2;
cvta.to.global.u64 %rd5, %rd2;
mul.wide.s32 %rd6, %r1, 4;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f2, [%rd7];
mul.f32 %f3, %f2, %f1;
cvta.to.global.u64 %rd8, %rd1;
add.s64 %rd9, %rd8, %rd6;
st.global.f32 [%rd9], %f3;
BB8_2:
ret;
}
// .globl vec_divScalarf
.visible .entry vec_divScalarf(
.param .u64 vec_divScalarf_param_0,
.param .u64 vec_divScalarf_param_1,
.param .u64 vec_divScalarf_param_2,
.param .f32 vec_divScalarf_param_3
)
{
.reg .pred %p<2>;
.reg .f32 %f<4>;
.reg .b32 %r<5>;
.reg .b64 %rd<10>;
ld.param.u64 %rd4, [vec_divScalarf_param_0];
ld.param.u64 %rd2, [vec_divScalarf_param_1];
ld.param.u64 %rd3, [vec_divScalarf_param_2];
ld.param.f32 %f1, [vec_divScalarf_param_3];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB9_2;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f2, [%rd7];
div.rn.f32 %f3, %f2, %f1;
cvta.to.global.u64 %rd8, %rd2;
add.s64 %rd9, %rd8, %rd6;
st.global.f32 [%rd9], %f3;
BB9_2:
ret;
}
// .globl vec_scalarAddf
.visible .entry vec_scalarAddf(
.param .u64 vec_scalarAddf_param_0,
.param .u64 vec_scalarAddf_param_1,
.param .f32 vec_scalarAddf_param_2,
.param .u64 vec_scalarAddf_param_3
)
{
.reg .pred %p<2>;
.reg .f32 %f<4>;
.reg .b32 %r<5>;
.reg .b64 %rd<10>;
ld.param.u64 %rd3, [vec_scalarAddf_param_0];
ld.param.u64 %rd1, [vec_scalarAddf_param_1];
ld.param.f32 %f1, [vec_scalarAddf_param_2];
ld.param.u64 %rd2, [vec_scalarAddf_param_3];
mov.u32 %r2, %tid.x;
mov.u32 %r3, %ntid.x;
mov.u32 %r4, %ctaid.x;
mad.lo.s32 %r1, %r3, %r4, %r2;
cvt.s64.s32 %rd4, %r1;
setp.ge.u64 %p1, %rd4, %rd3;
@%p1 bra BB10_2;
cvta.to.global.u64 %rd5, %rd2;
mul.wide.s32 %rd6, %r1, 4;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f2, [%rd7];
add.f32 %f3, %f2, %f1;
cvta.to.global.u64 %rd8, %rd1;
add.s64 %rd9, %rd8, %rd6;
st.global.f32 [%rd9], %f3;
BB10_2:
ret;
}
// .globl vec_scalarSubf
.visible .entry vec_scalarSubf(
.param .u64 vec_scalarSubf_param_0,
.param .u64 vec_scalarSubf_param_1,
.param .f32 vec_scalarSubf_param_2,
.param .u64 vec_scalarSubf_param_3
)
{
.reg .pred %p<2>;
.reg .f32 %f<4>;
.reg .b32 %r<5>;
.reg .b64 %rd<10>;
ld.param.u64 %rd3, [vec_scalarSubf_param_0];
ld.param.u64 %rd1, [vec_scalarSubf_param_1];
ld.param.f32 %f1, [vec_scalarSubf_param_2];
ld.param.u64 %rd2, [vec_scalarSubf_param_3];
mov.u32 %r2, %tid.x;
mov.u32 %r3, %ntid.x;
mov.u32 %r4, %ctaid.x;
mad.lo.s32 %r1, %r3, %r4, %r2;
cvt.s64.s32 %rd4, %r1;
setp.ge.u64 %p1, %rd4, %rd3;
@%p1 bra BB11_2;
cvta.to.global.u64 %rd5, %rd2;
mul.wide.s32 %rd6, %r1, 4;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f2, [%rd7];
sub.f32 %f3, %f1, %f2;
cvta.to.global.u64 %rd8, %rd1;
add.s64 %rd9, %rd8, %rd6;
st.global.f32 [%rd9], %f3;
BB11_2:
ret;
}
// .globl vec_scalarMulf
.visible .entry vec_scalarMulf(
.param .u64 vec_scalarMulf_param_0,
.param .u64 vec_scalarMulf_param_1,
.param .f32 vec_scalarMulf_param_2,
.param .u64 vec_scalarMulf_param_3
)
{
.reg .pred %p<2>;
.reg .f32 %f<4>;
.reg .b32 %r<5>;
.reg .b64 %rd<10>;
ld.param.u64 %rd3, [vec_scalarMulf_param_0];
ld.param.u64 %rd1, [vec_scalarMulf_param_1];
ld.param.f32 %f1, [vec_scalarMulf_param_2];
ld.param.u64 %rd2, [vec_scalarMulf_param_3];
mov.u32 %r2, %tid.x;
mov.u32 %r3, %ntid.x;
mov.u32 %r4, %ctaid.x;
mad.lo.s32 %r1, %r3, %r4, %r2;
cvt.s64.s32 %rd4, %r1;
setp.ge.u64 %p1, %rd4, %rd3;
@%p1 bra BB12_2;
cvta.to.global.u64 %rd5, %rd2;
mul.wide.s32 %rd6, %r1, 4;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f2, [%rd7];
mul.f32 %f3, %f2, %f1;
cvta.to.global.u64 %rd8, %rd1;
add.s64 %rd9, %rd8, %rd6;
st.global.f32 [%rd9], %f3;
BB12_2:
ret;
}
// .globl vec_scalarDivf
.visible .entry vec_scalarDivf(
.param .u64 vec_scalarDivf_param_0,
.param .u64 vec_scalarDivf_param_1,
.param .f32 vec_scalarDivf_param_2,
.param .u64 vec_scalarDivf_param_3
)
{
.reg .pred %p<2>;
.reg .f32 %f<4>;
.reg .b32 %r<5>;
.reg .b64 %rd<10>;
ld.param.u64 %rd4, [vec_scalarDivf_param_0];
ld.param.u64 %rd2, [vec_scalarDivf_param_1];
ld.param.f32 %f1, [vec_scalarDivf_param_2];
ld.param.u64 %rd3, [vec_scalarDivf_param_3];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB13_2;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f2, [%rd7];
div.rn.f32 %f3, %f1, %f2;
cvta.to.global.u64 %rd8, %rd2;
add.s64 %rd9, %rd8, %rd6;
st.global.f32 [%rd9], %f3;
BB13_2:
ret;
}
// .globl vec_ltf
.visible .entry vec_ltf(
.param .u64 vec_ltf_param_0,
.param .u64 vec_ltf_param_1,
.param .u64 vec_ltf_param_2,
.param .u64 vec_ltf_param_3
)
{
.reg .pred %p<3>;
.reg .f32 %f<4>;
.reg .b32 %r<5>;
.reg .b64 %rd<13>;
ld.param.u64 %rd4, [vec_ltf_param_0];
ld.param.u64 %rd1, [vec_ltf_param_1];
ld.param.u64 %rd2, [vec_ltf_param_2];
ld.param.u64 %rd3, [vec_ltf_param_3];
mov.u32 %r2, %tid.x;
mov.u32 %r3, %ntid.x;
mov.u32 %r4, %ctaid.x;
mad.lo.s32 %r1, %r3, %r4, %r2;
cvt.s64.s32 %rd5, %r1;
setp.ge.u64 %p1, %rd5, %rd4;
@%p1 bra BB14_2;
cvta.to.global.u64 %rd6, %rd2;
mul.wide.s32 %rd7, %r1, 4;
add.s64 %rd8, %rd6, %rd7;
cvta.to.global.u64 %rd9, %rd3;
add.s64 %rd10, %rd9, %rd7;
ld.global.f32 %f1, [%rd10];
ld.global.f32 %f2, [%rd8];
setp.lt.f32 %p2, %f2, %f1;
selp.f32 %f3, 0f3F800000, 0f00000000, %p2;
cvta.to.global.u64 %rd11, %rd1;
add.s64 %rd12, %rd11, %rd7;
st.global.f32 [%rd12], %f3;
BB14_2:
ret;
}
// .globl vec_ltef
.visible .entry vec_ltef(
.param .u64 vec_ltef_param_0,
.param .u64 vec_ltef_param_1,
.param .u64 vec_ltef_param_2,
.param .u64 vec_ltef_param_3
)
{
.reg .pred %p<3>;
.reg .f32 %f<4>;
.reg .b32 %r<5>;
.reg .b64 %rd<13>;
ld.param.u64 %rd4, [vec_ltef_param_0];
ld.param.u64 %rd1, [vec_ltef_param_1];
ld.param.u64 %rd2, [vec_ltef_param_2];
ld.param.u64 %rd3, [vec_ltef_param_3];
mov.u32 %r2, %tid.x;
mov.u32 %r3, %ntid.x;
mov.u32 %r4, %ctaid.x;
mad.lo.s32 %r1, %r3, %r4, %r2;
cvt.s64.s32 %rd5, %r1;
setp.ge.u64 %p1, %rd5, %rd4;
@%p1 bra BB15_2;
cvta.to.global.u64 %rd6, %rd2;
mul.wide.s32 %rd7, %r1, 4;
add.s64 %rd8, %rd6, %rd7;
cvta.to.global.u64 %rd9, %rd3;
add.s64 %rd10, %rd9, %rd7;
ld.global.f32 %f1, [%rd10];
ld.global.f32 %f2, [%rd8];
setp.gtu.f32 %p2, %f2, %f1;
selp.f32 %f3, 0f00000000, 0f3F800000, %p2;
cvta.to.global.u64 %rd11, %rd1;
add.s64 %rd12, %rd11, %rd7;
st.global.f32 [%rd12], %f3;
BB15_2:
ret;
}
// .globl vec_eqf
.visible .entry vec_eqf(
.param .u64 vec_eqf_param_0,
.param .u64 vec_eqf_param_1,
.param .u64 vec_eqf_param_2,
.param .u64 vec_eqf_param_3
)
{
.reg .pred %p<3>;
.reg .f32 %f<4>;
.reg .b32 %r<5>;
.reg .b64 %rd<13>;
ld.param.u64 %rd4, [vec_eqf_param_0];
ld.param.u64 %rd1, [vec_eqf_param_1];
ld.param.u64 %rd2, [vec_eqf_param_2];
ld.param.u64 %rd3, [vec_eqf_param_3];
mov.u32 %r2, %tid.x;
mov.u32 %r3, %ntid.x;
mov.u32 %r4, %ctaid.x;
mad.lo.s32 %r1, %r3, %r4, %r2;
cvt.s64.s32 %rd5, %r1;
setp.ge.u64 %p1, %rd5, %rd4;
@%p1 bra BB16_2;
cvta.to.global.u64 %rd6, %rd2;
mul.wide.s32 %rd7, %r1, 4;
add.s64 %rd8, %rd6, %rd7;
cvta.to.global.u64 %rd9, %rd3;
add.s64 %rd10, %rd9, %rd7;
ld.global.f32 %f1, [%rd10];
ld.global.f32 %f2, [%rd8];
setp.eq.f32 %p2, %f2, %f1;
selp.f32 %f3, 0f3F800000, 0f00000000, %p2;
cvta.to.global.u64 %rd11, %rd1;
add.s64 %rd12, %rd11, %rd7;
st.global.f32 [%rd12], %f3;
BB16_2:
ret;
}
// .globl vec_gtef
.visible .entry vec_gtef(
.param .u64 vec_gtef_param_0,
.param .u64 vec_gtef_param_1,
.param .u64 vec_gtef_param_2,
.param .u64 vec_gtef_param_3
)
{
.reg .pred %p<3>;
.reg .f32 %f<4>;
.reg .b32 %r<5>;
.reg .b64 %rd<13>;
ld.param.u64 %rd4, [vec_gtef_param_0];
ld.param.u64 %rd1, [vec_gtef_param_1];
ld.param.u64 %rd2, [vec_gtef_param_2];
ld.param.u64 %rd3, [vec_gtef_param_3];
mov.u32 %r2, %tid.x;
mov.u32 %r3, %ntid.x;
mov.u32 %r4, %ctaid.x;
mad.lo.s32 %r1, %r3, %r4, %r2;
cvt.s64.s32 %rd5, %r1;
setp.ge.u64 %p1, %rd5, %rd4;
@%p1 bra BB17_2;
cvta.to.global.u64 %rd6, %rd2;
mul.wide.s32 %rd7, %r1, 4;
add.s64 %rd8, %rd6, %rd7;
cvta.to.global.u64 %rd9, %rd3;
add.s64 %rd10, %rd9, %rd7;
ld.global.f32 %f1, [%rd10];
ld.global.f32 %f2, [%rd8];
setp.ltu.f32 %p2, %f2, %f1;
selp.f32 %f3, 0f00000000, 0f3F800000, %p2;
cvta.to.global.u64 %rd11, %rd1;
add.s64 %rd12, %rd11, %rd7;
st.global.f32 [%rd12], %f3;
BB17_2:
ret;
}
// .globl vec_gtf
.visible .entry vec_gtf(
.param .u64 vec_gtf_param_0,
.param .u64 vec_gtf_param_1,
.param .u64 vec_gtf_param_2,
.param .u64 vec_gtf_param_3
)
{
.reg .pred %p<3>;
.reg .f32 %f<4>;
.reg .b32 %r<5>;
.reg .b64 %rd<13>;
ld.param.u64 %rd4, [vec_gtf_param_0];
ld.param.u64 %rd1, [vec_gtf_param_1];
ld.param.u64 %rd2, [vec_gtf_param_2];
ld.param.u64 %rd3, [vec_gtf_param_3];
mov.u32 %r2, %tid.x;
mov.u32 %r3, %ntid.x;
mov.u32 %r4, %ctaid.x;
mad.lo.s32 %r1, %r3, %r4, %r2;
cvt.s64.s32 %rd5, %r1;
setp.ge.u64 %p1, %rd5, %rd4;
@%p1 bra BB18_2;
cvta.to.global.u64 %rd6, %rd2;
mul.wide.s32 %rd7, %r1, 4;
add.s64 %rd8, %rd6, %rd7;
cvta.to.global.u64 %rd9, %rd3;
add.s64 %rd10, %rd9, %rd7;
ld.global.f32 %f1, [%rd10];
ld.global.f32 %f2, [%rd8];
setp.gt.f32 %p2, %f2, %f1;
selp.f32 %f3, 0f3F800000, 0f00000000, %p2;
cvta.to.global.u64 %rd11, %rd1;
add.s64 %rd12, %rd11, %rd7;
st.global.f32 [%rd12], %f3;
BB18_2:
ret;
}
// .globl vec_nef
.visible .entry vec_nef(
.param .u64 vec_nef_param_0,
.param .u64 vec_nef_param_1,
.param .u64 vec_nef_param_2,
.param .u64 vec_nef_param_3
)
{
.reg .pred %p<3>;
.reg .f32 %f<4>;
.reg .b32 %r<5>;
.reg .b64 %rd<13>;
ld.param.u64 %rd4, [vec_nef_param_0];
ld.param.u64 %rd1, [vec_nef_param_1];
ld.param.u64 %rd2, [vec_nef_param_2];
ld.param.u64 %rd3, [vec_nef_param_3];
mov.u32 %r2, %tid.x;
mov.u32 %r3, %ntid.x;
mov.u32 %r4, %ctaid.x;
mad.lo.s32 %r1, %r3, %r4, %r2;
cvt.s64.s32 %rd5, %r1;
setp.ge.u64 %p1, %rd5, %rd4;
@%p1 bra BB19_2;
cvta.to.global.u64 %rd6, %rd2;
mul.wide.s32 %rd7, %r1, 4;
add.s64 %rd8, %rd6, %rd7;
cvta.to.global.u64 %rd9, %rd3;
add.s64 %rd10, %rd9, %rd7;
ld.global.f32 %f1, [%rd10];
ld.global.f32 %f2, [%rd8];
setp.neu.f32 %p2, %f2, %f1;
selp.f32 %f3, 0f3F800000, 0f00000000, %p2;
cvta.to.global.u64 %rd11, %rd1;
add.s64 %rd12, %rd11, %rd7;
st.global.f32 [%rd12], %f3;
BB19_2:
ret;
}
// .globl vec_ltScalarf
.visible .entry vec_ltScalarf(
.param .u64 vec_ltScalarf_param_0,
.param .u64 vec_ltScalarf_param_1,
.param .u64 vec_ltScalarf_param_2,
.param .f32 vec_ltScalarf_param_3
)
{
.reg .pred %p<3>;
.reg .f32 %f<4>;
.reg .b32 %r<5>;
.reg .b64 %rd<10>;
ld.param.u64 %rd3, [vec_ltScalarf_param_0];
ld.param.u64 %rd1, [vec_ltScalarf_param_1];
ld.param.u64 %rd2, [vec_ltScalarf_param_2];
ld.param.f32 %f1, [vec_ltScalarf_param_3];
mov.u32 %r2, %tid.x;
mov.u32 %r3, %ntid.x;
mov.u32 %r4, %ctaid.x;
mad.lo.s32 %r1, %r3, %r4, %r2;
cvt.s64.s32 %rd4, %r1;
setp.ge.u64 %p1, %rd4, %rd3;
@%p1 bra BB20_2;
cvta.to.global.u64 %rd5, %rd2;
mul.wide.s32 %rd6, %r1, 4;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f2, [%rd7];
setp.lt.f32 %p2, %f2, %f1;
selp.f32 %f3, 0f3F800000, 0f00000000, %p2;
cvta.to.global.u64 %rd8, %rd1;
add.s64 %rd9, %rd8, %rd6;
st.global.f32 [%rd9], %f3;
BB20_2:
ret;
}
// .globl vec_lteScalarf
.visible .entry vec_lteScalarf(
.param .u64 vec_lteScalarf_param_0,
.param .u64 vec_lteScalarf_param_1,
.param .u64 vec_lteScalarf_param_2,
.param .f32 vec_lteScalarf_param_3
)
{
.reg .pred %p<3>;
.reg .f32 %f<4>;
.reg .b32 %r<5>;
.reg .b64 %rd<10>;
ld.param.u64 %rd3, [vec_lteScalarf_param_0];
ld.param.u64 %rd1, [vec_lteScalarf_param_1];
ld.param.u64 %rd2, [vec_lteScalarf_param_2];
ld.param.f32 %f1, [vec_lteScalarf_param_3];
mov.u32 %r2, %tid.x;
mov.u32 %r3, %ntid.x;
mov.u32 %r4, %ctaid.x;
mad.lo.s32 %r1, %r3, %r4, %r2;
cvt.s64.s32 %rd4, %r1;
setp.ge.u64 %p1, %rd4, %rd3;
@%p1 bra BB21_2;
cvta.to.global.u64 %rd5, %rd2;
mul.wide.s32 %rd6, %r1, 4;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f2, [%rd7];
setp.gtu.f32 %p2, %f2, %f1;
selp.f32 %f3, 0f00000000, 0f3F800000, %p2;
cvta.to.global.u64 %rd8, %rd1;
add.s64 %rd9, %rd8, %rd6;
st.global.f32 [%rd9], %f3;
BB21_2:
ret;
}
// .globl vec_eqScalarf
.visible .entry vec_eqScalarf(
.param .u64 vec_eqScalarf_param_0,
.param .u64 vec_eqScalarf_param_1,
.param .u64 vec_eqScalarf_param_2,
.param .f32 vec_eqScalarf_param_3
)
{
.reg .pred %p<3>;
.reg .f32 %f<4>;
.reg .b32 %r<5>;
.reg .b64 %rd<10>;
ld.param.u64 %rd3, [vec_eqScalarf_param_0];
ld.param.u64 %rd1, [vec_eqScalarf_param_1];
ld.param.u64 %rd2, [vec_eqScalarf_param_2];
ld.param.f32 %f1, [vec_eqScalarf_param_3];
mov.u32 %r2, %tid.x;
mov.u32 %r3, %ntid.x;
mov.u32 %r4, %ctaid.x;
mad.lo.s32 %r1, %r3, %r4, %r2;
cvt.s64.s32 %rd4, %r1;
setp.ge.u64 %p1, %rd4, %rd3;
@%p1 bra BB22_2;
cvta.to.global.u64 %rd5, %rd2;
mul.wide.s32 %rd6, %r1, 4;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f2, [%rd7];
setp.eq.f32 %p2, %f2, %f1;
selp.f32 %f3, 0f3F800000, 0f00000000, %p2;
cvta.to.global.u64 %rd8, %rd1;
add.s64 %rd9, %rd8, %rd6;
st.global.f32 [%rd9], %f3;
BB22_2:
ret;
}
// .globl vec_gteScalarf
.visible .entry vec_gteScalarf(
.param .u64 vec_gteScalarf_param_0,
.param .u64 vec_gteScalarf_param_1,
.param .u64 vec_gteScalarf_param_2,
.param .f32 vec_gteScalarf_param_3
)
{
.reg .pred %p<3>;
.reg .f32 %f<4>;
.reg .b32 %r<5>;
.reg .b64 %rd<10>;
ld.param.u64 %rd3, [vec_gteScalarf_param_0];
ld.param.u64 %rd1, [vec_gteScalarf_param_1];
ld.param.u64 %rd2, [vec_gteScalarf_param_2];
ld.param.f32 %f1, [vec_gteScalarf_param_3];
mov.u32 %r2, %tid.x;
mov.u32 %r3, %ntid.x;
mov.u32 %r4, %ctaid.x;
mad.lo.s32 %r1, %r3, %r4, %r2;
cvt.s64.s32 %rd4, %r1;
setp.ge.u64 %p1, %rd4, %rd3;
@%p1 bra BB23_2;
cvta.to.global.u64 %rd5, %rd2;
mul.wide.s32 %rd6, %r1, 4;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f2, [%rd7];
setp.ltu.f32 %p2, %f2, %f1;
selp.f32 %f3, 0f00000000, 0f3F800000, %p2;
cvta.to.global.u64 %rd8, %rd1;
add.s64 %rd9, %rd8, %rd6;
st.global.f32 [%rd9], %f3;
BB23_2:
ret;
}
// .globl vec_gtScalarf
.visible .entry vec_gtScalarf(
.param .u64 vec_gtScalarf_param_0,
.param .u64 vec_gtScalarf_param_1,
.param .u64 vec_gtScalarf_param_2,
.param .f32 vec_gtScalarf_param_3
)
{
.reg .pred %p<3>;
.reg .f32 %f<4>;
.reg .b32 %r<5>;
.reg .b64 %rd<10>;
ld.param.u64 %rd3, [vec_gtScalarf_param_0];
ld.param.u64 %rd1, [vec_gtScalarf_param_1];
ld.param.u64 %rd2, [vec_gtScalarf_param_2];
ld.param.f32 %f1, [vec_gtScalarf_param_3];
mov.u32 %r2, %tid.x;
mov.u32 %r3, %ntid.x;
mov.u32 %r4, %ctaid.x;
mad.lo.s32 %r1, %r3, %r4, %r2;
cvt.s64.s32 %rd4, %r1;
setp.ge.u64 %p1, %rd4, %rd3;
@%p1 bra BB24_2;
cvta.to.global.u64 %rd5, %rd2;
mul.wide.s32 %rd6, %r1, 4;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f2, [%rd7];
setp.gt.f32 %p2, %f2, %f1;
selp.f32 %f3, 0f3F800000, 0f00000000, %p2;
cvta.to.global.u64 %rd8, %rd1;
add.s64 %rd9, %rd8, %rd6;
st.global.f32 [%rd9], %f3;
BB24_2:
ret;
}
// .globl vec_neScalarf
.visible .entry vec_neScalarf(
.param .u64 vec_neScalarf_param_0,
.param .u64 vec_neScalarf_param_1,
.param .u64 vec_neScalarf_param_2,
.param .f32 vec_neScalarf_param_3
)
{
.reg .pred %p<3>;
.reg .f32 %f<4>;
.reg .b32 %r<5>;
.reg .b64 %rd<10>;
ld.param.u64 %rd3, [vec_neScalarf_param_0];
ld.param.u64 %rd1, [vec_neScalarf_param_1];
ld.param.u64 %rd2, [vec_neScalarf_param_2];
ld.param.f32 %f1, [vec_neScalarf_param_3];
mov.u32 %r2, %tid.x;
mov.u32 %r3, %ntid.x;
mov.u32 %r4, %ctaid.x;
mad.lo.s32 %r1, %r3, %r4, %r2;
cvt.s64.s32 %rd4, %r1;
setp.ge.u64 %p1, %rd4, %rd3;
@%p1 bra BB25_2;
cvta.to.global.u64 %rd5, %rd2;
mul.wide.s32 %rd6, %r1, 4;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f2, [%rd7];
setp.neu.f32 %p2, %f2, %f1;
selp.f32 %f3, 0f3F800000, 0f00000000, %p2;
cvta.to.global.u64 %rd8, %rd1;
add.s64 %rd9, %rd8, %rd6;
st.global.f32 [%rd9], %f3;
BB25_2:
ret;
}
// .globl vec_acosf
.visible .entry vec_acosf(
.param .u64 vec_acosf_param_0,
.param .u64 vec_acosf_param_1,
.param .u64 vec_acosf_param_2
)
{
.reg .pred %p<4>;
.reg .f32 %f<27>;
.reg .b32 %r<5>;
.reg .b64 %rd<10>;
ld.param.u64 %rd4, [vec_acosf_param_0];
ld.param.u64 %rd2, [vec_acosf_param_1];
ld.param.u64 %rd3, [vec_acosf_param_2];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB26_2;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f1, [%rd7];
abs.f32 %f2, %f1;
mov.f32 %f3, 0f3F800000;
sub.f32 %f4, %f3, %f2;
mul.f32 %f5, %f4, 0f3F000000;
sqrt.rn.f32 %f6, %f5;
setp.gt.f32 %p2, %f2, 0f3F11EB85;
selp.f32 %f7, %f6, %f2, %p2;
mul.f32 %f8, %f7, %f7;
mov.f32 %f9, 0f3C94D2E9;
mov.f32 %f10, 0f3D53F941;
fma.rn.f32 %f11, %f10, %f8, %f9;
mov.f32 %f12, 0f3D3F841F;
fma.rn.f32 %f13, %f11, %f8, %f12;
mov.f32 %f14, 0f3D994929;
fma.rn.f32 %f15, %f13, %f8, %f14;
mov.f32 %f16, 0f3E2AAB94;
fma.rn.f32 %f17, %f15, %f8, %f16;
mul.f32 %f18, %f8, %f17;
fma.rn.f32 %f19, %f18, %f7, %f7;
add.f32 %f20, %f19, %f19;
mov.f32 %f21, 0f3FC90FDB;
sub.f32 %f22, %f21, %f19;
selp.f32 %f23, %f20, %f22, %p2;
setp.lt.f32 %p3, %f1, 0f00000000;
mov.f32 %f24, 0f40490FDB;
sub.f32 %f25, %f24, %f23;
selp.f32 %f26, %f25, %f23, %p3;
cvta.to.global.u64 %rd8, %rd2;
add.s64 %rd9, %rd8, %rd6;
st.global.f32 [%rd9], %f26;
BB26_2:
ret;
}
// .globl vec_acoshf
.visible .entry vec_acoshf(
.param .u64 vec_acoshf_param_0,
.param .u64 vec_acoshf_param_1,
.param .u64 vec_acoshf_param_2
)
{
.reg .pred %p<16>;
.reg .f32 %f<85>;
.reg .b32 %r<14>;
.reg .b64 %rd<11>;
ld.param.u64 %rd4, [vec_acoshf_param_0];
ld.param.u64 %rd2, [vec_acoshf_param_1];
ld.param.u64 %rd3, [vec_acoshf_param_2];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB27_12;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f1, [%rd7];
add.f32 %f2, %f1, 0fBF800000;
mov.b32 %r5, %f2;
setp.gt.u32 %p2, %r5, 1258291200;
@%p2 bra BB27_7;
bra.uni BB27_2;
BB27_7:
setp.gt.f32 %p11, %f2, 0f00000000;
setp.lt.f32 %p12, %f2, 0f7F800000;
and.pred %p13, %p11, %p12;
@%p13 bra BB27_9;
bra.uni BB27_8;
BB27_9:
setp.lt.f32 %p14, %f2, 0f00800000;
mul.f32 %f58, %f2, 0f4B800000;
selp.f32 %f59, %f58, %f2, %p14;
selp.f32 %f60, 0fC3170000, 0fC2FE0000, %p14;
mov.b32 %r10, %f59;
and.b32 %r11, %r10, 8388607;
or.b32 %r12, %r11, 1065353216;
mov.b32 %f61, %r12;
shr.u32 %r13, %r10, 23;
cvt.rn.f32.u32 %f62, %r13;
add.f32 %f63, %f60, %f62;
setp.gt.f32 %p15, %f61, 0f3FAE147B;
mul.f32 %f64, %f61, 0f3F000000;
add.f32 %f65, %f63, 0f3F800000;
selp.f32 %f66, %f64, %f61, %p15;
selp.f32 %f67, %f65, %f63, %p15;
add.f32 %f57, %f66, 0f3F800000;
add.f32 %f68, %f66, 0fBF800000;
// inline asm
rcp.approx.ftz.f32 %f56,%f57;
// inline asm
mul.f32 %f69, %f68, %f68;
neg.f32 %f70, %f69;
mul.rn.f32 %f71, %f56, %f70;
add.rn.f32 %f72, %f68, %f71;
mul.f32 %f73, %f72, %f72;
mov.f32 %f74, 0f3C4C6A36;
mov.f32 %f75, 0f3B1E94E6;
fma.rn.f32 %f76, %f75, %f73, %f74;
mov.f32 %f77, 0f3DAAAB1A;
fma.rn.f32 %f78, %f76, %f73, %f77;
mul.f32 %f79, %f73, %f78;
fma.rn.f32 %f80, %f79, %f72, %f71;
add.f32 %f81, %f68, %f80;
mov.f32 %f82, 0f3F317218;
fma.rn.f32 %f83, %f67, %f82, %f81;
bra.uni BB27_10;
BB27_2:
mul.rz.f32 %f13, %f1, %f2;
add.rn.f32 %f14, %f13, %f2;
sqrt.rn.f32 %f15, %f14;
add.f32 %f3, %f2, %f15;
setp.le.f32 %p3, %f3, 0f3F266666;
setp.ge.f32 %p4, %f3, 0fBEC9BA5E;
and.pred %p5, %p4, %p3;
@%p5 bra BB27_6;
bra.uni BB27_3;
BB27_6:
add.f32 %f43, %f3, 0f40000000;
div.approx.f32 %f44, %f3, %f43;
neg.f32 %f45, %f3;
mul.rn.f32 %f46, %f45, %f44;
add.rn.f32 %f47, %f3, %f46;
mul.f32 %f48, %f47, %f47;
mov.f32 %f49, 0f3C4C4BE0;
mov.f32 %f50, 0f3B2063C3;
fma.rn.f32 %f51, %f50, %f48, %f49;
mov.f32 %f52, 0f3DAAAB50;
fma.rn.f32 %f53, %f51, %f48, %f52;
mul.f32 %f54, %f48, %f53;
fma.rn.f32 %f55, %f54, %f47, %f46;
add.f32 %f84, %f3, %f55;
bra.uni BB27_11;
BB27_8:
lg2.approx.f32 %f83, %f2;
BB27_10:
add.f32 %f84, %f83, 0f3F317218;
bra.uni BB27_11;
BB27_3:
add.f32 %f4, %f3, 0f3F800000;
setp.gt.f32 %p6, %f4, 0f00000000;
setp.lt.f32 %p7, %f4, 0f7F800000;
and.pred %p8, %p6, %p7;
@%p8 bra BB27_5;
bra.uni BB27_4;
BB27_5:
setp.lt.f32 %p9, %f4, 0f00800000;
mul.f32 %f18, %f4, 0f4B800000;
selp.f32 %f19, %f18, %f4, %p9;
selp.f32 %f20, 0fC3170000, 0fC2FE0000, %p9;
mov.b32 %r6, %f19;
and.b32 %r7, %r6, 8388607;
or.b32 %r8, %r7, 1065353216;
mov.b32 %f21, %r8;
shr.u32 %r9, %r6, 23;
cvt.rn.f32.u32 %f22, %r9;
add.f32 %f23, %f20, %f22;
setp.gt.f32 %p10, %f21, 0f3FAE147B;
mul.f32 %f24, %f21, 0f3F000000;
add.f32 %f25, %f23, 0f3F800000;
selp.f32 %f26, %f24, %f21, %p10;
selp.f32 %f27, %f25, %f23, %p10;
add.f32 %f17, %f26, 0f3F800000;
add.f32 %f28, %f26, 0fBF800000;
// inline asm
rcp.approx.ftz.f32 %f16,%f17;
// inline asm
mul.f32 %f29, %f28, %f28;
neg.f32 %f30, %f29;
mul.rn.f32 %f31, %f16, %f30;
add.rn.f32 %f32, %f28, %f31;
mul.f32 %f33, %f32, %f32;
mov.f32 %f34, 0f3C4C6A36;
mov.f32 %f35, 0f3B1E94E6;
fma.rn.f32 %f36, %f35, %f33, %f34;
mov.f32 %f37, 0f3DAAAB1A;
fma.rn.f32 %f38, %f36, %f33, %f37;
mul.f32 %f39, %f33, %f38;
fma.rn.f32 %f40, %f39, %f32, %f31;
add.f32 %f41, %f28, %f40;
mov.f32 %f42, 0f3F317218;
fma.rn.f32 %f84, %f27, %f42, %f41;
bra.uni BB27_11;
BB27_4:
lg2.approx.f32 %f84, %f4;
BB27_11:
cvta.to.global.u64 %rd8, %rd2;
add.s64 %rd10, %rd8, %rd6;
st.global.f32 [%rd10], %f84;
BB27_12:
ret;
}
// .globl vec_asinf
.visible .entry vec_asinf(
.param .u64 vec_asinf_param_0,
.param .u64 vec_asinf_param_1,
.param .u64 vec_asinf_param_2
)
{
.reg .pred %p<4>;
.reg .f32 %f<26>;
.reg .b32 %r<9>;
.reg .b64 %rd<10>;
ld.param.u64 %rd4, [vec_asinf_param_0];
ld.param.u64 %rd2, [vec_asinf_param_1];
ld.param.u64 %rd3, [vec_asinf_param_2];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB28_2;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f1, [%rd7];
abs.f32 %f2, %f1;
mov.f32 %f3, 0f3F800000;
sub.f32 %f4, %f3, %f2;
mul.f32 %f5, %f4, 0f3F000000;
sqrt.rn.f32 %f6, %f5;
setp.gt.f32 %p2, %f2, 0f3F11EB85;
selp.f32 %f7, %f6, %f2, %p2;
mul.f32 %f8, %f7, %f7;
mov.f32 %f9, 0f3C94D2E9;
mov.f32 %f10, 0f3D53F941;
fma.rn.f32 %f11, %f10, %f8, %f9;
mov.f32 %f12, 0f3D3F841F;
fma.rn.f32 %f13, %f11, %f8, %f12;
mov.f32 %f14, 0f3D994929;
fma.rn.f32 %f15, %f13, %f8, %f14;
mov.f32 %f16, 0f3E2AAB94;
fma.rn.f32 %f17, %f15, %f8, %f16;
mul.f32 %f18, %f8, %f17;
fma.rn.f32 %f19, %f18, %f7, %f7;
mov.f32 %f20, 0f3FC90FDB;
mov.f32 %f21, 0fC0000000;
fma.rn.f32 %f22, %f21, %f19, %f20;
selp.f32 %f23, %f22, %f19, %p2;
setp.gtu.f32 %p3, %f23, 0f7F800000;
mov.b32 %r5, %f23;
mov.b32 %r6, %f1;
and.b32 %r7, %r6, -2147483648;
or.b32 %r8, %r5, %r7;
mov.b32 %f24, %r8;
selp.f32 %f25, %f23, %f24, %p3;
cvta.to.global.u64 %rd8, %rd2;
add.s64 %rd9, %rd8, %rd6;
st.global.f32 [%rd9], %f25;
BB28_2:
ret;
}
// .globl vec_asinhf
.visible .entry vec_asinhf(
.param .u64 vec_asinhf_param_0,
.param .u64 vec_asinhf_param_1,
.param .u64 vec_asinhf_param_2
)
{
.reg .pred %p<12>;
.reg .f32 %f<62>;
.reg .b32 %r<13>;
.reg .b64 %rd<11>;
ld.param.u64 %rd4, [vec_asinhf_param_0];
ld.param.u64 %rd2, [vec_asinhf_param_1];
ld.param.u64 %rd3, [vec_asinhf_param_2];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB29_9;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f1, [%rd7];
abs.f32 %f2, %f1;
setp.gt.f32 %p2, %f2, 0f7E800000;
@%p2 bra BB29_7;
bra.uni BB29_2;
BB29_7:
lg2.approx.f32 %f56, %f2;
mul.f32 %f57, %f56, 0f3F317218;
mov.f32 %f58, 0f3F317218;
add.rn.f32 %f61, %f58, %f57;
bra.uni BB29_8;
BB29_2:
rcp.rn.f32 %f12, %f2;
mov.f32 %f13, 0f3F800000;
fma.rn.f32 %f14, %f12, %f12, %f13;
sqrt.rn.f32 %f15, %f14;
add.f32 %f11, %f12, %f15;
// inline asm
rcp.approx.ftz.f32 %f10,%f11;
// inline asm
fma.rn.f32 %f3, %f2, %f10, %f2;
setp.le.f32 %p3, %f3, 0f3F266666;
setp.ge.f32 %p4, %f3, 0fBEC9BA5E;
and.pred %p5, %p4, %p3;
@%p5 bra BB29_6;
bra.uni BB29_3;
BB29_6:
add.f32 %f43, %f3, 0f40000000;
div.approx.f32 %f44, %f3, %f43;
neg.f32 %f45, %f3;
mul.rn.f32 %f46, %f45, %f44;
add.rn.f32 %f47, %f3, %f46;
mul.f32 %f48, %f47, %f47;
mov.f32 %f49, 0f3C4C4BE0;
mov.f32 %f50, 0f3B2063C3;
fma.rn.f32 %f51, %f50, %f48, %f49;
mov.f32 %f52, 0f3DAAAB50;
fma.rn.f32 %f53, %f51, %f48, %f52;
mul.f32 %f54, %f48, %f53;
fma.rn.f32 %f55, %f54, %f47, %f46;
add.f32 %f61, %f3, %f55;
bra.uni BB29_8;
BB29_3:
add.f32 %f4, %f3, 0f3F800000;
setp.gt.f32 %p6, %f4, 0f00000000;
setp.lt.f32 %p7, %f4, 0f7F800000;
and.pred %p8, %p6, %p7;
@%p8 bra BB29_5;
bra.uni BB29_4;
BB29_5:
setp.lt.f32 %p9, %f4, 0f00800000;
mul.f32 %f18, %f4, 0f4B800000;
selp.f32 %f19, %f18, %f4, %p9;
selp.f32 %f20, 0fC3170000, 0fC2FE0000, %p9;
mov.b32 %r5, %f19;
and.b32 %r6, %r5, 8388607;
or.b32 %r7, %r6, 1065353216;
mov.b32 %f21, %r7;
shr.u32 %r8, %r5, 23;
cvt.rn.f32.u32 %f22, %r8;
add.f32 %f23, %f20, %f22;
setp.gt.f32 %p10, %f21, 0f3FAE147B;
mul.f32 %f24, %f21, 0f3F000000;
add.f32 %f25, %f23, 0f3F800000;
selp.f32 %f26, %f24, %f21, %p10;
selp.f32 %f27, %f25, %f23, %p10;
add.f32 %f17, %f26, 0f3F800000;
add.f32 %f28, %f26, 0fBF800000;
// inline asm
rcp.approx.ftz.f32 %f16,%f17;
// inline asm
mul.f32 %f29, %f28, %f28;
neg.f32 %f30, %f29;
mul.rn.f32 %f31, %f16, %f30;
add.rn.f32 %f32, %f28, %f31;
mul.f32 %f33, %f32, %f32;
mov.f32 %f34, 0f3C4C6A36;
mov.f32 %f35, 0f3B1E94E6;
fma.rn.f32 %f36, %f35, %f33, %f34;
mov.f32 %f37, 0f3DAAAB1A;
fma.rn.f32 %f38, %f36, %f33, %f37;
mul.f32 %f39, %f33, %f38;
fma.rn.f32 %f40, %f39, %f32, %f31;
add.f32 %f41, %f28, %f40;
mov.f32 %f42, 0f3F317218;
fma.rn.f32 %f61, %f27, %f42, %f41;
bra.uni BB29_8;
BB29_4:
lg2.approx.f32 %f61, %f4;
BB29_8:
cvta.to.global.u64 %rd8, %rd2;
mov.b32 %r9, %f1;
and.b32 %r10, %r9, -2147483648;
mov.b32 %r11, %f61;
or.b32 %r12, %r11, %r10;
mov.b32 %f59, %r12;
setp.gtu.f32 %p11, %f2, 0f7F800000;
selp.f32 %f60, %f61, %f59, %p11;
add.s64 %rd10, %rd8, %rd6;
st.global.f32 [%rd10], %f60;
BB29_9:
ret;
}
// .globl vec_atanf
.visible .entry vec_atanf(
.param .u64 vec_atanf_param_0,
.param .u64 vec_atanf_param_1,
.param .u64 vec_atanf_param_2
)
{
.reg .pred %p<5>;
.reg .f32 %f<26>;
.reg .b32 %r<9>;
.reg .b64 %rd<11>;
ld.param.u64 %rd4, [vec_atanf_param_0];
ld.param.u64 %rd2, [vec_atanf_param_1];
ld.param.u64 %rd3, [vec_atanf_param_2];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB30_4;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f1, [%rd7];
abs.f32 %f2, %f1;
setp.leu.f32 %p2, %f2, 0f3F800000;
mov.f32 %f25, %f2;
@%p2 bra BB30_3;
rcp.rn.f32 %f3, %f2;
mov.f32 %f25, %f3;
BB30_3:
mov.f32 %f4, %f25;
cvta.to.global.u64 %rd8, %rd2;
mul.rn.f32 %f5, %f4, %f4;
mov.f32 %f6, 0fC0B59883;
mov.f32 %f7, 0fBF52C7EA;
fma.rn.f32 %f8, %f5, %f7, %f6;
mov.f32 %f9, 0fC0D21907;
fma.rn.f32 %f10, %f8, %f5, %f9;
mul.f32 %f11, %f5, %f10;
mul.f32 %f12, %f4, %f11;
add.f32 %f13, %f5, 0f41355DC0;
mov.f32 %f14, 0f41E6BD60;
fma.rn.f32 %f15, %f13, %f5, %f14;
mov.f32 %f16, 0f419D92C8;
fma.rn.f32 %f17, %f15, %f5, %f16;
rcp.rn.f32 %f18, %f17;
fma.rn.f32 %f19, %f12, %f18, %f4;
mov.f32 %f20, 0f3FC90FDB;
sub.f32 %f21, %f20, %f19;
setp.gt.f32 %p3, %f2, 0f3F800000;
selp.f32 %f22, %f21, %f19, %p3;
mov.b32 %r5, %f22;
mov.b32 %r6, %f1;
and.b32 %r7, %r6, -2147483648;
or.b32 %r8, %r5, %r7;
mov.b32 %f23, %r8;
setp.gtu.f32 %p4, %f2, 0f7F800000;
selp.f32 %f24, %f22, %f23, %p4;
add.s64 %rd10, %rd8, %rd6;
st.global.f32 [%rd10], %f24;
BB30_4:
ret;
}
// .globl vec_atanhf
.visible .entry vec_atanhf(
.param .u64 vec_atanhf_param_0,
.param .u64 vec_atanhf_param_1,
.param .u64 vec_atanhf_param_2
)
{
.reg .pred %p<12>;
.reg .f32 %f<59>;
.reg .b32 %r<13>;
.reg .b64 %rd<11>;
ld.param.u64 %rd4, [vec_atanhf_param_0];
ld.param.u64 %rd2, [vec_atanhf_param_1];
ld.param.u64 %rd3, [vec_atanhf_param_2];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB31_7;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f1, [%rd7];
abs.f32 %f10, %f1;
mov.f32 %f11, 0f3F800000;
sub.f32 %f9, %f11, %f10;
// inline asm
rcp.approx.ftz.f32 %f8,%f9;
// inline asm
add.f32 %f12, %f8, %f8;
mul.f32 %f13, %f10, %f12;
setp.gt.f32 %p2, %f10, 0f7E800000;
selp.f32 %f2, 0fC0000000, %f13, %p2;
setp.le.f32 %p3, %f2, 0f3F266666;
setp.ge.f32 %p4, %f2, 0fBEC9BA5E;
and.pred %p5, %p4, %p3;
@%p5 bra BB31_5;
bra.uni BB31_2;
BB31_5:
add.f32 %f41, %f2, 0f40000000;
div.approx.f32 %f42, %f2, %f41;
neg.f32 %f43, %f2;
mul.rn.f32 %f44, %f43, %f42;
add.rn.f32 %f45, %f2, %f44;
mul.f32 %f46, %f45, %f45;
mov.f32 %f47, 0f3C4C4BE0;
mov.f32 %f48, 0f3B2063C3;
fma.rn.f32 %f49, %f48, %f46, %f47;
mov.f32 %f50, 0f3DAAAB50;
fma.rn.f32 %f51, %f49, %f46, %f50;
mul.f32 %f52, %f46, %f51;
fma.rn.f32 %f53, %f52, %f45, %f44;
add.f32 %f58, %f2, %f53;
bra.uni BB31_6;
BB31_2:
add.f32 %f3, %f2, 0f3F800000;
setp.gt.f32 %p6, %f3, 0f00000000;
setp.lt.f32 %p7, %f3, 0f7F800000;
and.pred %p8, %p6, %p7;
@%p8 bra BB31_4;
bra.uni BB31_3;
BB31_4:
setp.lt.f32 %p9, %f3, 0f00800000;
mul.f32 %f16, %f3, 0f4B800000;
selp.f32 %f17, %f16, %f3, %p9;
selp.f32 %f18, 0fC3170000, 0fC2FE0000, %p9;
mov.b32 %r5, %f17;
and.b32 %r6, %r5, 8388607;
or.b32 %r7, %r6, 1065353216;
mov.b32 %f19, %r7;
shr.u32 %r8, %r5, 23;
cvt.rn.f32.u32 %f20, %r8;
add.f32 %f21, %f18, %f20;
setp.gt.f32 %p10, %f19, 0f3FAE147B;
mul.f32 %f22, %f19, 0f3F000000;
add.f32 %f23, %f21, 0f3F800000;
selp.f32 %f24, %f22, %f19, %p10;
selp.f32 %f25, %f23, %f21, %p10;
add.f32 %f15, %f24, 0f3F800000;
add.f32 %f26, %f24, 0fBF800000;
// inline asm
rcp.approx.ftz.f32 %f14,%f15;
// inline asm
mul.f32 %f27, %f26, %f26;
neg.f32 %f28, %f27;
mul.rn.f32 %f29, %f14, %f28;
add.rn.f32 %f30, %f26, %f29;
mul.f32 %f31, %f30, %f30;
mov.f32 %f32, 0f3C4C6A36;
mov.f32 %f33, 0f3B1E94E6;
fma.rn.f32 %f34, %f33, %f31, %f32;
mov.f32 %f35, 0f3DAAAB1A;
fma.rn.f32 %f36, %f34, %f31, %f35;
mul.f32 %f37, %f31, %f36;
fma.rn.f32 %f38, %f37, %f30, %f29;
add.f32 %f39, %f26, %f38;
mov.f32 %f40, 0f3F317218;
fma.rn.f32 %f58, %f25, %f40, %f39;
bra.uni BB31_6;
BB31_3:
lg2.approx.f32 %f58, %f3;
BB31_6:
cvta.to.global.u64 %rd8, %rd2;
mul.f32 %f54, %f58, 0f3F000000;
abs.f32 %f55, %f54;
setp.gtu.f32 %p11, %f55, 0f7F800000;
mov.b32 %r9, %f54;
mov.b32 %r10, %f1;
and.b32 %r11, %r10, -2147483648;
or.b32 %r12, %r9, %r11;
mov.b32 %f56, %r12;
selp.f32 %f57, %f54, %f56, %p11;
add.s64 %rd10, %rd8, %rd6;
st.global.f32 [%rd10], %f57;
BB31_7:
ret;
}
// .globl vec_cbrtf
.visible .entry vec_cbrtf(
.param .u64 vec_cbrtf_param_0,
.param .u64 vec_cbrtf_param_1,
.param .u64 vec_cbrtf_param_2
)
{
.reg .pred %p<4>;
.reg .f32 %f<16>;
.reg .b32 %r<6>;
.reg .b64 %rd<10>;
ld.param.u64 %rd4, [vec_cbrtf_param_0];
ld.param.u64 %rd2, [vec_cbrtf_param_1];
ld.param.u64 %rd3, [vec_cbrtf_param_2];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB32_2;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f5, [%rd7];
abs.f32 %f6, %f5;
lg2.approx.f32 %f7, %f6;
mul.f32 %f2, %f7, 0f3EAAAAAB;
// inline asm
ex2.approx.ftz.f32 %f1,%f2;
// inline asm
mul.f32 %f4, %f1, %f1;
// inline asm
rcp.approx.ftz.f32 %f3,%f4;
// inline asm
neg.f32 %f8, %f6;
fma.rn.f32 %f9, %f3, %f8, %f1;
mov.f32 %f10, 0fBEAAAAAB;
fma.rn.f32 %f11, %f9, %f10, %f1;
mov.b32 %r5, %f5;
setp.lt.s32 %p2, %r5, 0;
neg.f32 %f12, %f11;
selp.f32 %f13, %f12, %f11, %p2;
add.f32 %f14, %f5, %f5;
setp.eq.f32 %p3, %f14, %f5;
selp.f32 %f15, %f14, %f13, %p3;
cvta.to.global.u64 %rd8, %rd2;
add.s64 %rd9, %rd8, %rd6;
st.global.f32 [%rd9], %f15;
BB32_2:
ret;
}
// .globl vec_ceilf
.visible .entry vec_ceilf(
.param .u64 vec_ceilf_param_0,
.param .u64 vec_ceilf_param_1,
.param .u64 vec_ceilf_param_2
)
{
.reg .pred %p<2>;
.reg .f32 %f<3>;
.reg .b32 %r<5>;
.reg .b64 %rd<10>;
ld.param.u64 %rd4, [vec_ceilf_param_0];
ld.param.u64 %rd2, [vec_ceilf_param_1];
ld.param.u64 %rd3, [vec_ceilf_param_2];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB33_2;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f1, [%rd7];
cvt.rpi.f32.f32 %f2, %f1;
cvta.to.global.u64 %rd8, %rd2;
add.s64 %rd9, %rd8, %rd6;
st.global.f32 [%rd9], %f2;
BB33_2:
ret;
}
// .globl vec_cosf
.visible .entry vec_cosf(
.param .u64 vec_cosf_param_0,
.param .u64 vec_cosf_param_1,
.param .u64 vec_cosf_param_2
)
{
.local .align 4 .b8 __local_depot34[28];
.reg .b64 %SP;
.reg .b64 %SPL;
.reg .pred %p<15>;
.reg .f32 %f<48>;
.reg .b32 %r<97>;
.reg .b64 %rd<23>;
mov.u64 %rd22, __local_depot34;
cvta.local.u64 %SP, %rd22;
ld.param.u64 %rd9, [vec_cosf_param_0];
ld.param.u64 %rd7, [vec_cosf_param_1];
ld.param.u64 %rd8, [vec_cosf_param_2];
add.u64 %rd10, %SP, 0;
cvta.to.local.u64 %rd1, %rd10;
mov.u32 %r39, %ntid.x;
mov.u32 %r40, %ctaid.x;
mov.u32 %r41, %tid.x;
mad.lo.s32 %r1, %r39, %r40, %r41;
cvt.s64.s32 %rd11, %r1;
setp.ge.u64 %p1, %rd11, %rd9;
@%p1 bra BB34_24;
cvta.to.global.u64 %rd12, %rd8;
mul.wide.s32 %rd13, %r1, 4;
add.s64 %rd14, %rd12, %rd13;
ld.global.f32 %f43, [%rd14];
abs.f32 %f19, %f43;
setp.neu.f32 %p2, %f19, 0f7F800000;
@%p2 bra BB34_3;
mov.f32 %f20, 0f00000000;
mul.rn.f32 %f43, %f43, %f20;
BB34_3:
mul.f32 %f21, %f43, 0f3F22F983;
cvt.rni.s32.f32 %r96, %f21;
cvt.rn.f32.s32 %f22, %r96;
neg.f32 %f23, %f22;
mov.f32 %f24, 0f3FC90FDA;
fma.rn.f32 %f25, %f23, %f24, %f43;
mov.f32 %f26, 0f33A22168;
fma.rn.f32 %f27, %f23, %f26, %f25;
mov.f32 %f28, 0f27C234C5;
fma.rn.f32 %f44, %f23, %f28, %f27;
abs.f32 %f29, %f43;
setp.leu.f32 %p3, %f29, 0f47CE4780;
@%p3 bra BB34_13;
mov.b32 %r3, %f43;
shr.u32 %r4, %r3, 23;
bfe.u32 %r44, %r3, 23, 8;
add.s32 %r45, %r44, -128;
shl.b32 %r46, %r3, 8;
or.b32 %r5, %r46, -2147483648;
shr.u32 %r6, %r45, 5;
mov.u32 %r88, 0;
mov.u64 %rd20, __cudart_i2opi_f;
mov.u32 %r87, -6;
mov.u64 %rd21, %rd1;
BB34_5:
.pragma "nounroll";
mov.u64 %rd3, %rd21;
ld.const.u32 %r49, [%rd20];
// inline asm
{
mad.lo.cc.u32 %r47, %r49, %r5, %r88;
madc.hi.u32 %r88, %r49, %r5, 0;
}
// inline asm
st.local.u32 [%rd3], %r47;
add.s64 %rd4, %rd3, 4;
add.s64 %rd20, %rd20, 4;
add.s32 %r87, %r87, 1;
setp.ne.s32 %p4, %r87, 0;
mov.u64 %rd21, %rd4;
@%p4 bra BB34_5;
and.b32 %r11, %r3, -2147483648;
st.local.u32 [%rd1+24], %r88;
mov.u32 %r52, 6;
sub.s32 %r53, %r52, %r6;
mul.wide.s32 %rd16, %r53, 4;
add.s64 %rd6, %rd1, %rd16;
ld.local.u32 %r89, [%rd6];
ld.local.u32 %r90, [%rd6+-4];
and.b32 %r14, %r4, 31;
setp.eq.s32 %p5, %r14, 0;
@%p5 bra BB34_8;
mov.u32 %r54, 32;
sub.s32 %r55, %r54, %r14;
shr.u32 %r56, %r90, %r55;
shl.b32 %r57, %r89, %r14;
add.s32 %r89, %r56, %r57;
ld.local.u32 %r58, [%rd6+-8];
shr.u32 %r59, %r58, %r55;
shl.b32 %r60, %r90, %r14;
add.s32 %r90, %r59, %r60;
BB34_8:
shr.u32 %r61, %r90, 30;
shl.b32 %r62, %r89, 2;
add.s32 %r91, %r61, %r62;
shl.b32 %r20, %r90, 2;
shr.u32 %r63, %r91, 31;
shr.u32 %r64, %r89, 30;
add.s32 %r21, %r63, %r64;
setp.eq.s32 %p6, %r63, 0;
mov.u32 %r92, %r11;
mov.u32 %r93, %r20;
@%p6 bra BB34_10;
not.b32 %r65, %r91;
neg.s32 %r22, %r20;
setp.eq.s32 %p7, %r20, 0;
selp.u32 %r66, 1, 0, %p7;
add.s32 %r91, %r66, %r65;
xor.b32 %r24, %r11, -2147483648;
mov.u32 %r92, %r24;
mov.u32 %r93, %r22;
BB34_10:
mov.u32 %r26, %r92;
neg.s32 %r67, %r21;
setp.eq.s32 %p8, %r11, 0;
selp.b32 %r96, %r21, %r67, %p8;
clz.b32 %r95, %r91;
setp.eq.s32 %p9, %r95, 0;
shl.b32 %r68, %r91, %r95;
mov.u32 %r69, 32;
sub.s32 %r70, %r69, %r95;
shr.u32 %r71, %r93, %r70;
add.s32 %r72, %r71, %r68;
selp.b32 %r30, %r91, %r72, %p9;
mov.u32 %r73, -921707870;
mul.hi.u32 %r94, %r30, %r73;
setp.lt.s32 %p10, %r94, 1;
@%p10 bra BB34_12;
mul.lo.s32 %r74, %r30, -921707870;
shr.u32 %r75, %r74, 31;
shl.b32 %r76, %r94, 1;
add.s32 %r94, %r75, %r76;
add.s32 %r95, %r95, 1;
BB34_12:
mov.u32 %r77, 126;
sub.s32 %r78, %r77, %r95;
shl.b32 %r79, %r78, 23;
add.s32 %r80, %r94, 1;
shr.u32 %r81, %r80, 7;
add.s32 %r82, %r81, 1;
shr.u32 %r83, %r82, 1;
add.s32 %r84, %r83, %r79;
or.b32 %r85, %r84, %r26;
mov.b32 %f44, %r85;
BB34_13:
mul.rn.f32 %f7, %f44, %f44;
add.s32 %r37, %r96, 1;
and.b32 %r38, %r37, 1;
setp.eq.s32 %p11, %r38, 0;
@%p11 bra BB34_15;
mov.f32 %f30, 0fBAB6061A;
mov.f32 %f31, 0f37CCF5CE;
fma.rn.f32 %f45, %f31, %f7, %f30;
bra.uni BB34_16;
BB34_15:
mov.f32 %f32, 0f3C08839E;
mov.f32 %f33, 0fB94CA1F9;
fma.rn.f32 %f45, %f33, %f7, %f32;
BB34_16:
@%p11 bra BB34_18;
mov.f32 %f34, 0f3D2AAAA5;
fma.rn.f32 %f35, %f45, %f7, %f34;
mov.f32 %f36, 0fBF000000;
fma.rn.f32 %f46, %f35, %f7, %f36;
bra.uni BB34_19;
BB34_18:
mov.f32 %f37, 0fBE2AAAA3;
fma.rn.f32 %f38, %f45, %f7, %f37;
mov.f32 %f39, 0f00000000;
fma.rn.f32 %f46, %f38, %f7, %f39;
BB34_19:
fma.rn.f32 %f47, %f46, %f44, %f44;
@%p11 bra BB34_21;
mov.f32 %f40, 0f3F800000;
fma.rn.f32 %f47, %f46, %f7, %f40;
BB34_21:
and.b32 %r86, %r37, 2;
setp.eq.s32 %p14, %r86, 0;
@%p14 bra BB34_23;
mov.f32 %f41, 0f00000000;
mov.f32 %f42, 0fBF800000;
fma.rn.f32 %f47, %f47, %f42, %f41;
BB34_23:
cvta.to.global.u64 %rd17, %rd7;
add.s64 %rd19, %rd17, %rd13;
st.global.f32 [%rd19], %f47;
BB34_24:
ret;
}
// .globl vec_coshf
.visible .entry vec_coshf(
.param .u64 vec_coshf_param_0,
.param .u64 vec_coshf_param_1,
.param .u64 vec_coshf_param_2
)
{
.reg .pred %p<3>;
.reg .f32 %f<19>;
.reg .b32 %r<5>;
.reg .b64 %rd<10>;
ld.param.u64 %rd4, [vec_coshf_param_0];
ld.param.u64 %rd2, [vec_coshf_param_1];
ld.param.u64 %rd3, [vec_coshf_param_2];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB35_2;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f3, [%rd7];
abs.f32 %f4, %f3;
mul.f32 %f5, %f4, 0f3FB8AA3B;
cvt.rzi.f32.f32 %f6, %f5;
mov.f32 %f7, 0fBF317200;
fma.rn.f32 %f8, %f6, %f7, %f4;
mov.f32 %f9, 0fB5BFBE8E;
fma.rn.f32 %f10, %f6, %f9, %f8;
mul.f32 %f2, %f10, 0f3FB8AA3B;
// inline asm
ex2.approx.ftz.f32 %f1,%f2;
// inline asm
add.f32 %f11, %f6, 0fC0000000;
ex2.approx.f32 %f12, %f11;
mul.f32 %f13, %f1, %f12;
mov.f32 %f14, 0f3E000000;
div.approx.f32 %f15, %f14, %f13;
mov.f32 %f16, 0f40000000;
fma.rn.f32 %f17, %f16, %f13, %f15;
setp.ltu.f32 %p2, %f4, 0f42B40000;
selp.f32 %f18, %f17, 0f7F800000, %p2;
cvta.to.global.u64 %rd8, %rd2;
add.s64 %rd9, %rd8, %rd6;
st.global.f32 [%rd9], %f18;
BB35_2:
ret;
}
// .globl vec_cospif
.visible .entry vec_cospif(
.param .u64 vec_cospif_param_0,
.param .u64 vec_cospif_param_1,
.param .u64 vec_cospif_param_2
)
{
.reg .pred %p<7>;
.reg .f32 %f<43>;
.reg .b32 %r<9>;
.reg .b64 %rd<11>;
ld.param.u64 %rd4, [vec_cospif_param_0];
ld.param.u64 %rd2, [vec_cospif_param_1];
ld.param.u64 %rd3, [vec_cospif_param_2];
mov.u32 %r3, %tid.x;
mov.u32 %r4, %ntid.x;
mov.u32 %r5, %ctaid.x;
mad.lo.s32 %r6, %r4, %r5, %r3;
cvt.s64.s32 %rd1, %r6;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB36_14;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f39, [%rd7];
abs.f32 %f17, %f39;
setp.leu.f32 %p2, %f17, 0f4B800000;
@%p2 bra BB36_3;
mov.f32 %f18, 0f00000000;
mul.rn.f32 %f39, %f39, %f18;
BB36_3:
add.f32 %f19, %f39, %f39;
cvt.rni.f32.f32 %f20, %f19;
cvt.rzi.s32.f32 %r7, %f20;
neg.f32 %f21, %f20;
mov.f32 %f22, 0f3F000000;
fma.rn.f32 %f23, %f21, %f22, %f39;
mul.f32 %f24, %f23, 0f34222169;
mov.f32 %f25, 0f40490FDA;
fma.rn.f32 %f4, %f23, %f25, %f24;
add.s32 %r1, %r7, 1;
mul.rn.f32 %f5, %f4, %f4;
and.b32 %r2, %r1, 1;
setp.eq.s32 %p3, %r2, 0;
@%p3 bra BB36_5;
mov.f32 %f26, 0fBAB6061A;
mov.f32 %f27, 0f37CCF5CE;
fma.rn.f32 %f40, %f27, %f5, %f26;
bra.uni BB36_6;
BB36_5:
mov.f32 %f28, 0f3C08839E;
mov.f32 %f29, 0fB94CA1F9;
fma.rn.f32 %f40, %f29, %f5, %f28;
BB36_6:
@%p3 bra BB36_8;
mov.f32 %f30, 0f3D2AAAA5;
fma.rn.f32 %f31, %f40, %f5, %f30;
mov.f32 %f32, 0fBF000000;
fma.rn.f32 %f41, %f31, %f5, %f32;
bra.uni BB36_9;
BB36_8:
mov.f32 %f33, 0fBE2AAAA3;
fma.rn.f32 %f34, %f40, %f5, %f33;
mov.f32 %f35, 0f00000000;
fma.rn.f32 %f41, %f34, %f5, %f35;
BB36_9:
fma.rn.f32 %f42, %f41, %f4, %f4;
@%p3 bra BB36_11;
mov.f32 %f36, 0f3F800000;
fma.rn.f32 %f42, %f41, %f5, %f36;
BB36_11:
and.b32 %r8, %r1, 2;
setp.eq.s32 %p6, %r8, 0;
@%p6 bra BB36_13;
mov.f32 %f37, 0f00000000;
mov.f32 %f38, 0fBF800000;
fma.rn.f32 %f42, %f42, %f38, %f37;
BB36_13:
cvta.to.global.u64 %rd8, %rd2;
add.s64 %rd10, %rd8, %rd6;
st.global.f32 [%rd10], %f42;
BB36_14:
ret;
}
// .globl vec_erfcf
.visible .entry vec_erfcf(
.param .u64 vec_erfcf_param_0,
.param .u64 vec_erfcf_param_1,
.param .u64 vec_erfcf_param_2
)
{
.reg .pred %p<4>;
.reg .f32 %f<63>;
.reg .b32 %r<5>;
.reg .b64 %rd<10>;
ld.param.u64 %rd4, [vec_erfcf_param_0];
ld.param.u64 %rd2, [vec_erfcf_param_1];
ld.param.u64 %rd3, [vec_erfcf_param_2];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB37_2;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f7, [%rd7];
abs.f32 %f8, %f7;
add.f32 %f2, %f8, 0f40800000;
// inline asm
rcp.approx.ftz.f32 %f1,%f2;
// inline asm
add.f32 %f9, %f8, 0fC0800000;
mul.rn.f32 %f10, %f9, %f1;
add.f32 %f11, %f10, 0f3F800000;
mov.f32 %f12, 0fC0800000;
fma.rn.f32 %f13, %f12, %f11, %f8;
neg.f32 %f14, %f10;
fma.rn.f32 %f15, %f14, %f8, %f13;
fma.rn.f32 %f16, %f1, %f15, %f10;
mov.f32 %f17, 0f3BE6E05B;
mov.f32 %f18, 0f3A69A091;
fma.rn.f32 %f19, %f18, %f16, %f17;
mov.f32 %f20, 0fBC81FB4B;
fma.rn.f32 %f21, %f19, %f16, %f20;
mov.f32 %f22, 0f3D15373B;
fma.rn.f32 %f23, %f21, %f16, %f22;
mov.f32 %f24, 0fBD887C5A;
fma.rn.f32 %f25, %f23, %f16, %f24;
mov.f32 %f26, 0f3DC021D5;
fma.rn.f32 %f27, %f25, %f16, %f26;
mov.f32 %f28, 0fBDCED424;
fma.rn.f32 %f29, %f27, %f16, %f28;
mov.f32 %f30, 0f3D8B74DE;
fma.rn.f32 %f31, %f29, %f16, %f30;
mov.f32 %f32, 0f3C7BF170;
fma.rn.f32 %f33, %f31, %f16, %f32;
mov.f32 %f34, 0fBE0EF8D4;
fma.rn.f32 %f35, %f33, %f16, %f34;
mov.f32 %f36, 0f3F9DD2C9;
fma.rn.f32 %f37, %f35, %f16, %f36;
mov.f32 %f38, 0f3F800000;
mov.f32 %f39, 0f40000000;
fma.rn.f32 %f4, %f39, %f8, %f38;
// inline asm
rcp.approx.ftz.f32 %f3,%f4;
// inline asm
mul.rn.f32 %f40, %f37, %f3;
mul.f32 %f41, %f40, 0fC0000000;
fma.rn.f32 %f42, %f8, %f41, %f37;
sub.f32 %f43, %f42, %f40;
fma.rn.f32 %f44, %f43, %f3, %f40;
mul.f32 %f45, %f8, %f8;
neg.f32 %f46, %f45;
mul.f32 %f47, %f45, 0fBFB8AA3B;
cvt.rzi.f32.f32 %f48, %f47;
mov.f32 %f49, 0fBF317200;
fma.rn.f32 %f50, %f48, %f49, %f46;
mov.f32 %f51, 0fB5BFBE8E;
fma.rn.f32 %f52, %f48, %f51, %f50;
mul.f32 %f6, %f52, 0f3FB8AA3B;
// inline asm
ex2.approx.ftz.f32 %f5,%f6;
// inline asm
add.f32 %f53, %f48, 0f00000000;
ex2.approx.f32 %f54, %f53;
mul.f32 %f55, %f5, %f54;
neg.f32 %f56, %f8;
fma.rn.f32 %f57, %f56, %f8, %f45;
fma.rn.f32 %f58, %f55, %f57, %f55;
mul.f32 %f59, %f44, %f58;
setp.gt.f32 %p2, %f8, 0f4120E148;
selp.f32 %f60, 0f00000000, %f59, %p2;
setp.lt.f32 %p3, %f7, 0f00000000;
sub.f32 %f61, %f39, %f60;
selp.f32 %f62, %f61, %f60, %p3;
cvta.to.global.u64 %rd8, %rd2;
add.s64 %rd9, %rd8, %rd6;
st.global.f32 [%rd9], %f62;
BB37_2:
ret;
}
// .globl vec_erfcinvf
.visible .entry vec_erfcinvf(
.param .u64 vec_erfcinvf_param_0,
.param .u64 vec_erfcinvf_param_1,
.param .u64 vec_erfcinvf_param_2
)
{
.reg .pred %p<6>;
.reg .f32 %f<56>;
.reg .b32 %r<5>;
.reg .b64 %rd<11>;
ld.param.u64 %rd4, [vec_erfcinvf_param_0];
ld.param.u64 %rd2, [vec_erfcinvf_param_1];
ld.param.u64 %rd3, [vec_erfcinvf_param_2];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB38_5;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f1, [%rd7];
neg.f32 %f2, %f1;
mov.f32 %f7, 0f40000000;
add.rn.f32 %f3, %f7, %f2;
setp.le.f32 %p2, %f1, 0f3FFF9097;
setp.ge.f32 %p3, %f1, 0f3B5ED289;
and.pred %p4, %p3, %p2;
@%p4 bra BB38_3;
bra.uni BB38_2;
BB38_3:
mul.rn.f32 %f34, %f3, %f1;
// inline asm
lg2.approx.ftz.f32 %f33,%f34;
// inline asm
neg.f32 %f35, %f33;
mov.f32 %f36, 0f3221F645;
mov.f32 %f37, 0fAF8A6370;
fma.rn.f32 %f38, %f37, %f35, %f36;
mov.f32 %f39, 0fB4016FDA;
fma.rn.f32 %f40, %f38, %f35, %f39;
mov.f32 %f41, 0f3468F846;
fma.rn.f32 %f42, %f40, %f35, %f41;
mov.f32 %f43, 0f370742AA;
fma.rn.f32 %f44, %f42, %f35, %f43;
mov.f32 %f45, 0fB804DB4D;
fma.rn.f32 %f46, %f44, %f35, %f45;
mov.f32 %f47, 0fBA4AFEA1;
fma.rn.f32 %f48, %f46, %f35, %f47;
mov.f32 %f49, 0f3BB5C027;
fma.rn.f32 %f50, %f48, %f35, %f49;
mov.f32 %f51, 0f3E24AE0F;
fma.rn.f32 %f52, %f50, %f35, %f51;
mov.f32 %f53, 0f3F62DFC4;
fma.rn.f32 %f54, %f52, %f35, %f53;
fma.rn.f32 %f55, %f54, %f2, %f54;
bra.uni BB38_4;
BB38_2:
setp.gt.f32 %p5, %f1, 0f3F800000;
selp.f32 %f12, %f3, %f1, %p5;
lg2.approx.f32 %f13, %f12;
neg.f32 %f9, %f13;
// inline asm
rsqrt.approx.ftz.f32 %f8,%f9;
// inline asm
mov.f32 %f14, 0f42FEF829;
mov.f32 %f15, 0fC27C73F1;
fma.rn.f32 %f16, %f15, %f8, %f14;
mov.f32 %f17, 0fC2E4361C;
fma.rn.f32 %f18, %f16, %f8, %f17;
mov.f32 %f19, 0f42714D9B;
fma.rn.f32 %f20, %f18, %f8, %f19;
mov.f32 %f21, 0fC1AE51B3;
fma.rn.f32 %f22, %f20, %f8, %f21;
mov.f32 %f23, 0f40CEF504;
fma.rn.f32 %f24, %f22, %f8, %f23;
mov.f32 %f25, 0fBFEA9E05;
fma.rn.f32 %f26, %f24, %f8, %f25;
mov.f32 %f27, 0fBCF871F4;
fma.rn.f32 %f28, %f26, %f8, %f27;
mov.f32 %f29, 0f3F553775;
fma.rn.f32 %f30, %f28, %f8, %f29;
// inline asm
rcp.approx.ftz.f32 %f10,%f8;
// inline asm
mul.rn.f32 %f31, %f30, %f10;
neg.f32 %f32, %f31;
selp.f32 %f55, %f32, %f31, %p5;
BB38_4:
cvta.to.global.u64 %rd8, %rd2;
add.s64 %rd10, %rd8, %rd6;
st.global.f32 [%rd10], %f55;
BB38_5:
ret;
}
// .globl vec_erfcxf
.visible .entry vec_erfcxf(
.param .u64 vec_erfcxf_param_0,
.param .u64 vec_erfcxf_param_1,
.param .u64 vec_erfcxf_param_2
)
{
.reg .pred %p<7>;
.reg .f32 %f<81>;
.reg .b32 %r<5>;
.reg .b64 %rd<11>;
ld.param.u64 %rd4, [vec_erfcxf_param_0];
ld.param.u64 %rd2, [vec_erfcxf_param_1];
ld.param.u64 %rd3, [vec_erfcxf_param_2];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB39_7;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f1, [%rd7];
abs.f32 %f2, %f1;
setp.lt.f32 %p2, %f2, 0f4120E148;
@%p2 bra BB39_3;
bra.uni BB39_2;
BB39_3:
add.f32 %f23, %f2, 0f40800000;
// inline asm
rcp.approx.ftz.f32 %f22,%f23;
// inline asm
add.f32 %f26, %f2, 0fC0800000;
mul.rn.f32 %f27, %f26, %f22;
add.f32 %f28, %f27, 0f3F800000;
mov.f32 %f29, 0fC0800000;
fma.rn.f32 %f30, %f29, %f28, %f2;
neg.f32 %f31, %f27;
fma.rn.f32 %f32, %f31, %f2, %f30;
fma.rn.f32 %f33, %f22, %f32, %f27;
mov.f32 %f34, 0f3BE6E05B;
mov.f32 %f35, 0f3A69A091;
fma.rn.f32 %f36, %f35, %f33, %f34;
mov.f32 %f37, 0fBC81FB4B;
fma.rn.f32 %f38, %f36, %f33, %f37;
mov.f32 %f39, 0f3D15373B;
fma.rn.f32 %f40, %f38, %f33, %f39;
mov.f32 %f41, 0fBD887C5A;
fma.rn.f32 %f42, %f40, %f33, %f41;
mov.f32 %f43, 0f3DC021D5;
fma.rn.f32 %f44, %f42, %f33, %f43;
mov.f32 %f45, 0fBDCED424;
fma.rn.f32 %f46, %f44, %f33, %f45;
mov.f32 %f47, 0f3D8B74DE;
fma.rn.f32 %f48, %f46, %f33, %f47;
mov.f32 %f49, 0f3C7BF170;
fma.rn.f32 %f50, %f48, %f33, %f49;
mov.f32 %f51, 0fBE0EF8D4;
fma.rn.f32 %f52, %f50, %f33, %f51;
mov.f32 %f53, 0f3F9DD2C9;
fma.rn.f32 %f54, %f52, %f33, %f53;
mov.f32 %f55, 0f3F800000;
mov.f32 %f56, 0f40000000;
fma.rn.f32 %f25, %f56, %f2, %f55;
// inline asm
rcp.approx.ftz.f32 %f24,%f25;
// inline asm
mul.rn.f32 %f57, %f54, %f24;
mul.f32 %f58, %f57, 0fC0000000;
fma.rn.f32 %f59, %f2, %f58, %f54;
sub.f32 %f60, %f59, %f57;
fma.rn.f32 %f80, %f60, %f24, %f57;
bra.uni BB39_4;
BB39_2:
mul.f32 %f8, %f2, 0f3E800000;
mov.f32 %f9, 0f3E800000;
div.approx.f32 %f10, %f9, %f8;
mul.f32 %f11, %f10, %f10;
mov.f32 %f12, 0fBFF00000;
mov.f32 %f13, 0f40D20000;
fma.rn.f32 %f14, %f13, %f11, %f12;
mov.f32 %f15, 0f3F400000;
fma.rn.f32 %f16, %f14, %f11, %f15;
mov.f32 %f17, 0fBF000000;
fma.rn.f32 %f18, %f16, %f11, %f17;
mov.f32 %f19, 0f3F800000;
fma.rn.f32 %f20, %f18, %f11, %f19;
mul.f32 %f21, %f10, 0f3F106EBB;
mul.f32 %f80, %f21, %f20;
BB39_4:
setp.geu.f32 %p3, %f1, 0f00000000;
@%p3 bra BB39_6;
mul.rz.f32 %f63, %f2, %f2;
neg.f32 %f64, %f63;
fma.rn.f32 %f65, %f2, %f2, %f64;
mul.f32 %f66, %f63, 0f3FB8AA3B;
cvt.rzi.f32.f32 %f67, %f66;
mov.f32 %f68, 0fBF317200;
fma.rn.f32 %f69, %f67, %f68, %f63;
mov.f32 %f70, 0fB5BFBE8E;
fma.rn.f32 %f71, %f67, %f70, %f69;
mul.f32 %f62, %f71, 0f3FB8AA3B;
// inline asm
ex2.approx.ftz.f32 %f61,%f62;
// inline asm
add.f32 %f72, %f67, 0f00000000;
ex2.approx.f32 %f73, %f72;
mul.f32 %f74, %f61, %f73;
setp.lt.f32 %p4, %f63, 0fC2D20000;
selp.f32 %f75, 0f00000000, %f74, %p4;
setp.gt.f32 %p5, %f63, 0f42D20000;
selp.f32 %f76, 0f7F800000, %f75, %p5;
add.f32 %f77, %f76, %f76;
fma.rn.f32 %f78, %f77, %f65, %f77;
sub.f32 %f79, %f78, %f80;
setp.eq.f32 %p6, %f77, 0f7F800000;
selp.f32 %f80, %f77, %f79, %p6;
BB39_6:
cvta.to.global.u64 %rd8, %rd2;
add.s64 %rd10, %rd8, %rd6;
st.global.f32 [%rd10], %f80;
BB39_7:
ret;
}
// .globl vec_erff
.visible .entry vec_erff(
.param .u64 vec_erff_param_0,
.param .u64 vec_erff_param_1,
.param .u64 vec_erff_param_2
)
{
.reg .pred %p<4>;
.reg .f32 %f<37>;
.reg .b32 %r<10>;
.reg .b64 %rd<11>;
ld.param.u64 %rd4, [vec_erff_param_0];
ld.param.u64 %rd2, [vec_erff_param_1];
ld.param.u64 %rd3, [vec_erff_param_2];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB40_5;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f1, [%rd7];
abs.f32 %f2, %f1;
setp.ltu.f32 %p2, %f2, 0f3F800000;
@%p2 bra BB40_3;
bra.uni BB40_2;
BB40_3:
mul.f32 %f24, %f1, %f1;
mov.f32 %f25, 0f3BA0C9F8;
mov.f32 %f26, 0fBA1268FB;
fma.rn.f32 %f27, %f26, %f24, %f25;
mov.f32 %f28, 0fBCDABFD4;
fma.rn.f32 %f29, %f27, %f24, %f28;
mov.f32 %f30, 0f3DE70331;
fma.rn.f32 %f31, %f29, %f24, %f30;
mov.f32 %f32, 0fBEC09330;
fma.rn.f32 %f33, %f31, %f24, %f32;
mov.f32 %f34, 0f3F906EBA;
fma.rn.f32 %f35, %f33, %f24, %f34;
mul.f32 %f36, %f1, %f35;
bra.uni BB40_4;
BB40_2:
mov.f32 %f8, 0f3A03BB71;
mov.f32 %f9, 0fB7B730FB;
fma.rn.f32 %f10, %f9, %f2, %f8;
mov.f32 %f11, 0fBBACA3B3;
fma.rn.f32 %f12, %f10, %f2, %f11;
mov.f32 %f13, 0f3D0A7445;
fma.rn.f32 %f14, %f12, %f2, %f13;
mov.f32 %f15, 0fBE1B3B75;
fma.rn.f32 %f16, %f14, %f2, %f15;
mov.f32 %f17, 0fBF6B385A;
fma.rn.f32 %f18, %f16, %f2, %f17;
mov.f32 %f19, 0fBFD0316E;
fma.rn.f32 %f20, %f18, %f2, %f19;
mov.f32 %f21, 0fBA031CCE;
fma.rn.f32 %f7, %f20, %f2, %f21;
// inline asm
ex2.approx.ftz.f32 %f6,%f7;
// inline asm
mov.f32 %f22, 0f3F800000;
sub.f32 %f23, %f22, %f6;
mov.b32 %r5, %f23;
setp.ltu.f32 %p3, %f2, 0f407AD445;
selp.b32 %r6, %r5, 1065353216, %p3;
mov.b32 %r7, %f1;
and.b32 %r8, %r7, -2147483648;
or.b32 %r9, %r6, %r8;
mov.b32 %f36, %r9;
BB40_4:
cvta.to.global.u64 %rd8, %rd2;
add.s64 %rd10, %rd8, %rd6;
st.global.f32 [%rd10], %f36;
BB40_5:
ret;
}
// .globl vec_erfinvf
.visible .entry vec_erfinvf(
.param .u64 vec_erfinvf_param_0,
.param .u64 vec_erfinvf_param_1,
.param .u64 vec_erfinvf_param_2
)
{
.reg .pred %p<3>;
.reg .f32 %f<45>;
.reg .b32 %r<9>;
.reg .b64 %rd<11>;
ld.param.u64 %rd4, [vec_erfinvf_param_0];
ld.param.u64 %rd2, [vec_erfinvf_param_1];
ld.param.u64 %rd3, [vec_erfinvf_param_2];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB41_5;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f1, [%rd7];
neg.f32 %f8, %f1;
mov.f32 %f9, 0f3F800000;
fma.rn.f32 %f7, %f1, %f8, %f9;
// inline asm
lg2.approx.ftz.f32 %f6,%f7;
// inline asm
neg.f32 %f2, %f6;
setp.lt.f32 %p2, %f6, 0fC1033333;
@%p2 bra BB41_3;
bra.uni BB41_2;
BB41_3:
// inline asm
rsqrt.approx.ftz.f32 %f29,%f2;
// inline asm
mov.f32 %f31, 0fBF29BAA5;
mov.f32 %f32, 0fBF1704A1;
fma.rn.f32 %f33, %f32, %f29, %f31;
mov.f32 %f34, 0f3FCC6ADC;
fma.rn.f32 %f35, %f33, %f29, %f34;
mov.f32 %f36, 0fBF2CDAED;
fma.rn.f32 %f37, %f35, %f29, %f36;
mov.f32 %f38, 0fBDC30537;
fma.rn.f32 %f39, %f37, %f29, %f38;
mov.f32 %f40, 0f3F55D9B9;
fma.rn.f32 %f41, %f39, %f29, %f40;
rcp.rn.f32 %f42, %f29;
mul.f32 %f43, %f41, %f42;
mov.b32 %r5, %f43;
mov.b32 %r6, %f1;
and.b32 %r7, %r6, -2147483648;
or.b32 %r8, %r5, %r7;
mov.b32 %f44, %r8;
bra.uni BB41_4;
BB41_2:
mov.f32 %f10, 0f3221F645;
mov.f32 %f11, 0fAF8A6370;
fma.rn.f32 %f12, %f11, %f2, %f10;
mov.f32 %f13, 0fB4016FDA;
fma.rn.f32 %f14, %f12, %f2, %f13;
mov.f32 %f15, 0f3468F846;
fma.rn.f32 %f16, %f14, %f2, %f15;
mov.f32 %f17, 0f370742AA;
fma.rn.f32 %f18, %f16, %f2, %f17;
mov.f32 %f19, 0fB804DB4D;
fma.rn.f32 %f20, %f18, %f2, %f19;
mov.f32 %f21, 0fBA4AFEA1;
fma.rn.f32 %f22, %f20, %f2, %f21;
mov.f32 %f23, 0f3BB5C027;
fma.rn.f32 %f24, %f22, %f2, %f23;
mov.f32 %f25, 0f3E24AE0F;
fma.rn.f32 %f26, %f24, %f2, %f25;
mov.f32 %f27, 0f3F62DFC4;
fma.rn.f32 %f28, %f26, %f2, %f27;
mul.f32 %f44, %f1, %f28;
BB41_4:
cvta.to.global.u64 %rd8, %rd2;
add.s64 %rd10, %rd8, %rd6;
st.global.f32 [%rd10], %f44;
BB41_5:
ret;
}
// .globl vec_exp10f
.visible .entry vec_exp10f(
.param .u64 vec_exp10f_param_0,
.param .u64 vec_exp10f_param_1,
.param .u64 vec_exp10f_param_2
)
{
.reg .pred %p<4>;
.reg .f32 %f<14>;
.reg .b32 %r<5>;
.reg .b64 %rd<10>;
ld.param.u64 %rd4, [vec_exp10f_param_0];
ld.param.u64 %rd2, [vec_exp10f_param_1];
ld.param.u64 %rd3, [vec_exp10f_param_2];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB42_2;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f3, [%rd7];
mul.f32 %f4, %f3, 0f40549A78;
cvt.rzi.f32.f32 %f5, %f4;
mov.f32 %f6, 0fBE9A2080;
fma.rn.f32 %f7, %f5, %f6, %f3;
mov.f32 %f8, 0fB55427DE;
fma.rn.f32 %f9, %f5, %f8, %f7;
mul.f32 %f2, %f9, 0f40549A78;
// inline asm
ex2.approx.ftz.f32 %f1,%f2;
// inline asm
ex2.approx.f32 %f10, %f5;
mul.f32 %f11, %f1, %f10;
setp.lt.f32 %p2, %f3, 0fC2380000;
selp.f32 %f12, 0f00000000, %f11, %p2;
setp.gt.f32 %p3, %f3, 0f42380000;
selp.f32 %f13, 0f7F800000, %f12, %p3;
cvta.to.global.u64 %rd8, %rd2;
add.s64 %rd9, %rd8, %rd6;
st.global.f32 [%rd9], %f13;
BB42_2:
ret;
}
// .globl vec_exp2f
.visible .entry vec_exp2f(
.param .u64 vec_exp2f_param_0,
.param .u64 vec_exp2f_param_1,
.param .u64 vec_exp2f_param_2
)
{
.reg .pred %p<2>;
.reg .f32 %f<3>;
.reg .b32 %r<5>;
.reg .b64 %rd<10>;
ld.param.u64 %rd4, [vec_exp2f_param_0];
ld.param.u64 %rd2, [vec_exp2f_param_1];
ld.param.u64 %rd3, [vec_exp2f_param_2];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB43_2;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f1, [%rd7];
ex2.approx.f32 %f2, %f1;
cvta.to.global.u64 %rd8, %rd2;
add.s64 %rd9, %rd8, %rd6;
st.global.f32 [%rd9], %f2;
BB43_2:
ret;
}
// .globl vec_expf
.visible .entry vec_expf(
.param .u64 vec_expf_param_0,
.param .u64 vec_expf_param_1,
.param .u64 vec_expf_param_2
)
{
.reg .pred %p<4>;
.reg .f32 %f<15>;
.reg .b32 %r<5>;
.reg .b64 %rd<10>;
ld.param.u64 %rd4, [vec_expf_param_0];
ld.param.u64 %rd2, [vec_expf_param_1];
ld.param.u64 %rd3, [vec_expf_param_2];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB44_2;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f3, [%rd7];
mul.f32 %f4, %f3, 0f3FB8AA3B;
cvt.rzi.f32.f32 %f5, %f4;
mov.f32 %f6, 0fBF317200;
fma.rn.f32 %f7, %f5, %f6, %f3;
mov.f32 %f8, 0fB5BFBE8E;
fma.rn.f32 %f9, %f5, %f8, %f7;
mul.f32 %f2, %f9, 0f3FB8AA3B;
// inline asm
ex2.approx.ftz.f32 %f1,%f2;
// inline asm
add.f32 %f10, %f5, 0f00000000;
ex2.approx.f32 %f11, %f10;
mul.f32 %f12, %f1, %f11;
setp.lt.f32 %p2, %f3, 0fC2D20000;
selp.f32 %f13, 0f00000000, %f12, %p2;
setp.gt.f32 %p3, %f3, 0f42D20000;
selp.f32 %f14, 0f7F800000, %f13, %p3;
cvta.to.global.u64 %rd8, %rd2;
add.s64 %rd9, %rd8, %rd6;
st.global.f32 [%rd9], %f14;
BB44_2:
ret;
}
// .globl vec_expm1f
.visible .entry vec_expm1f(
.param .u64 vec_expm1f_param_0,
.param .u64 vec_expm1f_param_1,
.param .u64 vec_expm1f_param_2
)
{
.reg .pred %p<7>;
.reg .f32 %f<33>;
.reg .b32 %r<5>;
.reg .b64 %rd<10>;
ld.param.u64 %rd4, [vec_expm1f_param_0];
ld.param.u64 %rd2, [vec_expm1f_param_1];
ld.param.u64 %rd3, [vec_expm1f_param_2];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB45_2;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f1, [%rd7];
mul.f32 %f2, %f1, 0f3FB8AA3B;
cvt.rni.f32.f32 %f3, %f2;
abs.f32 %f4, %f1;
setp.lt.f32 %p2, %f4, 0f3ED1EB85;
selp.f32 %f5, 0f00000000, %f3, %p2;
neg.f32 %f6, %f5;
mov.f32 %f7, 0f3F317200;
fma.rn.f32 %f8, %f6, %f7, %f1;
mov.f32 %f9, 0f35BFBE8E;
fma.rn.f32 %f10, %f6, %f9, %f8;
setp.eq.f32 %p3, %f5, 0f43000000;
add.f32 %f11, %f5, 0fBF800000;
selp.f32 %f12, %f11, %f5, %p3;
mov.f32 %f13, 0f3C095663;
mov.f32 %f14, 0f3AB5EBE6;
fma.rn.f32 %f15, %f14, %f10, %f13;
mov.f32 %f16, 0f3D2AABE3;
fma.rn.f32 %f17, %f15, %f10, %f16;
mov.f32 %f18, 0f3E2AA9F6;
fma.rn.f32 %f19, %f17, %f10, %f18;
mov.f32 %f20, 0f3EFFFFFE;
fma.rn.f32 %f21, %f19, %f10, %f20;
mul.f32 %f22, %f10, %f21;
fma.rn.f32 %f23, %f22, %f10, %f10;
ex2.approx.f32 %f24, %f12;
add.f32 %f25, %f24, 0fBF800000;
fma.rn.f32 %f26, %f23, %f24, %f25;
add.f32 %f27, %f26, %f26;
selp.f32 %f28, %f27, %f26, %p3;
setp.gt.f32 %p4, %f12, 0f43000000;
selp.f32 %f29, 0f7F800000, %f28, %p4;
setp.lt.f32 %p5, %f12, 0fC1C80000;
selp.f32 %f30, 0fBF800000, %f29, %p5;
setp.eq.f32 %p6, %f1, 0f00000000;
add.f32 %f31, %f1, %f1;
selp.f32 %f32, %f31, %f30, %p6;
cvta.to.global.u64 %rd8, %rd2;
add.s64 %rd9, %rd8, %rd6;
st.global.f32 [%rd9], %f32;
BB45_2:
ret;
}
// .globl vec_fabsf
.visible .entry vec_fabsf(
.param .u64 vec_fabsf_param_0,
.param .u64 vec_fabsf_param_1,
.param .u64 vec_fabsf_param_2
)
{
.reg .pred %p<2>;
.reg .f32 %f<3>;
.reg .b32 %r<5>;
.reg .b64 %rd<10>;
ld.param.u64 %rd4, [vec_fabsf_param_0];
ld.param.u64 %rd2, [vec_fabsf_param_1];
ld.param.u64 %rd3, [vec_fabsf_param_2];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB46_2;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f1, [%rd7];
abs.f32 %f2, %f1;
cvta.to.global.u64 %rd8, %rd2;
add.s64 %rd9, %rd8, %rd6;
st.global.f32 [%rd9], %f2;
BB46_2:
ret;
}
// .globl vec_floorf
.visible .entry vec_floorf(
.param .u64 vec_floorf_param_0,
.param .u64 vec_floorf_param_1,
.param .u64 vec_floorf_param_2
)
{
.reg .pred %p<2>;
.reg .f32 %f<3>;
.reg .b32 %r<5>;
.reg .b64 %rd<10>;
ld.param.u64 %rd4, [vec_floorf_param_0];
ld.param.u64 %rd2, [vec_floorf_param_1];
ld.param.u64 %rd3, [vec_floorf_param_2];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB47_2;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f1, [%rd7];
cvt.rmi.f32.f32 %f2, %f1;
cvta.to.global.u64 %rd8, %rd2;
add.s64 %rd9, %rd8, %rd6;
st.global.f32 [%rd9], %f2;
BB47_2:
ret;
}
// .globl vec_j0f
.visible .entry vec_j0f(
.param .u64 vec_j0f_param_0,
.param .u64 vec_j0f_param_1,
.param .u64 vec_j0f_param_2
)
{
.local .align 4 .b8 __local_depot48[28];
.reg .b64 %SP;
.reg .b64 %SPL;
.reg .pred %p<25>;
.reg .f32 %f<129>;
.reg .b32 %r<186>;
.reg .b64 %rd<36>;
mov.u64 %rd35, __local_depot48;
cvta.local.u64 %SP, %rd35;
ld.param.u64 %rd15, [vec_j0f_param_0];
ld.param.u64 %rd13, [vec_j0f_param_1];
ld.param.u64 %rd14, [vec_j0f_param_2];
add.u64 %rd16, %SP, 0;
cvta.to.local.u64 %rd1, %rd16;
mov.u32 %r73, %ntid.x;
mov.u32 %r74, %ctaid.x;
mov.u32 %r75, %tid.x;
mad.lo.s32 %r1, %r73, %r74, %r75;
cvt.s64.s32 %rd17, %r1;
setp.ge.u64 %p1, %rd17, %rd15;
@%p1 bra BB48_38;
cvta.to.global.u64 %rd18, %rd14;
mul.wide.s32 %rd19, %r1, 4;
add.s64 %rd20, %rd18, %rd19;
ld.global.f32 %f28, [%rd20];
abs.f32 %f1, %f28;
setp.gtu.f32 %p2, %f1, 0f41000000;
@%p2 bra BB48_3;
bra.uni BB48_2;
BB48_3:
abs.f32 %f65, %f1;
mov.f32 %f128, 0f00000000;
setp.eq.f32 %p3, %f65, 0f7F800000;
@%p3 bra BB48_37;
// inline asm
rcp.approx.ftz.f32 %f66,%f1;
// inline asm
mul.f32 %f68, %f66, %f66;
mov.f32 %f69, 0fBF03B7C2;
mov.f32 %f70, 0f4056FE93;
fma.rn.f32 %f71, %f70, %f68, %f69;
mov.f32 %f72, 0f3DD3B3F3;
fma.rn.f32 %f73, %f71, %f68, %f72;
mov.f32 %f74, 0fBD7FFFB6;
fma.rn.f32 %f75, %f73, %f68, %f74;
mov.f32 %f76, 0f3F800000;
fma.rn.f32 %f77, %f75, %f68, %f76;
mov.f32 %f78, 0fBE52412D;
mov.f32 %f79, 0f3F91E009;
fma.rn.f32 %f80, %f79, %f68, %f78;
mov.f32 %f81, 0f3D854ED1;
fma.rn.f32 %f82, %f80, %f68, %f81;
mov.f32 %f83, 0fBDFFFFFF;
fma.rn.f32 %f84, %f82, %f68, %f83;
fma.rn.f32 %f3, %f84, %f66, %f1;
rsqrt.approx.f32 %f85, %f1;
mul.f32 %f86, %f85, 0f3F4C422A;
mul.f32 %f4, %f77, %f86;
mul.f32 %f87, %f3, 0f3F22F983;
cvt.rni.s32.f32 %r175, %f87;
cvt.rn.f32.s32 %f88, %r175;
neg.f32 %f89, %f88;
mov.f32 %f90, 0f3FC90FDA;
fma.rn.f32 %f91, %f89, %f90, %f3;
mov.f32 %f92, 0f33A22168;
fma.rn.f32 %f93, %f89, %f92, %f91;
mov.f32 %f94, 0f27C234C5;
fma.rn.f32 %f122, %f89, %f94, %f93;
abs.f32 %f95, %f3;
setp.leu.f32 %p4, %f95, 0f47CE4780;
@%p4 bra BB48_14;
add.s64 %rd2, %rd1, 24;
mov.b32 %r3, %f3;
bfe.u32 %r78, %r3, 23, 8;
add.s32 %r79, %r78, -128;
shl.b32 %r80, %r3, 8;
or.b32 %r4, %r80, -2147483648;
shr.u32 %r5, %r79, 5;
mov.u32 %r167, 0;
mov.u64 %rd30, __cudart_i2opi_f;
mov.u32 %r166, -6;
mov.u64 %rd34, %rd1;
BB48_6:
.pragma "nounroll";
mov.u64 %rd4, %rd34;
ld.const.u32 %r83, [%rd30];
// inline asm
{
mad.lo.cc.u32 %r81, %r83, %r4, %r167;
madc.hi.u32 %r167, %r83, %r4, 0;
}
// inline asm
st.local.u32 [%rd4], %r81;
add.s64 %rd5, %rd4, 4;
add.s64 %rd30, %rd30, 4;
add.s32 %r166, %r166, 1;
setp.ne.s32 %p5, %r166, 0;
mov.u64 %rd34, %rd5;
@%p5 bra BB48_6;
and.b32 %r10, %r3, -2147483648;
st.local.u32 [%rd2], %r167;
bfe.u32 %r11, %r3, 23, 5;
mov.u32 %r86, 6;
sub.s32 %r87, %r86, %r5;
mul.wide.s32 %rd22, %r87, 4;
add.s64 %rd7, %rd1, %rd22;
ld.local.u32 %r168, [%rd7];
ld.local.u32 %r169, [%rd7+-4];
setp.eq.s32 %p6, %r11, 0;
@%p6 bra BB48_9;
mov.u32 %r88, 32;
sub.s32 %r89, %r88, %r11;
shr.u32 %r90, %r169, %r89;
shl.b32 %r91, %r168, %r11;
add.s32 %r168, %r90, %r91;
ld.local.u32 %r92, [%rd7+-8];
shr.u32 %r93, %r92, %r89;
shl.b32 %r94, %r169, %r11;
add.s32 %r169, %r93, %r94;
BB48_9:
shr.u32 %r95, %r169, 30;
shl.b32 %r96, %r168, 2;
add.s32 %r170, %r95, %r96;
shl.b32 %r19, %r169, 2;
shr.u32 %r97, %r170, 31;
shr.u32 %r98, %r168, 30;
add.s32 %r20, %r97, %r98;
setp.eq.s32 %p7, %r97, 0;
mov.u32 %r171, %r10;
mov.u32 %r172, %r19;
@%p7 bra BB48_11;
not.b32 %r99, %r170;
neg.s32 %r21, %r19;
setp.eq.s32 %p8, %r19, 0;
selp.u32 %r100, 1, 0, %p8;
add.s32 %r170, %r100, %r99;
xor.b32 %r23, %r10, -2147483648;
mov.u32 %r171, %r23;
mov.u32 %r172, %r21;
BB48_11:
mov.u32 %r25, %r171;
neg.s32 %r101, %r20;
setp.eq.s32 %p9, %r10, 0;
selp.b32 %r175, %r20, %r101, %p9;
clz.b32 %r174, %r170;
setp.eq.s32 %p10, %r174, 0;
shl.b32 %r102, %r170, %r174;
mov.u32 %r103, 32;
sub.s32 %r104, %r103, %r174;
shr.u32 %r105, %r172, %r104;
add.s32 %r106, %r105, %r102;
selp.b32 %r29, %r170, %r106, %p10;
mov.u32 %r107, -921707870;
mul.hi.u32 %r173, %r29, %r107;
setp.lt.s32 %p11, %r173, 1;
@%p11 bra BB48_13;
mul.lo.s32 %r108, %r29, -921707870;
shr.u32 %r109, %r108, 31;
shl.b32 %r110, %r173, 1;
add.s32 %r173, %r109, %r110;
add.s32 %r174, %r174, 1;
BB48_13:
mov.u32 %r111, 126;
sub.s32 %r112, %r111, %r174;
shl.b32 %r113, %r112, 23;
add.s32 %r114, %r173, 1;
shr.u32 %r115, %r114, 7;
add.s32 %r116, %r115, 1;
shr.u32 %r117, %r116, 1;
add.s32 %r118, %r117, %r113;
or.b32 %r119, %r118, %r25;
mov.b32 %f122, %r119;
BB48_14:
and.b32 %r120, %r175, 3;
cvt.rn.f32.s32 %f96, %r120;
add.f32 %f97, %f122, 0fBF490FDB;
fma.rn.f32 %f123, %f96, 0f3FC90FDB, %f97;
abs.f32 %f98, %f123;
setp.neu.f32 %p12, %f98, 0f7F800000;
@%p12 bra BB48_16;
mov.f32 %f99, 0f00000000;
mul.rn.f32 %f123, %f123, %f99;
BB48_16:
mul.f32 %f100, %f123, 0f3F22F983;
cvt.rni.s32.f32 %r185, %f100;
cvt.rn.f32.s32 %f101, %r185;
neg.f32 %f102, %f101;
fma.rn.f32 %f104, %f102, %f90, %f123;
fma.rn.f32 %f106, %f102, %f92, %f104;
fma.rn.f32 %f124, %f102, %f94, %f106;
abs.f32 %f108, %f123;
setp.leu.f32 %p13, %f108, 0f47CE4780;
@%p13 bra BB48_26;
mov.b32 %r37, %f123;
shr.u32 %r38, %r37, 23;
bfe.u32 %r123, %r37, 23, 8;
add.s32 %r124, %r123, -128;
shl.b32 %r125, %r37, 8;
or.b32 %r39, %r125, -2147483648;
shr.u32 %r40, %r124, 5;
mov.u32 %r177, 0;
mov.u64 %rd31, __cudart_i2opi_f;
mov.u32 %r176, -6;
mov.u64 %rd33, %rd1;
BB48_18:
.pragma "nounroll";
ld.const.u32 %r128, [%rd31];
// inline asm
{
mad.lo.cc.u32 %r126, %r128, %r39, %r177;
madc.hi.u32 %r177, %r128, %r39, 0;
}
// inline asm
st.local.u32 [%rd33], %r126;
add.s64 %rd33, %rd33, 4;
add.s64 %rd31, %rd31, 4;
add.s32 %r176, %r176, 1;
setp.ne.s32 %p14, %r176, 0;
@%p14 bra BB48_18;
and.b32 %r45, %r37, -2147483648;
cvta.to.local.u64 %rd25, %rd16;
st.local.u32 [%rd25+24], %r177;
mov.u32 %r131, 6;
sub.s32 %r132, %r131, %r40;
mul.wide.s32 %rd26, %r132, 4;
add.s64 %rd12, %rd25, %rd26;
ld.local.u32 %r178, [%rd12];
ld.local.u32 %r179, [%rd12+-4];
and.b32 %r48, %r38, 31;
setp.eq.s32 %p15, %r48, 0;
@%p15 bra BB48_21;
mov.u32 %r133, 32;
sub.s32 %r134, %r133, %r48;
shr.u32 %r135, %r179, %r134;
shl.b32 %r136, %r178, %r48;
add.s32 %r178, %r135, %r136;
ld.local.u32 %r137, [%rd12+-8];
shr.u32 %r138, %r137, %r134;
shl.b32 %r139, %r179, %r48;
add.s32 %r179, %r138, %r139;
BB48_21:
shr.u32 %r140, %r179, 30;
shl.b32 %r141, %r178, 2;
add.s32 %r180, %r140, %r141;
shl.b32 %r54, %r179, 2;
shr.u32 %r142, %r180, 31;
shr.u32 %r143, %r178, 30;
add.s32 %r55, %r142, %r143;
setp.eq.s32 %p16, %r142, 0;
mov.u32 %r181, %r45;
mov.u32 %r182, %r54;
@%p16 bra BB48_23;
not.b32 %r144, %r180;
neg.s32 %r56, %r54;
setp.eq.s32 %p17, %r54, 0;
selp.u32 %r145, 1, 0, %p17;
add.s32 %r180, %r145, %r144;
xor.b32 %r58, %r45, -2147483648;
mov.u32 %r181, %r58;
mov.u32 %r182, %r56;
BB48_23:
mov.u32 %r60, %r181;
neg.s32 %r146, %r55;
setp.eq.s32 %p18, %r45, 0;
selp.b32 %r185, %r55, %r146, %p18;
clz.b32 %r184, %r180;
setp.eq.s32 %p19, %r184, 0;
shl.b32 %r147, %r180, %r184;
mov.u32 %r148, 32;
sub.s32 %r149, %r148, %r184;
shr.u32 %r150, %r182, %r149;
add.s32 %r151, %r150, %r147;
selp.b32 %r64, %r180, %r151, %p19;
mov.u32 %r152, -921707870;
mul.hi.u32 %r183, %r64, %r152;
setp.lt.s32 %p20, %r183, 1;
@%p20 bra BB48_25;
mul.lo.s32 %r153, %r64, -921707870;
shr.u32 %r154, %r153, 31;
shl.b32 %r155, %r183, 1;
add.s32 %r183, %r154, %r155;
add.s32 %r184, %r184, 1;
BB48_25:
mov.u32 %r156, 126;
sub.s32 %r157, %r156, %r184;
shl.b32 %r158, %r157, 23;
add.s32 %r159, %r183, 1;
shr.u32 %r160, %r159, 7;
add.s32 %r161, %r160, 1;
shr.u32 %r162, %r161, 1;
add.s32 %r163, %r162, %r158;
or.b32 %r164, %r163, %r60;
mov.b32 %f124, %r164;
BB48_26:
mul.rn.f32 %f14, %f124, %f124;
add.s32 %r71, %r185, 1;
and.b32 %r72, %r71, 1;
setp.eq.s32 %p21, %r72, 0;
@%p21 bra BB48_28;
mov.f32 %f109, 0fBAB6061A;
mov.f32 %f110, 0f37CCF5CE;
fma.rn.f32 %f125, %f110, %f14, %f109;
bra.uni BB48_29;
BB48_2:
add.f32 %f29, %f1, 0fC019E8A9;
add.f32 %f30, %f29, 0fB3E971B3;
mov.f32 %f31, 0fA9ACA9B3;
mov.f32 %f32, 0fA6B3B8E7;
fma.rn.f32 %f33, %f32, %f30, %f31;
mov.f32 %f34, 0f2C3F0E18;
fma.rn.f32 %f35, %f33, %f30, %f34;
mov.f32 %f36, 0fACD41781;
fma.rn.f32 %f37, %f35, %f30, %f36;
mov.f32 %f38, 0fAFE90F38;
fma.rn.f32 %f39, %f37, %f30, %f38;
mov.f32 %f40, 0f3020305B;
fma.rn.f32 %f41, %f39, %f30, %f40;
mov.f32 %f42, 0f33797143;
fma.rn.f32 %f43, %f41, %f30, %f42;
mov.f32 %f44, 0f30F76F85;
fma.rn.f32 %f45, %f43, %f30, %f44;
mov.f32 %f46, 0fB6B6DFC6;
fma.rn.f32 %f47, %f45, %f30, %f46;
mov.f32 %f48, 0fB6F665C9;
fma.rn.f32 %f49, %f47, %f30, %f48;
mov.f32 %f50, 0f399E2DEB;
fma.rn.f32 %f51, %f49, %f30, %f50;
mov.f32 %f52, 0f3A4AE334;
fma.rn.f32 %f53, %f51, %f30, %f52;
mov.f32 %f54, 0fBBEEAA1B;
fma.rn.f32 %f55, %f53, %f30, %f54;
mov.f32 %f56, 0fBCDA7747;
fma.rn.f32 %f57, %f55, %f30, %f56;
mul.f32 %f58, %f30, %f57;
add.f32 %f59, %f1, 0fC0B0A47B;
add.f32 %f60, %f59, 0f339A7A37;
mul.f32 %f61, %f60, %f58;
add.f32 %f62, %f1, 0fC10A75AB;
add.f32 %f63, %f62, 0fB4CCCDED;
mul.f32 %f128, %f63, %f61;
bra.uni BB48_37;
BB48_28:
mov.f32 %f111, 0f3C08839E;
mov.f32 %f112, 0fB94CA1F9;
fma.rn.f32 %f125, %f112, %f14, %f111;
BB48_29:
@%p21 bra BB48_31;
mov.f32 %f113, 0f3D2AAAA5;
fma.rn.f32 %f114, %f125, %f14, %f113;
mov.f32 %f115, 0fBF000000;
fma.rn.f32 %f126, %f114, %f14, %f115;
bra.uni BB48_32;
BB48_31:
mov.f32 %f116, 0fBE2AAAA3;
fma.rn.f32 %f117, %f125, %f14, %f116;
mov.f32 %f118, 0f00000000;
fma.rn.f32 %f126, %f117, %f14, %f118;
BB48_32:
fma.rn.f32 %f127, %f126, %f124, %f124;
@%p21 bra BB48_34;
fma.rn.f32 %f127, %f126, %f14, %f76;
BB48_34:
and.b32 %r165, %r71, 2;
setp.eq.s32 %p24, %r165, 0;
@%p24 bra BB48_36;
mov.f32 %f120, 0f00000000;
mov.f32 %f121, 0fBF800000;
fma.rn.f32 %f127, %f127, %f121, %f120;
BB48_36:
mul.f32 %f128, %f4, %f127;
BB48_37:
cvta.to.global.u64 %rd27, %rd13;
add.s64 %rd29, %rd27, %rd19;
st.global.f32 [%rd29], %f128;
BB48_38:
ret;
}
// .globl vec_j1f
.visible .entry vec_j1f(
.param .u64 vec_j1f_param_0,
.param .u64 vec_j1f_param_1,
.param .u64 vec_j1f_param_2
)
{
.local .align 4 .b8 __local_depot49[28];
.reg .b64 %SP;
.reg .b64 %SPL;
.reg .pred %p<27>;
.reg .f32 %f<129>;
.reg .b32 %r<195>;
.reg .b64 %rd<38>;
mov.u64 %rd37, __local_depot49;
cvta.local.u64 %SP, %rd37;
ld.param.u64 %rd15, [vec_j1f_param_0];
ld.param.u64 %rd13, [vec_j1f_param_1];
ld.param.u64 %rd14, [vec_j1f_param_2];
mov.u32 %r73, %ntid.x;
mov.u32 %r74, %ctaid.x;
mov.u32 %r75, %tid.x;
mad.lo.s32 %r1, %r73, %r74, %r75;
cvt.s64.s32 %rd16, %r1;
setp.ge.u64 %p1, %rd16, %rd15;
@%p1 bra BB49_38;
cvta.to.global.u64 %rd17, %rd14;
mul.wide.s32 %rd18, %r1, 4;
add.s64 %rd19, %rd17, %rd18;
ld.global.f32 %f1, [%rd19];
abs.f32 %f2, %f1;
setp.gtu.f32 %p2, %f2, 0f40FB3333;
@%p2 bra BB49_3;
bra.uni BB49_2;
BB49_3:
abs.f32 %f63, %f2;
mov.f32 %f128, 0f00000000;
setp.eq.f32 %p3, %f63, 0f7F800000;
@%p3 bra BB49_37;
// inline asm
rcp.approx.ftz.f32 %f64,%f2;
// inline asm
mul.f32 %f66, %f64, %f64;
mov.f32 %f67, 0f3F3FF7E9;
mov.f32 %f68, 0fC082CB37;
fma.rn.f32 %f69, %f68, %f66, %f67;
mov.f32 %f70, 0fBE458BAE;
fma.rn.f32 %f71, %f69, %f66, %f70;
mov.f32 %f72, 0f3E3FFF8B;
fma.rn.f32 %f73, %f71, %f66, %f72;
mov.f32 %f74, 0f3F800000;
fma.rn.f32 %f4, %f73, %f66, %f74;
mov.f32 %f75, 0f3EB914AD;
mov.f32 %f76, 0fBFCA3BA2;
fma.rn.f32 %f77, %f76, %f66, %f75;
mov.f32 %f78, 0fBE27F2EC;
fma.rn.f32 %f79, %f77, %f66, %f78;
mov.f32 %f80, 0f3EBFFFFD;
fma.rn.f32 %f81, %f79, %f66, %f80;
fma.rn.f32 %f5, %f81, %f64, %f2;
rsqrt.approx.f32 %f6, %f2;
mul.f32 %f82, %f5, 0f3F22F983;
cvt.rni.s32.f32 %r184, %f82;
cvt.rn.f32.s32 %f83, %r184;
neg.f32 %f84, %f83;
mov.f32 %f85, 0f3FC90FDA;
fma.rn.f32 %f86, %f84, %f85, %f5;
mov.f32 %f87, 0f33A22168;
fma.rn.f32 %f88, %f84, %f87, %f86;
mov.f32 %f89, 0f27C234C5;
fma.rn.f32 %f122, %f84, %f89, %f88;
abs.f32 %f90, %f5;
setp.leu.f32 %p4, %f90, 0f47CE4780;
@%p4 bra BB49_14;
add.u64 %rd21, %SP, 0;
cvta.to.local.u64 %rd34, %rd21;
mov.b32 %r3, %f5;
bfe.u32 %r78, %r3, 23, 8;
add.s32 %r79, %r78, -128;
shl.b32 %r80, %r3, 8;
or.b32 %r4, %r80, -2147483648;
shr.u32 %r5, %r79, 5;
mov.u32 %r176, 0;
mov.u64 %rd33, __cudart_i2opi_f;
mov.u32 %r175, -6;
BB49_6:
.pragma "nounroll";
ld.const.u32 %r83, [%rd33];
// inline asm
{
mad.lo.cc.u32 %r81, %r83, %r4, %r176;
madc.hi.u32 %r176, %r83, %r4, 0;
}
// inline asm
st.local.u32 [%rd34], %r81;
add.s64 %rd34, %rd34, 4;
add.s64 %rd33, %rd33, 4;
add.s32 %r175, %r175, 1;
setp.ne.s32 %p5, %r175, 0;
@%p5 bra BB49_6;
and.b32 %r10, %r3, -2147483648;
cvta.to.local.u64 %rd23, %rd21;
st.local.u32 [%rd23+24], %r176;
bfe.u32 %r11, %r3, 23, 5;
mov.u32 %r86, 6;
sub.s32 %r87, %r86, %r5;
mul.wide.s32 %rd24, %r87, 4;
add.s64 %rd6, %rd23, %rd24;
ld.local.u32 %r177, [%rd6];
ld.local.u32 %r178, [%rd6+-4];
setp.eq.s32 %p6, %r11, 0;
@%p6 bra BB49_9;
mov.u32 %r88, 32;
sub.s32 %r89, %r88, %r11;
shr.u32 %r90, %r178, %r89;
shl.b32 %r91, %r177, %r11;
add.s32 %r177, %r90, %r91;
ld.local.u32 %r92, [%rd6+-8];
shr.u32 %r93, %r92, %r89;
shl.b32 %r94, %r178, %r11;
add.s32 %r178, %r93, %r94;
BB49_9:
shr.u32 %r95, %r178, 30;
shl.b32 %r96, %r177, 2;
add.s32 %r179, %r95, %r96;
shl.b32 %r19, %r178, 2;
shr.u32 %r97, %r179, 31;
shr.u32 %r98, %r177, 30;
add.s32 %r20, %r97, %r98;
setp.eq.s32 %p7, %r97, 0;
mov.u32 %r180, %r10;
mov.u32 %r181, %r19;
@%p7 bra BB49_11;
not.b32 %r99, %r179;
neg.s32 %r21, %r19;
setp.eq.s32 %p8, %r19, 0;
selp.u32 %r100, 1, 0, %p8;
add.s32 %r179, %r100, %r99;
xor.b32 %r23, %r10, -2147483648;
mov.u32 %r180, %r23;
mov.u32 %r181, %r21;
BB49_11:
mov.u32 %r25, %r180;
neg.s32 %r101, %r20;
setp.eq.s32 %p9, %r10, 0;
selp.b32 %r184, %r20, %r101, %p9;
clz.b32 %r183, %r179;
setp.eq.s32 %p10, %r183, 0;
shl.b32 %r102, %r179, %r183;
mov.u32 %r103, 32;
sub.s32 %r104, %r103, %r183;
shr.u32 %r105, %r181, %r104;
add.s32 %r106, %r105, %r102;
selp.b32 %r29, %r179, %r106, %p10;
mov.u32 %r107, -921707870;
mul.hi.u32 %r182, %r29, %r107;
setp.lt.s32 %p11, %r182, 1;
@%p11 bra BB49_13;
mul.lo.s32 %r108, %r29, -921707870;
shr.u32 %r109, %r108, 31;
shl.b32 %r110, %r182, 1;
add.s32 %r182, %r109, %r110;
add.s32 %r183, %r183, 1;
BB49_13:
mov.u32 %r111, 126;
sub.s32 %r112, %r111, %r183;
shl.b32 %r113, %r112, 23;
add.s32 %r114, %r182, 1;
shr.u32 %r115, %r114, 7;
add.s32 %r116, %r115, 1;
shr.u32 %r117, %r116, 1;
add.s32 %r118, %r117, %r113;
or.b32 %r119, %r118, %r25;
mov.b32 %f122, %r119;
BB49_14:
mul.f32 %f91, %f6, 0f3F4C422A;
mul.f32 %f10, %f4, %f91;
and.b32 %r120, %r184, 3;
cvt.rn.f32.s32 %f92, %r120;
add.f32 %f93, %f122, 0fC016CBE4;
fma.rn.f32 %f123, %f92, 0f3FC90FDB, %f93;
abs.f32 %f94, %f123;
setp.neu.f32 %p12, %f94, 0f7F800000;
@%p12 bra BB49_16;
mov.f32 %f95, 0f00000000;
mul.rn.f32 %f123, %f123, %f95;
BB49_16:
mul.f32 %f96, %f123, 0f3F22F983;
cvt.rni.s32.f32 %r194, %f96;
cvt.rn.f32.s32 %f97, %r194;
neg.f32 %f98, %f97;
fma.rn.f32 %f100, %f98, %f85, %f123;
fma.rn.f32 %f102, %f98, %f87, %f100;
fma.rn.f32 %f124, %f98, %f89, %f102;
abs.f32 %f104, %f123;
setp.leu.f32 %p13, %f104, 0f47CE4780;
@%p13 bra BB49_26;
mov.b32 %r37, %f123;
shr.u32 %r38, %r37, 23;
bfe.u32 %r123, %r37, 23, 8;
add.s32 %r124, %r123, -128;
shl.b32 %r125, %r37, 8;
or.b32 %r39, %r125, -2147483648;
shr.u32 %r40, %r124, 5;
add.u64 %rd26, %SP, 0;
cvta.to.local.u64 %rd36, %rd26;
mov.u32 %r186, 0;
mov.u64 %rd35, __cudart_i2opi_f;
mov.u32 %r185, -6;
BB49_18:
.pragma "nounroll";
ld.const.u32 %r128, [%rd35];
// inline asm
{
mad.lo.cc.u32 %r126, %r128, %r39, %r186;
madc.hi.u32 %r186, %r128, %r39, 0;
}
// inline asm
st.local.u32 [%rd36], %r126;
add.s64 %rd36, %rd36, 4;
add.s64 %rd35, %rd35, 4;
add.s32 %r185, %r185, 1;
setp.ne.s32 %p14, %r185, 0;
@%p14 bra BB49_18;
and.b32 %r45, %r37, -2147483648;
cvta.to.local.u64 %rd28, %rd26;
st.local.u32 [%rd28+24], %r186;
mov.u32 %r131, 6;
sub.s32 %r132, %r131, %r40;
mul.wide.s32 %rd29, %r132, 4;
add.s64 %rd12, %rd28, %rd29;
ld.local.u32 %r187, [%rd12];
ld.local.u32 %r188, [%rd12+-4];
and.b32 %r48, %r38, 31;
setp.eq.s32 %p15, %r48, 0;
@%p15 bra BB49_21;
mov.u32 %r133, 32;
sub.s32 %r134, %r133, %r48;
shr.u32 %r135, %r188, %r134;
shl.b32 %r136, %r187, %r48;
add.s32 %r187, %r135, %r136;
ld.local.u32 %r137, [%rd12+-8];
shr.u32 %r138, %r137, %r134;
shl.b32 %r139, %r188, %r48;
add.s32 %r188, %r138, %r139;
BB49_21:
shr.u32 %r140, %r188, 30;
shl.b32 %r141, %r187, 2;
add.s32 %r189, %r140, %r141;
shl.b32 %r54, %r188, 2;
shr.u32 %r142, %r189, 31;
shr.u32 %r143, %r187, 30;
add.s32 %r55, %r142, %r143;
setp.eq.s32 %p16, %r142, 0;
mov.u32 %r190, %r45;
mov.u32 %r191, %r54;
@%p16 bra BB49_23;
not.b32 %r144, %r189;
neg.s32 %r56, %r54;
setp.eq.s32 %p17, %r54, 0;
selp.u32 %r145, 1, 0, %p17;
add.s32 %r189, %r145, %r144;
xor.b32 %r58, %r45, -2147483648;
mov.u32 %r190, %r58;
mov.u32 %r191, %r56;
BB49_23:
mov.u32 %r60, %r190;
neg.s32 %r146, %r55;
setp.eq.s32 %p18, %r45, 0;
selp.b32 %r194, %r55, %r146, %p18;
clz.b32 %r193, %r189;
setp.eq.s32 %p19, %r193, 0;
shl.b32 %r147, %r189, %r193;
mov.u32 %r148, 32;
sub.s32 %r149, %r148, %r193;
shr.u32 %r150, %r191, %r149;
add.s32 %r151, %r150, %r147;
selp.b32 %r64, %r189, %r151, %p19;
mov.u32 %r152, -921707870;
mul.hi.u32 %r192, %r64, %r152;
setp.lt.s32 %p20, %r192, 1;
@%p20 bra BB49_25;
mul.lo.s32 %r153, %r64, -921707870;
shr.u32 %r154, %r153, 31;
shl.b32 %r155, %r192, 1;
add.s32 %r192, %r154, %r155;
add.s32 %r193, %r193, 1;
BB49_25:
mov.u32 %r156, 126;
sub.s32 %r157, %r156, %r193;
shl.b32 %r158, %r157, 23;
add.s32 %r159, %r192, 1;
shr.u32 %r160, %r159, 7;
add.s32 %r161, %r160, 1;
shr.u32 %r162, %r161, 1;
add.s32 %r163, %r162, %r158;
or.b32 %r164, %r163, %r60;
mov.b32 %f124, %r164;
BB49_26:
mul.rn.f32 %f17, %f124, %f124;
add.s32 %r71, %r194, 1;
and.b32 %r72, %r71, 1;
setp.eq.s32 %p21, %r72, 0;
@%p21 bra BB49_28;
mov.f32 %f105, 0fBAB6061A;
mov.f32 %f106, 0f37CCF5CE;
fma.rn.f32 %f125, %f106, %f17, %f105;
bra.uni BB49_29;
BB49_2:
add.f32 %f31, %f2, 0fC0753AAC;
add.f32 %f32, %f31, 0f33A5090F;
mov.f32 %f33, 0f2B81BF42;
mov.f32 %f34, 0f29AF3463;
fma.rn.f32 %f35, %f34, %f32, %f33;
mov.f32 %f36, 0fADE21EC1;
fma.rn.f32 %f37, %f35, %f32, %f36;
mov.f32 %f38, 0fAF5DDEFF;
fma.rn.f32 %f39, %f37, %f32, %f38;
mov.f32 %f40, 0f319B0C9D;
fma.rn.f32 %f41, %f39, %f32, %f40;
mov.f32 %f42, 0f32E81173;
fma.rn.f32 %f43, %f41, %f32, %f42;
mov.f32 %f44, 0fB50F8DC8;
fma.rn.f32 %f45, %f43, %f32, %f44;
mov.f32 %f46, 0fB61E653D;
fma.rn.f32 %f47, %f45, %f32, %f46;
mov.f32 %f48, 0f382CD9C5;
fma.rn.f32 %f49, %f47, %f32, %f48;
mov.f32 %f50, 0f38F9EB10;
fma.rn.f32 %f51, %f49, %f32, %f50;
mov.f32 %f52, 0fBAECEB9C;
fma.rn.f32 %f53, %f51, %f32, %f52;
mov.f32 %f54, 0fBB276FFD;
fma.rn.f32 %f55, %f53, %f32, %f54;
mov.f32 %f56, 0f3D073993;
fma.rn.f32 %f57, %f55, %f32, %f56;
add.f32 %f58, %f2, 0fC0E07FB0;
add.f32 %f59, %f58, 0f3444B8DB;
mul.f32 %f60, %f59, %f57;
mul.f32 %f61, %f32, %f60;
mul.f32 %f128, %f2, %f61;
bra.uni BB49_37;
BB49_28:
mov.f32 %f107, 0f3C08839E;
mov.f32 %f108, 0fB94CA1F9;
fma.rn.f32 %f125, %f108, %f17, %f107;
BB49_29:
@%p21 bra BB49_31;
mov.f32 %f109, 0f3D2AAAA5;
fma.rn.f32 %f110, %f125, %f17, %f109;
mov.f32 %f111, 0fBF000000;
fma.rn.f32 %f126, %f110, %f17, %f111;
bra.uni BB49_32;
BB49_31:
mov.f32 %f112, 0fBE2AAAA3;
fma.rn.f32 %f113, %f125, %f17, %f112;
mov.f32 %f114, 0f00000000;
fma.rn.f32 %f126, %f113, %f17, %f114;
BB49_32:
fma.rn.f32 %f127, %f126, %f124, %f124;
@%p21 bra BB49_34;
fma.rn.f32 %f127, %f126, %f17, %f74;
BB49_34:
and.b32 %r165, %r71, 2;
setp.eq.s32 %p24, %r165, 0;
@%p24 bra BB49_36;
mov.f32 %f116, 0f00000000;
mov.f32 %f117, 0fBF800000;
fma.rn.f32 %f127, %f127, %f117, %f116;
BB49_36:
mul.f32 %f128, %f10, %f127;
BB49_37:
neg.f32 %f118, %f128;
setp.lt.f32 %p25, %f1, 0f00000000;
selp.f32 %f119, %f118, %f128, %p25;
mov.b32 %r166, %f1;
and.b32 %r167, %r166, -2147483648;
mov.b32 %r168, %f119;
and.b32 %r169, %r168, 2147483647;
or.b32 %r170, %r169, %r167;
mov.b32 %f120, %r170;
setp.lt.f32 %p26, %f2, 0f0DA24260;
selp.f32 %f121, %f120, %f119, %p26;
cvta.to.global.u64 %rd30, %rd13;
add.s64 %rd32, %rd30, %rd18;
st.global.f32 [%rd32], %f121;
BB49_38:
ret;
}
// .globl vec_lgammaf
.visible .entry vec_lgammaf(
.param .u64 vec_lgammaf_param_0,
.param .u64 vec_lgammaf_param_1,
.param .u64 vec_lgammaf_param_2
)
{
.reg .pred %p<34>;
.reg .f32 %f<271>;
.reg .b32 %r<24>;
.reg .b64 %rd<11>;
ld.param.u64 %rd4, [vec_lgammaf_param_0];
ld.param.u64 %rd2, [vec_lgammaf_param_1];
ld.param.u64 %rd3, [vec_lgammaf_param_2];
mov.u32 %r3, %tid.x;
mov.u32 %r4, %ntid.x;
mov.u32 %r5, %ctaid.x;
mad.lo.s32 %r6, %r4, %r5, %r3;
cvt.s64.s32 %rd1, %r6;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB50_38;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f1, [%rd7];
abs.f32 %f2, %f1;
setp.ltu.f32 %p2, %f2, 0f40400000;
@%p2 bra BB50_8;
bra.uni BB50_2;
BB50_8:
setp.ltu.f32 %p10, %f2, 0f3FC00000;
@%p10 bra BB50_10;
bra.uni BB50_9;
BB50_10:
setp.ltu.f32 %p11, %f2, 0f3F333333;
@%p11 bra BB50_12;
bra.uni BB50_11;
BB50_12:
mov.f32 %f143, 0fBBB34878;
mov.f32 %f144, 0f3B6B1C86;
fma.rn.f32 %f145, %f144, %f2, %f143;
mov.f32 %f146, 0fBD36CAEF;
fma.rn.f32 %f147, %f145, %f2, %f146;
mov.f32 %f148, 0f3E2B5555;
fma.rn.f32 %f149, %f147, %f2, %f148;
mov.f32 %f150, 0fBD2C96C7;
fma.rn.f32 %f151, %f149, %f2, %f150;
mov.f32 %f152, 0fBF27E6EB;
fma.rn.f32 %f153, %f151, %f2, %f152;
mov.f32 %f154, 0f3F13C463;
fma.rn.f32 %f155, %f153, %f2, %f154;
mul.f32 %f156, %f2, %f155;
fma.rn.f32 %f11, %f156, %f2, %f2;
setp.gt.f32 %p12, %f11, 0f00000000;
setp.lt.f32 %p13, %f11, 0f7F800000;
and.pred %p14, %p12, %p13;
@%p14 bra BB50_14;
bra.uni BB50_13;
BB50_14:
setp.lt.f32 %p15, %f11, 0f00800000;
mul.f32 %f159, %f11, 0f4B800000;
selp.f32 %f160, %f159, %f11, %p15;
selp.f32 %f161, 0fC3170000, 0fC2FE0000, %p15;
mov.b32 %r11, %f160;
and.b32 %r12, %r11, 8388607;
or.b32 %r13, %r12, 1065353216;
mov.b32 %f162, %r13;
shr.u32 %r14, %r11, 23;
cvt.rn.f32.u32 %f163, %r14;
add.f32 %f164, %f161, %f163;
setp.gt.f32 %p16, %f162, 0f3FAE147B;
mul.f32 %f165, %f162, 0f3F000000;
add.f32 %f166, %f164, 0f3F800000;
selp.f32 %f167, %f165, %f162, %p16;
selp.f32 %f168, %f166, %f164, %p16;
add.f32 %f158, %f167, 0f3F800000;
add.f32 %f169, %f167, 0fBF800000;
// inline asm
rcp.approx.ftz.f32 %f157,%f158;
// inline asm
mul.f32 %f170, %f169, %f169;
neg.f32 %f171, %f170;
mul.rn.f32 %f172, %f157, %f171;
add.rn.f32 %f173, %f169, %f172;
mul.f32 %f174, %f173, %f173;
mov.f32 %f175, 0f3C4C6A36;
mov.f32 %f176, 0f3B1E94E6;
fma.rn.f32 %f177, %f176, %f174, %f175;
mov.f32 %f178, 0f3DAAAB1A;
fma.rn.f32 %f179, %f177, %f174, %f178;
mul.f32 %f180, %f174, %f179;
fma.rn.f32 %f181, %f180, %f173, %f172;
add.f32 %f182, %f169, %f181;
mov.f32 %f183, 0f3F317218;
fma.rn.f32 %f262, %f168, %f183, %f182;
bra.uni BB50_15;
BB50_2:
setp.ltu.f32 %p3, %f2, 0f40F9999A;
@%p3 bra BB50_7;
bra.uni BB50_3;
BB50_7:
add.f32 %f84, %f2, 0fC0400000;
mov.f32 %f85, 0fC640F6F8;
mov.f32 %f86, 0fC43B38FB;
fma.rn.f32 %f87, %f86, %f84, %f85;
mov.f32 %f88, 0fC7206560;
fma.rn.f32 %f89, %f87, %f84, %f88;
mov.f32 %f90, 0fC73CB6AA;
fma.rn.f32 %f91, %f89, %f84, %f90;
mov.f32 %f92, 0fC80BAE5A;
fma.rn.f32 %f93, %f91, %f84, %f92;
add.f32 %f94, %f84, 0fC381A020;
mov.f32 %f95, 0fC62864B8;
fma.rn.f32 %f96, %f94, %f84, %f95;
mov.f32 %f97, 0fC7B50686;
fma.rn.f32 %f98, %f96, %f84, %f97;
mov.f32 %f99, 0fC8498465;
fma.rn.f32 %f83, %f98, %f84, %f99;
// inline asm
rcp.approx.ftz.f32 %f82,%f83;
// inline asm
fma.rn.f32 %f270, %f93, %f82, %f84;
bra.uni BB50_16;
BB50_9:
add.f32 %f100, %f2, 0fC0000000;
mov.f32 %f101, 0fB967A002;
mov.f32 %f102, 0f385007FA;
fma.rn.f32 %f103, %f102, %f100, %f101;
mov.f32 %f104, 0f3A0DE6FC;
fma.rn.f32 %f105, %f103, %f100, %f104;
mov.f32 %f106, 0fBA9DE0E2;
fma.rn.f32 %f107, %f105, %f100, %f106;
mov.f32 %f108, 0f3B3D05B7;
fma.rn.f32 %f109, %f107, %f100, %f108;
mov.f32 %f110, 0fBBF1EB10;
fma.rn.f32 %f111, %f109, %f100, %f110;
mov.f32 %f112, 0f3CA89A28;
fma.rn.f32 %f113, %f111, %f100, %f112;
mov.f32 %f114, 0fBD89F01A;
fma.rn.f32 %f115, %f113, %f100, %f114;
mov.f32 %f116, 0f3EA51A66;
fma.rn.f32 %f117, %f115, %f100, %f116;
mov.f32 %f118, 0f3ED87730;
fma.rn.f32 %f119, %f117, %f100, %f118;
mul.f32 %f270, %f100, %f119;
bra.uni BB50_16;
BB50_3:
// inline asm
rcp.approx.ftz.f32 %f40,%f2;
// inline asm
mul.f32 %f42, %f40, %f40;
mov.f32 %f43, 0fBB360953;
mov.f32 %f44, 0f3A4BE755;
fma.rn.f32 %f45, %f44, %f42, %f43;
mov.f32 %f46, 0f3DAAAAA3;
fma.rn.f32 %f47, %f45, %f42, %f46;
mov.f32 %f48, 0f3F6B3F8E;
fma.rn.f32 %f3, %f47, %f40, %f48;
setp.lt.f32 %p4, %f2, 0f7F800000;
setp.gt.f32 %p5, %f2, 0f00000000;
and.pred %p6, %p5, %p4;
@%p6 bra BB50_5;
bra.uni BB50_4;
BB50_5:
setp.lt.f32 %p7, %f2, 0f00800000;
mul.f32 %f51, %f2, 0f4B800000;
selp.f32 %f52, %f51, %f2, %p7;
selp.f32 %f53, 0fC3170000, 0fC2FE0000, %p7;
mov.b32 %r7, %f52;
and.b32 %r8, %r7, 8388607;
or.b32 %r9, %r8, 1065353216;
mov.b32 %f54, %r9;
shr.u32 %r10, %r7, 23;
cvt.rn.f32.u32 %f55, %r10;
add.f32 %f56, %f53, %f55;
setp.gt.f32 %p8, %f54, 0f3FAE147B;
mul.f32 %f57, %f54, 0f3F000000;
add.f32 %f58, %f56, 0f3F800000;
selp.f32 %f59, %f57, %f54, %p8;
selp.f32 %f60, %f58, %f56, %p8;
add.f32 %f50, %f59, 0f3F800000;
add.f32 %f61, %f59, 0fBF800000;
// inline asm
rcp.approx.ftz.f32 %f49,%f50;
// inline asm
mul.f32 %f62, %f61, %f61;
neg.f32 %f63, %f62;
mul.rn.f32 %f64, %f49, %f63;
add.rn.f32 %f65, %f61, %f64;
mul.f32 %f66, %f65, %f65;
mov.f32 %f67, 0f3C4C6A36;
mov.f32 %f68, 0f3B1E94E6;
fma.rn.f32 %f69, %f68, %f66, %f67;
mov.f32 %f70, 0f3DAAAB1A;
fma.rn.f32 %f71, %f69, %f66, %f70;
mul.f32 %f72, %f66, %f71;
fma.rn.f32 %f73, %f72, %f65, %f64;
add.f32 %f74, %f61, %f73;
mov.f32 %f75, 0f3F317218;
fma.rn.f32 %f261, %f60, %f75, %f74;
bra.uni BB50_6;
BB50_11:
mov.f32 %f120, 0f3F800000;
sub.f32 %f121, %f120, %f2;
mov.f32 %f122, 0f3DD47577;
mov.f32 %f123, 0f3D3BEF76;
fma.rn.f32 %f124, %f123, %f121, %f122;
mov.f32 %f125, 0f3DFB8079;
fma.rn.f32 %f126, %f124, %f121, %f125;
mov.f32 %f127, 0f3E0295B5;
fma.rn.f32 %f128, %f126, %f121, %f127;
mov.f32 %f129, 0f3E12A765;
fma.rn.f32 %f130, %f128, %f121, %f129;
mov.f32 %f131, 0f3E2D6867;
fma.rn.f32 %f132, %f130, %f121, %f131;
mov.f32 %f133, 0f3E5462BF;
fma.rn.f32 %f134, %f132, %f121, %f133;
mov.f32 %f135, 0f3E8A8A72;
fma.rn.f32 %f136, %f134, %f121, %f135;
mov.f32 %f137, 0f3ECD26A4;
fma.rn.f32 %f138, %f136, %f121, %f137;
mov.f32 %f139, 0f3F528D32;
fma.rn.f32 %f140, %f138, %f121, %f139;
mov.f32 %f141, 0f3F13C468;
fma.rn.f32 %f142, %f140, %f121, %f141;
mul.f32 %f270, %f121, %f142;
bra.uni BB50_16;
BB50_4:
lg2.approx.f32 %f261, %f2;
BB50_6:
mul.f32 %f76, %f261, 0f3F000000;
add.f32 %f77, %f2, 0fBF000000;
mul.rn.f32 %f78, %f76, %f77;
sub.f32 %f79, %f78, %f2;
add.rn.f32 %f80, %f78, %f3;
add.f32 %f81, %f79, %f80;
setp.eq.f32 %p9, %f2, 0f7F800000;
selp.f32 %f270, %f2, %f81, %p9;
bra.uni BB50_16;
BB50_13:
lg2.approx.f32 %f262, %f11;
BB50_15:
neg.f32 %f270, %f262;
BB50_16:
mov.f32 %f16, %f270;
setp.ge.f32 %p17, %f1, 0f00000000;
mov.f32 %f269, %f16;
@%p17 bra BB50_37;
cvt.rmi.f32.f32 %f185, %f2;
setp.eq.f32 %p18, %f2, %f185;
mov.f32 %f184, 0f7F800000;
mov.f32 %f269, %f184;
@%p18 bra BB50_37;
setp.lt.f32 %p19, %f2, 0f1FEC1E4A;
@%p19 bra BB50_33;
bra.uni BB50_19;
BB50_33:
setp.gt.f32 %p29, %f2, 0f00000000;
setp.lt.f32 %p30, %f2, 0f7F800000;
and.pred %p31, %p29, %p30;
@%p31 bra BB50_35;
bra.uni BB50_34;
BB50_35:
setp.lt.f32 %p32, %f2, 0f00800000;
mul.f32 %f236, %f2, 0f4B800000;
selp.f32 %f237, %f236, %f2, %p32;
selp.f32 %f238, 0fC3170000, 0fC2FE0000, %p32;
mov.b32 %r20, %f237;
and.b32 %r21, %r20, 8388607;
or.b32 %r22, %r21, 1065353216;
mov.b32 %f239, %r22;
shr.u32 %r23, %r20, 23;
cvt.rn.f32.u32 %f240, %r23;
add.f32 %f241, %f238, %f240;
setp.gt.f32 %p33, %f239, 0f3FAE147B;
mul.f32 %f242, %f239, 0f3F000000;
add.f32 %f243, %f241, 0f3F800000;
selp.f32 %f244, %f242, %f239, %p33;
selp.f32 %f245, %f243, %f241, %p33;
add.f32 %f235, %f244, 0f3F800000;
add.f32 %f246, %f244, 0fBF800000;
// inline asm
rcp.approx.ftz.f32 %f234,%f235;
// inline asm
mul.f32 %f247, %f246, %f246;
neg.f32 %f248, %f247;
mul.rn.f32 %f249, %f234, %f248;
add.rn.f32 %f250, %f246, %f249;
mul.f32 %f251, %f250, %f250;
mov.f32 %f252, 0f3C4C6A36;
mov.f32 %f253, 0f3B1E94E6;
fma.rn.f32 %f254, %f253, %f251, %f252;
mov.f32 %f255, 0f3DAAAB1A;
fma.rn.f32 %f256, %f254, %f251, %f255;
mul.f32 %f257, %f251, %f256;
fma.rn.f32 %f258, %f257, %f250, %f249;
add.f32 %f259, %f246, %f258;
mov.f32 %f260, 0f3F317218;
fma.rn.f32 %f267, %f245, %f260, %f259;
bra.uni BB50_36;
BB50_19:
add.f32 %f186, %f2, %f2;
cvt.rni.f32.f32 %f187, %f186;
cvt.rzi.s32.f32 %r1, %f187;
neg.f32 %f188, %f187;
mov.f32 %f189, 0f3F000000;
fma.rn.f32 %f190, %f188, %f189, %f2;
mul.f32 %f17, %f190, 0f40490FDB;
mul.rn.f32 %f18, %f17, %f17;
and.b32 %r2, %r1, 1;
setp.eq.s32 %p20, %r2, 0;
@%p20 bra BB50_21;
mov.f32 %f191, 0fBAB6061A;
mov.f32 %f192, 0f37CCF5CE;
fma.rn.f32 %f263, %f192, %f18, %f191;
bra.uni BB50_22;
BB50_34:
lg2.approx.f32 %f267, %f2;
BB50_36:
neg.f32 %f269, %f267;
bra.uni BB50_37;
BB50_21:
mov.f32 %f193, 0f3C08839E;
mov.f32 %f194, 0fB94CA1F9;
fma.rn.f32 %f263, %f194, %f18, %f193;
BB50_22:
@%p20 bra BB50_24;
mov.f32 %f195, 0f3D2AAAA5;
fma.rn.f32 %f196, %f263, %f18, %f195;
mov.f32 %f197, 0fBF000000;
fma.rn.f32 %f264, %f196, %f18, %f197;
bra.uni BB50_25;
BB50_24:
mov.f32 %f198, 0fBE2AAAA3;
fma.rn.f32 %f199, %f263, %f18, %f198;
mov.f32 %f200, 0f00000000;
fma.rn.f32 %f264, %f199, %f18, %f200;
BB50_25:
fma.rn.f32 %f265, %f264, %f17, %f17;
@%p20 bra BB50_27;
mov.f32 %f201, 0f3F800000;
fma.rn.f32 %f265, %f264, %f18, %f201;
BB50_27:
and.b32 %r15, %r1, 2;
setp.eq.s32 %p23, %r15, 0;
@%p23 bra BB50_29;
mov.f32 %f202, 0f00000000;
mov.f32 %f203, 0fBF800000;
fma.rn.f32 %f265, %f265, %f203, %f202;
BB50_29:
abs.f32 %f204, %f265;
mul.f32 %f30, %f2, %f204;
setp.gt.f32 %p24, %f30, 0f00000000;
setp.lt.f32 %p25, %f30, 0f7F800000;
and.pred %p26, %p24, %p25;
@%p26 bra BB50_31;
bra.uni BB50_30;
BB50_31:
setp.lt.f32 %p27, %f30, 0f00800000;
mul.f32 %f207, %f30, 0f4B800000;
selp.f32 %f208, %f207, %f30, %p27;
selp.f32 %f209, 0fC3170000, 0fC2FE0000, %p27;
mov.b32 %r16, %f208;
and.b32 %r17, %r16, 8388607;
or.b32 %r18, %r17, 1065353216;
mov.b32 %f210, %r18;
shr.u32 %r19, %r16, 23;
cvt.rn.f32.u32 %f211, %r19;
add.f32 %f212, %f209, %f211;
setp.gt.f32 %p28, %f210, 0f3FAE147B;
mul.f32 %f213, %f210, 0f3F000000;
add.f32 %f214, %f212, 0f3F800000;
selp.f32 %f215, %f213, %f210, %p28;
selp.f32 %f216, %f214, %f212, %p28;
add.f32 %f206, %f215, 0f3F800000;
add.f32 %f217, %f215, 0fBF800000;
// inline asm
rcp.approx.ftz.f32 %f205,%f206;
// inline asm
mul.f32 %f218, %f217, %f217;
neg.f32 %f219, %f218;
mul.rn.f32 %f220, %f205, %f219;
add.rn.f32 %f221, %f217, %f220;
mul.f32 %f222, %f221, %f221;
mov.f32 %f223, 0f3C4C6A36;
mov.f32 %f224, 0f3B1E94E6;
fma.rn.f32 %f225, %f224, %f222, %f223;
mov.f32 %f226, 0f3DAAAB1A;
fma.rn.f32 %f227, %f225, %f222, %f226;
mul.f32 %f228, %f222, %f227;
fma.rn.f32 %f229, %f228, %f221, %f220;
add.f32 %f230, %f217, %f229;
mov.f32 %f231, 0f3F317218;
fma.rn.f32 %f266, %f216, %f231, %f230;
bra.uni BB50_32;
BB50_30:
lg2.approx.f32 %f266, %f30;
BB50_32:
mov.f32 %f232, 0f3F928682;
sub.f32 %f233, %f232, %f266;
sub.f32 %f269, %f233, %f16;
BB50_37:
cvta.to.global.u64 %rd8, %rd2;
add.s64 %rd10, %rd8, %rd6;
st.global.f32 [%rd10], %f269;
BB50_38:
ret;
}
// .globl vec_log10f
.visible .entry vec_log10f(
.param .u64 vec_log10f_param_0,
.param .u64 vec_log10f_param_1,
.param .u64 vec_log10f_param_2
)
{
.reg .pred %p<7>;
.reg .f32 %f<34>;
.reg .b32 %r<9>;
.reg .b64 %rd<11>;
ld.param.u64 %rd4, [vec_log10f_param_0];
ld.param.u64 %rd2, [vec_log10f_param_1];
ld.param.u64 %rd3, [vec_log10f_param_2];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB51_5;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f1, [%rd7];
setp.gt.f32 %p2, %f1, 0f00000000;
setp.lt.f32 %p3, %f1, 0f7F800000;
and.pred %p4, %p2, %p3;
@%p4 bra BB51_3;
bra.uni BB51_2;
BB51_3:
setp.lt.f32 %p5, %f1, 0f00800000;
mul.f32 %f7, %f1, 0f4B800000;
selp.f32 %f8, %f7, %f1, %p5;
selp.f32 %f9, 0fC3170000, 0fC2FE0000, %p5;
mov.b32 %r5, %f8;
and.b32 %r6, %r5, 8388607;
or.b32 %r7, %r6, 1065353216;
mov.b32 %f10, %r7;
shr.u32 %r8, %r5, 23;
cvt.rn.f32.u32 %f11, %r8;
add.f32 %f12, %f9, %f11;
setp.gt.f32 %p6, %f10, 0f3FAE147B;
mul.f32 %f13, %f10, 0f3F000000;
add.f32 %f14, %f12, 0f3F800000;
selp.f32 %f15, %f13, %f10, %p6;
selp.f32 %f16, %f14, %f12, %p6;
add.f32 %f6, %f15, 0f3F800000;
add.f32 %f17, %f15, 0fBF800000;
// inline asm
rcp.approx.ftz.f32 %f5,%f6;
// inline asm
mul.f32 %f18, %f17, %f17;
neg.f32 %f19, %f18;
mul.rn.f32 %f20, %f5, %f19;
add.rn.f32 %f21, %f17, %f20;
mul.f32 %f22, %f21, %f21;
mov.f32 %f23, 0f3C4C6A36;
mov.f32 %f24, 0f3B1E94E6;
fma.rn.f32 %f25, %f24, %f22, %f23;
mov.f32 %f26, 0f3DAAAB1A;
fma.rn.f32 %f27, %f25, %f22, %f26;
mul.f32 %f28, %f22, %f27;
fma.rn.f32 %f29, %f28, %f21, %f20;
add.f32 %f30, %f17, %f29;
mov.f32 %f31, 0f3F317218;
fma.rn.f32 %f33, %f16, %f31, %f30;
bra.uni BB51_4;
BB51_2:
lg2.approx.f32 %f33, %f1;
BB51_4:
cvta.to.global.u64 %rd8, %rd2;
add.s64 %rd10, %rd8, %rd6;
mul.f32 %f32, %f33, 0f3EDE5BD9;
st.global.f32 [%rd10], %f32;
BB51_5:
ret;
}
// .globl vec_log1pf
.visible .entry vec_log1pf(
.param .u64 vec_log1pf_param_0,
.param .u64 vec_log1pf_param_1,
.param .u64 vec_log1pf_param_2
)
{
.reg .pred %p<10>;
.reg .f32 %f<48>;
.reg .b32 %r<9>;
.reg .b64 %rd<11>;
ld.param.u64 %rd4, [vec_log1pf_param_0];
ld.param.u64 %rd2, [vec_log1pf_param_1];
ld.param.u64 %rd3, [vec_log1pf_param_2];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB52_7;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f1, [%rd7];
setp.le.f32 %p2, %f1, 0f3F266666;
setp.ge.f32 %p3, %f1, 0fBEC9BA5E;
and.pred %p4, %p3, %p2;
@%p4 bra BB52_5;
bra.uni BB52_2;
BB52_5:
add.f32 %f34, %f1, 0f40000000;
div.approx.f32 %f35, %f1, %f34;
neg.f32 %f36, %f1;
mul.rn.f32 %f37, %f36, %f35;
add.rn.f32 %f38, %f1, %f37;
mul.f32 %f39, %f38, %f38;
mov.f32 %f40, 0f3C4C4BE0;
mov.f32 %f41, 0f3B2063C3;
fma.rn.f32 %f42, %f41, %f39, %f40;
mov.f32 %f43, 0f3DAAAB50;
fma.rn.f32 %f44, %f42, %f39, %f43;
mul.f32 %f45, %f39, %f44;
fma.rn.f32 %f46, %f45, %f38, %f37;
add.f32 %f47, %f1, %f46;
bra.uni BB52_6;
BB52_2:
add.f32 %f2, %f1, 0f3F800000;
setp.gt.f32 %p5, %f2, 0f00000000;
setp.lt.f32 %p6, %f2, 0f7F800000;
and.pred %p7, %p5, %p6;
@%p7 bra BB52_4;
bra.uni BB52_3;
BB52_4:
setp.lt.f32 %p8, %f2, 0f00800000;
mul.f32 %f9, %f2, 0f4B800000;
selp.f32 %f10, %f9, %f2, %p8;
selp.f32 %f11, 0fC3170000, 0fC2FE0000, %p8;
mov.b32 %r5, %f10;
and.b32 %r6, %r5, 8388607;
or.b32 %r7, %r6, 1065353216;
mov.b32 %f12, %r7;
shr.u32 %r8, %r5, 23;
cvt.rn.f32.u32 %f13, %r8;
add.f32 %f14, %f11, %f13;
setp.gt.f32 %p9, %f12, 0f3FAE147B;
mul.f32 %f15, %f12, 0f3F000000;
add.f32 %f16, %f14, 0f3F800000;
selp.f32 %f17, %f15, %f12, %p9;
selp.f32 %f18, %f16, %f14, %p9;
add.f32 %f8, %f17, 0f3F800000;
add.f32 %f19, %f17, 0fBF800000;
// inline asm
rcp.approx.ftz.f32 %f7,%f8;
// inline asm
mul.f32 %f20, %f19, %f19;
neg.f32 %f21, %f20;
mul.rn.f32 %f22, %f7, %f21;
add.rn.f32 %f23, %f19, %f22;
mul.f32 %f24, %f23, %f23;
mov.f32 %f25, 0f3C4C6A36;
mov.f32 %f26, 0f3B1E94E6;
fma.rn.f32 %f27, %f26, %f24, %f25;
mov.f32 %f28, 0f3DAAAB1A;
fma.rn.f32 %f29, %f27, %f24, %f28;
mul.f32 %f30, %f24, %f29;
fma.rn.f32 %f31, %f30, %f23, %f22;
add.f32 %f32, %f19, %f31;
mov.f32 %f33, 0f3F317218;
fma.rn.f32 %f47, %f18, %f33, %f32;
bra.uni BB52_6;
BB52_3:
lg2.approx.f32 %f47, %f2;
BB52_6:
cvta.to.global.u64 %rd8, %rd2;
add.s64 %rd10, %rd8, %rd6;
st.global.f32 [%rd10], %f47;
BB52_7:
ret;
}
// .globl vec_log2f
.visible .entry vec_log2f(
.param .u64 vec_log2f_param_0,
.param .u64 vec_log2f_param_1,
.param .u64 vec_log2f_param_2
)
{
.reg .pred %p<7>;
.reg .f32 %f<34>;
.reg .b32 %r<9>;
.reg .b64 %rd<11>;
ld.param.u64 %rd4, [vec_log2f_param_0];
ld.param.u64 %rd2, [vec_log2f_param_1];
ld.param.u64 %rd3, [vec_log2f_param_2];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB53_5;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f1, [%rd7];
setp.gt.f32 %p2, %f1, 0f00000000;
setp.lt.f32 %p3, %f1, 0f7F800000;
and.pred %p4, %p2, %p3;
@%p4 bra BB53_3;
bra.uni BB53_2;
BB53_3:
setp.lt.f32 %p5, %f1, 0f00800000;
mul.f32 %f7, %f1, 0f4B800000;
selp.f32 %f8, %f7, %f1, %p5;
selp.f32 %f9, 0fC3170000, 0fC2FE0000, %p5;
mov.b32 %r5, %f8;
and.b32 %r6, %r5, 8388607;
or.b32 %r7, %r6, 1065353216;
mov.b32 %f10, %r7;
shr.u32 %r8, %r5, 23;
cvt.rn.f32.u32 %f11, %r8;
add.f32 %f12, %f9, %f11;
setp.gt.f32 %p6, %f10, 0f3FAE147B;
mul.f32 %f13, %f10, 0f3F000000;
add.f32 %f14, %f12, 0f3F800000;
selp.f32 %f15, %f13, %f10, %p6;
selp.f32 %f16, %f14, %f12, %p6;
add.f32 %f6, %f15, 0f3F800000;
add.f32 %f17, %f15, 0fBF800000;
// inline asm
rcp.approx.ftz.f32 %f5,%f6;
// inline asm
mul.f32 %f18, %f17, %f17;
neg.f32 %f19, %f18;
mul.rn.f32 %f20, %f5, %f19;
add.rn.f32 %f21, %f17, %f20;
mul.f32 %f22, %f21, %f21;
mov.f32 %f23, 0f3C4C6A36;
mov.f32 %f24, 0f3B1E94E6;
fma.rn.f32 %f25, %f24, %f22, %f23;
mov.f32 %f26, 0f3DAAAB1A;
fma.rn.f32 %f27, %f25, %f22, %f26;
mul.f32 %f28, %f22, %f27;
fma.rn.f32 %f29, %f28, %f21, %f20;
add.f32 %f30, %f17, %f29;
mov.f32 %f31, 0f3F317218;
fma.rn.f32 %f33, %f16, %f31, %f30;
bra.uni BB53_4;
BB53_2:
lg2.approx.f32 %f33, %f1;
BB53_4:
cvta.to.global.u64 %rd8, %rd2;
add.s64 %rd10, %rd8, %rd6;
mul.f32 %f32, %f33, 0f3FB8AA3B;
st.global.f32 [%rd10], %f32;
BB53_5:
ret;
}
// .globl vec_logbf
.visible .entry vec_logbf(
.param .u64 vec_logbf_param_0,
.param .u64 vec_logbf_param_1,
.param .u64 vec_logbf_param_2
)
{
.reg .pred %p<5>;
.reg .f32 %f<10>;
.reg .b32 %r<11>;
.reg .b64 %rd<11>;
ld.param.u64 %rd4, [vec_logbf_param_0];
ld.param.u64 %rd2, [vec_logbf_param_1];
ld.param.u64 %rd3, [vec_logbf_param_2];
mov.u32 %r2, %tid.x;
mov.u32 %r3, %ntid.x;
mov.u32 %r4, %ctaid.x;
mad.lo.s32 %r5, %r3, %r4, %r2;
cvt.s64.s32 %rd1, %r5;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB54_5;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f1, [%rd7];
abs.f32 %f5, %f1;
mov.b32 %r1, %f5;
setp.lt.u32 %p2, %r1, 8388608;
@%p2 bra BB54_3;
bra.uni BB54_2;
BB54_3:
clz.b32 %r8, %r1;
mov.u32 %r9, -118;
sub.s32 %r10, %r9, %r8;
cvt.rn.f32.s32 %f8, %r10;
setp.eq.f32 %p4, %f1, 0f00000000;
selp.f32 %f9, 0fFF800000, %f8, %p4;
bra.uni BB54_4;
BB54_2:
shr.u32 %r6, %r1, 23;
add.s32 %r7, %r6, -127;
cvt.rn.f32.s32 %f6, %r7;
mul.f32 %f7, %f1, %f1;
setp.gt.u32 %p3, %r1, 2139095039;
selp.f32 %f9, %f7, %f6, %p3;
BB54_4:
cvta.to.global.u64 %rd8, %rd2;
add.s64 %rd10, %rd8, %rd6;
st.global.f32 [%rd10], %f9;
BB54_5:
ret;
}
// .globl vec_logf
.visible .entry vec_logf(
.param .u64 vec_logf_param_0,
.param .u64 vec_logf_param_1,
.param .u64 vec_logf_param_2
)
{
.reg .pred %p<7>;
.reg .f32 %f<33>;
.reg .b32 %r<9>;
.reg .b64 %rd<11>;
ld.param.u64 %rd4, [vec_logf_param_0];
ld.param.u64 %rd2, [vec_logf_param_1];
ld.param.u64 %rd3, [vec_logf_param_2];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB55_5;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f1, [%rd7];
setp.gt.f32 %p2, %f1, 0f00000000;
setp.lt.f32 %p3, %f1, 0f7F800000;
and.pred %p4, %p2, %p3;
@%p4 bra BB55_3;
bra.uni BB55_2;
BB55_3:
setp.lt.f32 %p5, %f1, 0f00800000;
mul.f32 %f7, %f1, 0f4B800000;
selp.f32 %f8, %f7, %f1, %p5;
selp.f32 %f9, 0fC3170000, 0fC2FE0000, %p5;
mov.b32 %r5, %f8;
and.b32 %r6, %r5, 8388607;
or.b32 %r7, %r6, 1065353216;
mov.b32 %f10, %r7;
shr.u32 %r8, %r5, 23;
cvt.rn.f32.u32 %f11, %r8;
add.f32 %f12, %f9, %f11;
setp.gt.f32 %p6, %f10, 0f3FAE147B;
mul.f32 %f13, %f10, 0f3F000000;
add.f32 %f14, %f12, 0f3F800000;
selp.f32 %f15, %f13, %f10, %p6;
selp.f32 %f16, %f14, %f12, %p6;
add.f32 %f6, %f15, 0f3F800000;
add.f32 %f17, %f15, 0fBF800000;
// inline asm
rcp.approx.ftz.f32 %f5,%f6;
// inline asm
mul.f32 %f18, %f17, %f17;
neg.f32 %f19, %f18;
mul.rn.f32 %f20, %f5, %f19;
add.rn.f32 %f21, %f17, %f20;
mul.f32 %f22, %f21, %f21;
mov.f32 %f23, 0f3C4C6A36;
mov.f32 %f24, 0f3B1E94E6;
fma.rn.f32 %f25, %f24, %f22, %f23;
mov.f32 %f26, 0f3DAAAB1A;
fma.rn.f32 %f27, %f25, %f22, %f26;
mul.f32 %f28, %f22, %f27;
fma.rn.f32 %f29, %f28, %f21, %f20;
add.f32 %f30, %f17, %f29;
mov.f32 %f31, 0f3F317218;
fma.rn.f32 %f32, %f16, %f31, %f30;
bra.uni BB55_4;
BB55_2:
lg2.approx.f32 %f32, %f1;
BB55_4:
cvta.to.global.u64 %rd8, %rd2;
add.s64 %rd10, %rd8, %rd6;
st.global.f32 [%rd10], %f32;
BB55_5:
ret;
}
// .globl vec_normcdff
.visible .entry vec_normcdff(
.param .u64 vec_normcdff_param_0,
.param .u64 vec_normcdff_param_1,
.param .u64 vec_normcdff_param_2
)
{
.reg .pred %p<6>;
.reg .f32 %f<81>;
.reg .b32 %r<8>;
.reg .b64 %rd<11>;
ld.param.u64 %rd4, [vec_normcdff_param_0];
ld.param.u64 %rd2, [vec_normcdff_param_1];
ld.param.u64 %rd3, [vec_normcdff_param_2];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB56_4;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f13, [%rd7];
abs.f32 %f14, %f13;
setp.gt.f32 %p2, %f14, 0f41680000;
mov.b32 %r5, %f13;
and.b32 %r6, %r5, -2147483648;
or.b32 %r7, %r6, 1097334784;
mov.b32 %f15, %r7;
selp.f32 %f16, %f15, %f13, %p2;
mov.f32 %f17, 0fBF3504F3;
mul.rn.f32 %f1, %f16, %f17;
neg.f32 %f18, %f1;
fma.rn.f32 %f19, %f16, %f17, %f18;
mov.f32 %f20, 0fB24FE77A;
fma.rn.f32 %f2, %f16, %f20, %f19;
add.rn.f32 %f3, %f1, %f2;
abs.f32 %f21, %f3;
add.f32 %f8, %f21, 0f40800000;
// inline asm
rcp.approx.ftz.f32 %f7,%f8;
// inline asm
add.f32 %f22, %f21, 0fC0800000;
mul.rn.f32 %f23, %f22, %f7;
add.f32 %f24, %f23, 0f3F800000;
mov.f32 %f25, 0fC0800000;
fma.rn.f32 %f26, %f25, %f24, %f21;
neg.f32 %f27, %f23;
fma.rn.f32 %f28, %f27, %f21, %f26;
fma.rn.f32 %f29, %f7, %f28, %f23;
mov.f32 %f30, 0f3BE6E05B;
mov.f32 %f31, 0f3A69A091;
fma.rn.f32 %f32, %f31, %f29, %f30;
mov.f32 %f33, 0fBC81FB4B;
fma.rn.f32 %f34, %f32, %f29, %f33;
mov.f32 %f35, 0f3D15373B;
fma.rn.f32 %f36, %f34, %f29, %f35;
mov.f32 %f37, 0fBD887C5A;
fma.rn.f32 %f38, %f36, %f29, %f37;
mov.f32 %f39, 0f3DC021D5;
fma.rn.f32 %f40, %f38, %f29, %f39;
mov.f32 %f41, 0fBDCED424;
fma.rn.f32 %f42, %f40, %f29, %f41;
mov.f32 %f43, 0f3D8B74DE;
fma.rn.f32 %f44, %f42, %f29, %f43;
mov.f32 %f45, 0f3C7BF170;
fma.rn.f32 %f46, %f44, %f29, %f45;
mov.f32 %f47, 0fBE0EF8D4;
fma.rn.f32 %f48, %f46, %f29, %f47;
mov.f32 %f49, 0f3F9DD2C9;
fma.rn.f32 %f50, %f48, %f29, %f49;
mov.f32 %f51, 0f3F800000;
mov.f32 %f52, 0f40000000;
fma.rn.f32 %f10, %f52, %f21, %f51;
// inline asm
rcp.approx.ftz.f32 %f9,%f10;
// inline asm
mul.rn.f32 %f53, %f50, %f9;
mul.f32 %f54, %f53, 0fC0000000;
fma.rn.f32 %f55, %f21, %f54, %f50;
sub.f32 %f56, %f55, %f53;
fma.rn.f32 %f57, %f56, %f9, %f53;
mul.f32 %f58, %f21, %f21;
neg.f32 %f59, %f58;
mul.f32 %f60, %f58, 0fBFB8AA3B;
cvt.rzi.f32.f32 %f61, %f60;
mov.f32 %f62, 0fBF317200;
fma.rn.f32 %f63, %f61, %f62, %f59;
mov.f32 %f64, 0fB5BFBE8E;
fma.rn.f32 %f65, %f61, %f64, %f63;
mul.f32 %f12, %f65, 0f3FB8AA3B;
// inline asm
ex2.approx.ftz.f32 %f11,%f12;
// inline asm
add.f32 %f66, %f61, 0f00000000;
ex2.approx.f32 %f67, %f66;
mul.f32 %f68, %f11, %f67;
neg.f32 %f69, %f21;
fma.rn.f32 %f70, %f69, %f21, %f58;
fma.rn.f32 %f71, %f68, %f70, %f68;
mul.f32 %f72, %f57, %f71;
setp.gt.f32 %p3, %f21, 0f4120E148;
selp.f32 %f73, 0f00000000, %f72, %p3;
setp.lt.f32 %p4, %f3, 0f00000000;
sub.f32 %f74, %f52, %f73;
selp.f32 %f80, %f74, %f73, %p4;
setp.geu.f32 %p5, %f16, 0fBF800000;
@%p5 bra BB56_3;
sub.f32 %f75, %f1, %f3;
add.rn.f32 %f76, %f75, %f2;
mul.f32 %f77, %f3, 0fC0000000;
mul.f32 %f78, %f77, %f80;
fma.rn.f32 %f80, %f78, %f76, %f80;
BB56_3:
cvta.to.global.u64 %rd8, %rd2;
add.s64 %rd10, %rd8, %rd6;
mul.f32 %f79, %f80, 0f3F000000;
st.global.f32 [%rd10], %f79;
BB56_4:
ret;
}
// .globl vec_normcdfinvf
.visible .entry vec_normcdfinvf(
.param .u64 vec_normcdfinvf_param_0,
.param .u64 vec_normcdfinvf_param_1,
.param .u64 vec_normcdfinvf_param_2
)
{
.reg .pred %p<6>;
.reg .f32 %f<60>;
.reg .b32 %r<5>;
.reg .b64 %rd<11>;
ld.param.u64 %rd4, [vec_normcdfinvf_param_0];
ld.param.u64 %rd2, [vec_normcdfinvf_param_1];
ld.param.u64 %rd3, [vec_normcdfinvf_param_2];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB57_5;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f7, [%rd7];
add.f32 %f1, %f7, %f7;
neg.f32 %f2, %f1;
mov.f32 %f8, 0f40000000;
add.rn.f32 %f3, %f8, %f2;
setp.le.f32 %p2, %f1, 0f3FFF9097;
setp.ge.f32 %p3, %f1, 0f3B5ED289;
and.pred %p4, %p3, %p2;
@%p4 bra BB57_3;
bra.uni BB57_2;
BB57_3:
mul.rn.f32 %f35, %f3, %f1;
// inline asm
lg2.approx.ftz.f32 %f34,%f35;
// inline asm
neg.f32 %f36, %f34;
mov.f32 %f37, 0f3221F645;
mov.f32 %f38, 0fAF8A6370;
fma.rn.f32 %f39, %f38, %f36, %f37;
mov.f32 %f40, 0fB4016FDA;
fma.rn.f32 %f41, %f39, %f36, %f40;
mov.f32 %f42, 0f3468F846;
fma.rn.f32 %f43, %f41, %f36, %f42;
mov.f32 %f44, 0f370742AA;
fma.rn.f32 %f45, %f43, %f36, %f44;
mov.f32 %f46, 0fB804DB4D;
fma.rn.f32 %f47, %f45, %f36, %f46;
mov.f32 %f48, 0fBA4AFEA1;
fma.rn.f32 %f49, %f47, %f36, %f48;
mov.f32 %f50, 0f3BB5C027;
fma.rn.f32 %f51, %f49, %f36, %f50;
mov.f32 %f52, 0f3E24AE0F;
fma.rn.f32 %f53, %f51, %f36, %f52;
mov.f32 %f54, 0f3F62DFC4;
fma.rn.f32 %f55, %f53, %f36, %f54;
fma.rn.f32 %f59, %f55, %f2, %f55;
bra.uni BB57_4;
BB57_2:
setp.gt.f32 %p5, %f1, 0f3F800000;
selp.f32 %f13, %f3, %f1, %p5;
lg2.approx.f32 %f14, %f13;
neg.f32 %f10, %f14;
// inline asm
rsqrt.approx.ftz.f32 %f9,%f10;
// inline asm
mov.f32 %f15, 0f42FEF829;
mov.f32 %f16, 0fC27C73F1;
fma.rn.f32 %f17, %f16, %f9, %f15;
mov.f32 %f18, 0fC2E4361C;
fma.rn.f32 %f19, %f17, %f9, %f18;
mov.f32 %f20, 0f42714D9B;
fma.rn.f32 %f21, %f19, %f9, %f20;
mov.f32 %f22, 0fC1AE51B3;
fma.rn.f32 %f23, %f21, %f9, %f22;
mov.f32 %f24, 0f40CEF504;
fma.rn.f32 %f25, %f23, %f9, %f24;
mov.f32 %f26, 0fBFEA9E05;
fma.rn.f32 %f27, %f25, %f9, %f26;
mov.f32 %f28, 0fBCF871F4;
fma.rn.f32 %f29, %f27, %f9, %f28;
mov.f32 %f30, 0f3F553775;
fma.rn.f32 %f31, %f29, %f9, %f30;
// inline asm
rcp.approx.ftz.f32 %f11,%f9;
// inline asm
mul.rn.f32 %f32, %f31, %f11;
neg.f32 %f33, %f32;
selp.f32 %f59, %f33, %f32, %p5;
BB57_4:
cvta.to.global.u64 %rd8, %rd2;
mul.f32 %f56, %f59, 0fBFB504F3;
mov.f32 %f57, 0f00000000;
add.rn.f32 %f58, %f56, %f57;
add.s64 %rd10, %rd8, %rd6;
st.global.f32 [%rd10], %f58;
BB57_5:
ret;
}
// .globl vec_rcbrtf
.visible .entry vec_rcbrtf(
.param .u64 vec_rcbrtf_param_0,
.param .u64 vec_rcbrtf_param_1,
.param .u64 vec_rcbrtf_param_2
)
{
.reg .pred %p<4>;
.reg .f32 %f<21>;
.reg .b32 %r<6>;
.reg .b64 %rd<11>;
ld.param.u64 %rd4, [vec_rcbrtf_param_0];
ld.param.u64 %rd2, [vec_rcbrtf_param_1];
ld.param.u64 %rd3, [vec_rcbrtf_param_2];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB58_4;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f1, [%rd7];
abs.f32 %f7, %f1;
lg2.approx.f32 %f8, %f7;
mul.f32 %f6, %f8, 0fBEAAAAAB;
// inline asm
ex2.approx.ftz.f32 %f5,%f6;
// inline asm
mul.f32 %f9, %f7, %f5;
neg.f32 %f10, %f9;
mul.f32 %f11, %f5, %f5;
mov.f32 %f12, 0f3F800000;
fma.rn.f32 %f13, %f11, %f10, %f12;
mul.f32 %f14, %f5, 0f3EAAAAAB;
fma.rn.f32 %f15, %f13, %f14, %f5;
mov.b32 %r5, %f1;
setp.lt.s32 %p2, %r5, 0;
neg.f32 %f16, %f15;
selp.f32 %f20, %f16, %f15, %p2;
add.f32 %f17, %f1, %f1;
setp.neu.f32 %p3, %f17, %f1;
@%p3 bra BB58_3;
// inline asm
rcp.approx.ftz.f32 %f20,%f1;
// inline asm
BB58_3:
cvta.to.global.u64 %rd8, %rd2;
add.s64 %rd10, %rd8, %rd6;
st.global.f32 [%rd10], %f20;
BB58_4:
ret;
}
// .globl vec_rintf
.visible .entry vec_rintf(
.param .u64 vec_rintf_param_0,
.param .u64 vec_rintf_param_1,
.param .u64 vec_rintf_param_2
)
{
.reg .pred %p<2>;
.reg .f32 %f<3>;
.reg .b32 %r<5>;
.reg .b64 %rd<10>;
ld.param.u64 %rd4, [vec_rintf_param_0];
ld.param.u64 %rd2, [vec_rintf_param_1];
ld.param.u64 %rd3, [vec_rintf_param_2];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB59_2;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f1, [%rd7];
cvt.rni.f32.f32 %f2, %f1;
cvta.to.global.u64 %rd8, %rd2;
add.s64 %rd9, %rd8, %rd6;
st.global.f32 [%rd9], %f2;
BB59_2:
ret;
}
// .globl vec_roundf
.visible .entry vec_roundf(
.param .u64 vec_roundf_param_0,
.param .u64 vec_roundf_param_1,
.param .u64 vec_roundf_param_2
)
{
.reg .pred %p<4>;
.reg .f32 %f<10>;
.reg .b32 %r<8>;
.reg .b64 %rd<11>;
ld.param.u64 %rd4, [vec_roundf_param_0];
ld.param.u64 %rd2, [vec_roundf_param_1];
ld.param.u64 %rd3, [vec_roundf_param_2];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB60_4;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f1, [%rd7];
abs.f32 %f5, %f1;
mov.b32 %r5, %f1;
and.b32 %r6, %r5, -2147483648;
or.b32 %r7, %r6, 1056964608;
mov.b32 %f6, %r7;
add.f32 %f7, %f1, %f6;
cvt.rzi.f32.f32 %f8, %f7;
setp.gt.f32 %p2, %f5, 0f4B000000;
selp.f32 %f9, %f1, %f8, %p2;
setp.geu.f32 %p3, %f5, 0f3F000000;
@%p3 bra BB60_3;
cvt.rzi.f32.f32 %f9, %f1;
BB60_3:
cvta.to.global.u64 %rd8, %rd2;
add.s64 %rd10, %rd8, %rd6;
st.global.f32 [%rd10], %f9;
BB60_4:
ret;
}
// .globl vec_rsqrtf
.visible .entry vec_rsqrtf(
.param .u64 vec_rsqrtf_param_0,
.param .u64 vec_rsqrtf_param_1,
.param .u64 vec_rsqrtf_param_2
)
{
.reg .pred %p<2>;
.reg .f32 %f<3>;
.reg .b32 %r<5>;
.reg .b64 %rd<10>;
ld.param.u64 %rd4, [vec_rsqrtf_param_0];
ld.param.u64 %rd2, [vec_rsqrtf_param_1];
ld.param.u64 %rd3, [vec_rsqrtf_param_2];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB61_2;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f1, [%rd7];
rsqrt.approx.f32 %f2, %f1;
cvta.to.global.u64 %rd8, %rd2;
add.s64 %rd9, %rd8, %rd6;
st.global.f32 [%rd9], %f2;
BB61_2:
ret;
}
// .globl vec_sinf
.visible .entry vec_sinf(
.param .u64 vec_sinf_param_0,
.param .u64 vec_sinf_param_1,
.param .u64 vec_sinf_param_2
)
{
.local .align 4 .b8 __local_depot62[28];
.reg .b64 %SP;
.reg .b64 %SPL;
.reg .pred %p<15>;
.reg .f32 %f<48>;
.reg .b32 %r<96>;
.reg .b64 %rd<23>;
mov.u64 %rd22, __local_depot62;
cvta.local.u64 %SP, %rd22;
ld.param.u64 %rd9, [vec_sinf_param_0];
ld.param.u64 %rd7, [vec_sinf_param_1];
ld.param.u64 %rd8, [vec_sinf_param_2];
add.u64 %rd10, %SP, 0;
cvta.to.local.u64 %rd1, %rd10;
mov.u32 %r38, %ntid.x;
mov.u32 %r39, %ctaid.x;
mov.u32 %r40, %tid.x;
mad.lo.s32 %r1, %r38, %r39, %r40;
cvt.s64.s32 %rd11, %r1;
setp.ge.u64 %p1, %rd11, %rd9;
@%p1 bra BB62_24;
cvta.to.global.u64 %rd12, %rd8;
mul.wide.s32 %rd13, %r1, 4;
add.s64 %rd14, %rd12, %rd13;
ld.global.f32 %f43, [%rd14];
abs.f32 %f19, %f43;
setp.neu.f32 %p2, %f19, 0f7F800000;
@%p2 bra BB62_3;
mov.f32 %f20, 0f00000000;
mul.rn.f32 %f43, %f43, %f20;
BB62_3:
mul.f32 %f21, %f43, 0f3F22F983;
cvt.rni.s32.f32 %r95, %f21;
cvt.rn.f32.s32 %f22, %r95;
neg.f32 %f23, %f22;
mov.f32 %f24, 0f3FC90FDA;
fma.rn.f32 %f25, %f23, %f24, %f43;
mov.f32 %f26, 0f33A22168;
fma.rn.f32 %f27, %f23, %f26, %f25;
mov.f32 %f28, 0f27C234C5;
fma.rn.f32 %f44, %f23, %f28, %f27;
abs.f32 %f29, %f43;
setp.leu.f32 %p3, %f29, 0f47CE4780;
@%p3 bra BB62_13;
mov.b32 %r3, %f43;
shr.u32 %r4, %r3, 23;
bfe.u32 %r43, %r3, 23, 8;
add.s32 %r44, %r43, -128;
shl.b32 %r45, %r3, 8;
or.b32 %r5, %r45, -2147483648;
shr.u32 %r6, %r44, 5;
mov.u32 %r87, 0;
mov.u64 %rd20, __cudart_i2opi_f;
mov.u32 %r86, -6;
mov.u64 %rd21, %rd1;
BB62_5:
.pragma "nounroll";
mov.u64 %rd3, %rd21;
ld.const.u32 %r48, [%rd20];
// inline asm
{
mad.lo.cc.u32 %r46, %r48, %r5, %r87;
madc.hi.u32 %r87, %r48, %r5, 0;
}
// inline asm
st.local.u32 [%rd3], %r46;
add.s64 %rd4, %rd3, 4;
add.s64 %rd20, %rd20, 4;
add.s32 %r86, %r86, 1;
setp.ne.s32 %p4, %r86, 0;
mov.u64 %rd21, %rd4;
@%p4 bra BB62_5;
and.b32 %r11, %r3, -2147483648;
st.local.u32 [%rd1+24], %r87;
mov.u32 %r51, 6;
sub.s32 %r52, %r51, %r6;
mul.wide.s32 %rd16, %r52, 4;
add.s64 %rd6, %rd1, %rd16;
ld.local.u32 %r88, [%rd6];
ld.local.u32 %r89, [%rd6+-4];
and.b32 %r14, %r4, 31;
setp.eq.s32 %p5, %r14, 0;
@%p5 bra BB62_8;
mov.u32 %r53, 32;
sub.s32 %r54, %r53, %r14;
shr.u32 %r55, %r89, %r54;
shl.b32 %r56, %r88, %r14;
add.s32 %r88, %r55, %r56;
ld.local.u32 %r57, [%rd6+-8];
shr.u32 %r58, %r57, %r54;
shl.b32 %r59, %r89, %r14;
add.s32 %r89, %r58, %r59;
BB62_8:
shr.u32 %r60, %r89, 30;
shl.b32 %r61, %r88, 2;
add.s32 %r90, %r60, %r61;
shl.b32 %r20, %r89, 2;
shr.u32 %r62, %r90, 31;
shr.u32 %r63, %r88, 30;
add.s32 %r21, %r62, %r63;
setp.eq.s32 %p6, %r62, 0;
mov.u32 %r91, %r11;
mov.u32 %r92, %r20;
@%p6 bra BB62_10;
not.b32 %r64, %r90;
neg.s32 %r22, %r20;
setp.eq.s32 %p7, %r20, 0;
selp.u32 %r65, 1, 0, %p7;
add.s32 %r90, %r65, %r64;
xor.b32 %r24, %r11, -2147483648;
mov.u32 %r91, %r24;
mov.u32 %r92, %r22;
BB62_10:
mov.u32 %r26, %r91;
neg.s32 %r66, %r21;
setp.eq.s32 %p8, %r11, 0;
selp.b32 %r95, %r21, %r66, %p8;
clz.b32 %r94, %r90;
setp.eq.s32 %p9, %r94, 0;
shl.b32 %r67, %r90, %r94;
mov.u32 %r68, 32;
sub.s32 %r69, %r68, %r94;
shr.u32 %r70, %r92, %r69;
add.s32 %r71, %r70, %r67;
selp.b32 %r30, %r90, %r71, %p9;
mov.u32 %r72, -921707870;
mul.hi.u32 %r93, %r30, %r72;
setp.lt.s32 %p10, %r93, 1;
@%p10 bra BB62_12;
mul.lo.s32 %r73, %r30, -921707870;
shr.u32 %r74, %r73, 31;
shl.b32 %r75, %r93, 1;
add.s32 %r93, %r74, %r75;
add.s32 %r94, %r94, 1;
BB62_12:
mov.u32 %r76, 126;
sub.s32 %r77, %r76, %r94;
shl.b32 %r78, %r77, 23;
add.s32 %r79, %r93, 1;
shr.u32 %r80, %r79, 7;
add.s32 %r81, %r80, 1;
shr.u32 %r82, %r81, 1;
add.s32 %r83, %r82, %r78;
or.b32 %r84, %r83, %r26;
mov.b32 %f44, %r84;
BB62_13:
mul.rn.f32 %f7, %f44, %f44;
and.b32 %r37, %r95, 1;
setp.eq.s32 %p11, %r37, 0;
@%p11 bra BB62_15;
mov.f32 %f30, 0fBAB6061A;
mov.f32 %f31, 0f37CCF5CE;
fma.rn.f32 %f45, %f31, %f7, %f30;
bra.uni BB62_16;
BB62_15:
mov.f32 %f32, 0f3C08839E;
mov.f32 %f33, 0fB94CA1F9;
fma.rn.f32 %f45, %f33, %f7, %f32;
BB62_16:
@%p11 bra BB62_18;
mov.f32 %f34, 0f3D2AAAA5;
fma.rn.f32 %f35, %f45, %f7, %f34;
mov.f32 %f36, 0fBF000000;
fma.rn.f32 %f46, %f35, %f7, %f36;
bra.uni BB62_19;
BB62_18:
mov.f32 %f37, 0fBE2AAAA3;
fma.rn.f32 %f38, %f45, %f7, %f37;
mov.f32 %f39, 0f00000000;
fma.rn.f32 %f46, %f38, %f7, %f39;
BB62_19:
fma.rn.f32 %f47, %f46, %f44, %f44;
@%p11 bra BB62_21;
mov.f32 %f40, 0f3F800000;
fma.rn.f32 %f47, %f46, %f7, %f40;
BB62_21:
and.b32 %r85, %r95, 2;
setp.eq.s32 %p14, %r85, 0;
@%p14 bra BB62_23;
mov.f32 %f41, 0f00000000;
mov.f32 %f42, 0fBF800000;
fma.rn.f32 %f47, %f47, %f42, %f41;
BB62_23:
cvta.to.global.u64 %rd17, %rd7;
add.s64 %rd19, %rd17, %rd13;
st.global.f32 [%rd19], %f47;
BB62_24:
ret;
}
// .globl vec_sinhf
.visible .entry vec_sinhf(
.param .u64 vec_sinhf_param_0,
.param .u64 vec_sinhf_param_1,
.param .u64 vec_sinhf_param_2
)
{
.reg .pred %p<4>;
.reg .f32 %f<32>;
.reg .b32 %r<10>;
.reg .b64 %rd<11>;
ld.param.u64 %rd4, [vec_sinhf_param_0];
ld.param.u64 %rd2, [vec_sinhf_param_1];
ld.param.u64 %rd3, [vec_sinhf_param_2];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB63_5;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f1, [%rd7];
abs.f32 %f2, %f1;
setp.ltu.f32 %p2, %f2, 0f3F800000;
@%p2 bra BB63_3;
bra.uni BB63_2;
BB63_3:
mul.f32 %f22, %f1, %f1;
mov.f32 %f23, 0f394FFF49;
mov.f32 %f24, 0f363D0ADA;
fma.rn.f32 %f25, %f24, %f22, %f23;
mov.f32 %f26, 0f3C08889A;
fma.rn.f32 %f27, %f25, %f22, %f26;
mov.f32 %f28, 0f3E2AAAAB;
fma.rn.f32 %f29, %f27, %f22, %f28;
mul.f32 %f30, %f22, %f29;
fma.rn.f32 %f31, %f30, %f1, %f1;
bra.uni BB63_4;
BB63_2:
mul.f32 %f8, %f2, 0f3FB8AA3B;
cvt.rzi.f32.f32 %f9, %f8;
mov.f32 %f10, 0fBF317200;
fma.rn.f32 %f11, %f9, %f10, %f2;
mov.f32 %f12, 0fB5BFBE8E;
fma.rn.f32 %f13, %f9, %f12, %f11;
mul.f32 %f7, %f13, 0f3FB8AA3B;
// inline asm
ex2.approx.ftz.f32 %f6,%f7;
// inline asm
add.f32 %f14, %f9, 0fC0000000;
ex2.approx.f32 %f15, %f14;
mul.f32 %f16, %f6, %f15;
mov.f32 %f17, 0f3E000000;
div.approx.f32 %f18, %f17, %f16;
neg.f32 %f19, %f18;
mov.f32 %f20, 0f40000000;
fma.rn.f32 %f21, %f20, %f16, %f19;
mov.b32 %r5, %f21;
setp.ltu.f32 %p3, %f2, 0f42B40000;
selp.b32 %r6, %r5, 2139095040, %p3;
mov.b32 %r7, %f1;
and.b32 %r8, %r7, -2147483648;
or.b32 %r9, %r6, %r8;
mov.b32 %f31, %r9;
BB63_4:
cvta.to.global.u64 %rd8, %rd2;
add.s64 %rd10, %rd8, %rd6;
st.global.f32 [%rd10], %f31;
BB63_5:
ret;
}
// .globl vec_sinpif
.visible .entry vec_sinpif(
.param .u64 vec_sinpif_param_0,
.param .u64 vec_sinpif_param_1,
.param .u64 vec_sinpif_param_2
)
{
.reg .pred %p<7>;
.reg .f32 %f<42>;
.reg .b32 %r<8>;
.reg .b64 %rd<11>;
ld.param.u64 %rd4, [vec_sinpif_param_0];
ld.param.u64 %rd2, [vec_sinpif_param_1];
ld.param.u64 %rd3, [vec_sinpif_param_2];
mov.u32 %r3, %tid.x;
mov.u32 %r4, %ntid.x;
mov.u32 %r5, %ctaid.x;
mad.lo.s32 %r6, %r4, %r5, %r3;
cvt.s64.s32 %rd1, %r6;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB64_14;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f1, [%rd7];
add.f32 %f17, %f1, %f1;
cvt.rni.f32.f32 %f18, %f17;
cvt.rzi.s32.f32 %r1, %f18;
neg.f32 %f19, %f18;
mov.f32 %f20, 0f3F000000;
fma.rn.f32 %f21, %f19, %f20, %f1;
mul.f32 %f22, %f21, 0f34222169;
mov.f32 %f23, 0f40490FDA;
fma.rn.f32 %f2, %f21, %f23, %f22;
mul.rn.f32 %f3, %f2, %f2;
and.b32 %r2, %r1, 1;
setp.eq.s32 %p2, %r2, 0;
@%p2 bra BB64_3;
mov.f32 %f24, 0fBAB6061A;
mov.f32 %f25, 0f37CCF5CE;
fma.rn.f32 %f39, %f25, %f3, %f24;
bra.uni BB64_4;
BB64_3:
mov.f32 %f26, 0f3C08839E;
mov.f32 %f27, 0fB94CA1F9;
fma.rn.f32 %f39, %f27, %f3, %f26;
BB64_4:
@%p2 bra BB64_6;
mov.f32 %f28, 0f3D2AAAA5;
fma.rn.f32 %f29, %f39, %f3, %f28;
mov.f32 %f30, 0fBF000000;
fma.rn.f32 %f40, %f29, %f3, %f30;
bra.uni BB64_7;
BB64_6:
mov.f32 %f31, 0fBE2AAAA3;
fma.rn.f32 %f32, %f39, %f3, %f31;
mov.f32 %f33, 0f00000000;
fma.rn.f32 %f40, %f32, %f3, %f33;
BB64_7:
fma.rn.f32 %f41, %f40, %f2, %f2;
@%p2 bra BB64_9;
mov.f32 %f34, 0f3F800000;
fma.rn.f32 %f41, %f40, %f3, %f34;
BB64_9:
and.b32 %r7, %r1, 2;
setp.eq.s32 %p5, %r7, 0;
@%p5 bra BB64_11;
mov.f32 %f35, 0f00000000;
mov.f32 %f36, 0fBF800000;
fma.rn.f32 %f41, %f41, %f36, %f35;
BB64_11:
cvt.rzi.f32.f32 %f37, %f1;
setp.neu.f32 %p6, %f37, %f1;
@%p6 bra BB64_13;
mov.f32 %f38, 0f00000000;
mul.rn.f32 %f41, %f1, %f38;
BB64_13:
cvta.to.global.u64 %rd8, %rd2;
add.s64 %rd10, %rd8, %rd6;
st.global.f32 [%rd10], %f41;
BB64_14:
ret;
}
// .globl vec_sqrtf
.visible .entry vec_sqrtf(
.param .u64 vec_sqrtf_param_0,
.param .u64 vec_sqrtf_param_1,
.param .u64 vec_sqrtf_param_2
)
{
.reg .pred %p<2>;
.reg .f32 %f<3>;
.reg .b32 %r<5>;
.reg .b64 %rd<10>;
ld.param.u64 %rd4, [vec_sqrtf_param_0];
ld.param.u64 %rd2, [vec_sqrtf_param_1];
ld.param.u64 %rd3, [vec_sqrtf_param_2];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB65_2;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f1, [%rd7];
sqrt.rn.f32 %f2, %f1;
cvta.to.global.u64 %rd8, %rd2;
add.s64 %rd9, %rd8, %rd6;
st.global.f32 [%rd9], %f2;
BB65_2:
ret;
}
// .globl vec_tanf
.visible .entry vec_tanf(
.param .u64 vec_tanf_param_0,
.param .u64 vec_tanf_param_1,
.param .u64 vec_tanf_param_2
)
{
.local .align 4 .b8 __local_depot66[28];
.reg .b64 %SP;
.reg .b64 %SPL;
.reg .pred %p<12>;
.reg .f32 %f<33>;
.reg .b32 %r<95>;
.reg .b64 %rd<23>;
mov.u64 %rd22, __local_depot66;
cvta.local.u64 %SP, %rd22;
ld.param.u64 %rd9, [vec_tanf_param_0];
ld.param.u64 %rd7, [vec_tanf_param_1];
ld.param.u64 %rd8, [vec_tanf_param_2];
add.u64 %rd10, %SP, 0;
cvta.to.local.u64 %rd1, %rd10;
mov.u32 %r37, %ntid.x;
mov.u32 %r38, %ctaid.x;
mov.u32 %r39, %tid.x;
mad.lo.s32 %r1, %r37, %r38, %r39;
cvt.s64.s32 %rd11, %r1;
setp.ge.u64 %p1, %rd11, %rd9;
@%p1 bra BB66_16;
cvta.to.global.u64 %rd12, %rd8;
mul.wide.s32 %rd13, %r1, 4;
add.s64 %rd14, %rd12, %rd13;
ld.global.f32 %f30, [%rd14];
abs.f32 %f10, %f30;
setp.neu.f32 %p2, %f10, 0f7F800000;
@%p2 bra BB66_3;
mov.f32 %f11, 0f00000000;
mul.rn.f32 %f30, %f30, %f11;
BB66_3:
mul.f32 %f12, %f30, 0f3F22F983;
cvt.rni.s32.f32 %r94, %f12;
cvt.rn.f32.s32 %f13, %r94;
neg.f32 %f14, %f13;
mov.f32 %f15, 0f3FC90FDA;
fma.rn.f32 %f16, %f14, %f15, %f30;
mov.f32 %f17, 0f33A22168;
fma.rn.f32 %f18, %f14, %f17, %f16;
mov.f32 %f19, 0f27C234C5;
fma.rn.f32 %f31, %f14, %f19, %f18;
abs.f32 %f20, %f30;
setp.leu.f32 %p3, %f20, 0f47CE4780;
@%p3 bra BB66_13;
mov.b32 %r3, %f30;
shr.u32 %r4, %r3, 23;
bfe.u32 %r42, %r3, 23, 8;
add.s32 %r43, %r42, -128;
shl.b32 %r44, %r3, 8;
or.b32 %r5, %r44, -2147483648;
shr.u32 %r6, %r43, 5;
mov.u32 %r86, 0;
mov.u64 %rd20, __cudart_i2opi_f;
mov.u32 %r85, -6;
mov.u64 %rd21, %rd1;
BB66_5:
.pragma "nounroll";
mov.u64 %rd3, %rd21;
ld.const.u32 %r47, [%rd20];
// inline asm
{
mad.lo.cc.u32 %r45, %r47, %r5, %r86;
madc.hi.u32 %r86, %r47, %r5, 0;
}
// inline asm
st.local.u32 [%rd3], %r45;
add.s64 %rd4, %rd3, 4;
add.s64 %rd20, %rd20, 4;
add.s32 %r85, %r85, 1;
setp.ne.s32 %p4, %r85, 0;
mov.u64 %rd21, %rd4;
@%p4 bra BB66_5;
and.b32 %r11, %r3, -2147483648;
st.local.u32 [%rd1+24], %r86;
mov.u32 %r50, 6;
sub.s32 %r51, %r50, %r6;
mul.wide.s32 %rd16, %r51, 4;
add.s64 %rd6, %rd1, %rd16;
ld.local.u32 %r87, [%rd6];
ld.local.u32 %r88, [%rd6+-4];
and.b32 %r14, %r4, 31;
setp.eq.s32 %p5, %r14, 0;
@%p5 bra BB66_8;
mov.u32 %r52, 32;
sub.s32 %r53, %r52, %r14;
shr.u32 %r54, %r88, %r53;
shl.b32 %r55, %r87, %r14;
add.s32 %r87, %r54, %r55;
ld.local.u32 %r56, [%rd6+-8];
shr.u32 %r57, %r56, %r53;
shl.b32 %r58, %r88, %r14;
add.s32 %r88, %r57, %r58;
BB66_8:
shr.u32 %r59, %r88, 30;
shl.b32 %r60, %r87, 2;
add.s32 %r89, %r59, %r60;
shl.b32 %r20, %r88, 2;
shr.u32 %r61, %r89, 31;
shr.u32 %r62, %r87, 30;
add.s32 %r21, %r61, %r62;
setp.eq.s32 %p6, %r61, 0;
mov.u32 %r90, %r11;
mov.u32 %r91, %r20;
@%p6 bra BB66_10;
not.b32 %r63, %r89;
neg.s32 %r22, %r20;
setp.eq.s32 %p7, %r20, 0;
selp.u32 %r64, 1, 0, %p7;
add.s32 %r89, %r64, %r63;
xor.b32 %r24, %r11, -2147483648;
mov.u32 %r90, %r24;
mov.u32 %r91, %r22;
BB66_10:
mov.u32 %r26, %r90;
neg.s32 %r65, %r21;
setp.eq.s32 %p8, %r11, 0;
selp.b32 %r94, %r21, %r65, %p8;
clz.b32 %r93, %r89;
setp.eq.s32 %p9, %r93, 0;
shl.b32 %r66, %r89, %r93;
mov.u32 %r67, 32;
sub.s32 %r68, %r67, %r93;
shr.u32 %r69, %r91, %r68;
add.s32 %r70, %r69, %r66;
selp.b32 %r30, %r89, %r70, %p9;
mov.u32 %r71, -921707870;
mul.hi.u32 %r92, %r30, %r71;
setp.lt.s32 %p10, %r92, 1;
@%p10 bra BB66_12;
mul.lo.s32 %r72, %r30, -921707870;
shr.u32 %r73, %r72, 31;
shl.b32 %r74, %r92, 1;
add.s32 %r92, %r73, %r74;
add.s32 %r93, %r93, 1;
BB66_12:
mov.u32 %r75, 126;
sub.s32 %r76, %r75, %r93;
shl.b32 %r77, %r76, 23;
add.s32 %r78, %r92, 1;
shr.u32 %r79, %r78, 7;
add.s32 %r80, %r79, 1;
shr.u32 %r81, %r80, 1;
add.s32 %r82, %r81, %r77;
or.b32 %r83, %r82, %r26;
mov.b32 %f31, %r83;
BB66_13:
mul.f32 %f21, %f31, %f31;
mov.f32 %f22, 0fBF52B7F4;
mov.f32 %f23, 0f3B86D46D;
fma.rn.f32 %f24, %f23, %f21, %f22;
add.f32 %f25, %f21, 0fC01E09D0;
rcp.rn.f32 %f26, %f25;
mul.f32 %f27, %f24, %f26;
mul.f32 %f28, %f21, %f27;
fma.rn.f32 %f32, %f28, %f31, %f31;
and.b32 %r84, %r94, 1;
setp.eq.b32 %p11, %r84, 1;
@!%p11 bra BB66_15;
bra.uni BB66_14;
BB66_14:
mov.f32 %f29, 0fBF800000;
div.rn.f32 %f32, %f29, %f32;
BB66_15:
cvta.to.global.u64 %rd17, %rd7;
add.s64 %rd19, %rd17, %rd13;
st.global.f32 [%rd19], %f32;
BB66_16:
ret;
}
// .globl vec_tanhf
.visible .entry vec_tanhf(
.param .u64 vec_tanhf_param_0,
.param .u64 vec_tanhf_param_1,
.param .u64 vec_tanhf_param_2
)
{
.reg .pred %p<5>;
.reg .f32 %f<33>;
.reg .b32 %r<10>;
.reg .b64 %rd<11>;
ld.param.u64 %rd4, [vec_tanhf_param_0];
ld.param.u64 %rd2, [vec_tanhf_param_1];
ld.param.u64 %rd3, [vec_tanhf_param_2];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB67_5;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f1, [%rd7];
abs.f32 %f2, %f1;
setp.ltu.f32 %p2, %f2, 0f3F0CCCCD;
@%p2 bra BB67_3;
bra.uni BB67_2;
BB67_3:
mul.f32 %f21, %f1, %f1;
mov.f32 %f22, 0fBD57BE66;
mov.f32 %f23, 0f3C86A81B;
fma.rn.f32 %f24, %f23, %f21, %f22;
mov.f32 %f25, 0f3E08677B;
fma.rn.f32 %f26, %f24, %f21, %f25;
mov.f32 %f27, 0fBEAAAA29;
fma.rn.f32 %f28, %f26, %f21, %f27;
mul.f32 %f29, %f21, %f28;
fma.rn.f32 %f30, %f29, %f1, %f1;
add.f32 %f31, %f1, %f1;
setp.eq.f32 %p4, %f1, 0f00000000;
selp.f32 %f32, %f31, %f30, %p4;
bra.uni BB67_4;
BB67_2:
add.f32 %f10, %f2, %f2;
mul.f32 %f11, %f10, 0f3FB8AA3B;
cvt.rzi.f32.f32 %f12, %f11;
mov.f32 %f13, 0fBF317200;
fma.rn.f32 %f14, %f12, %f13, %f10;
mov.f32 %f15, 0fB5BFBE8E;
fma.rn.f32 %f16, %f12, %f15, %f14;
mul.f32 %f7, %f16, 0f3FB8AA3B;
// inline asm
ex2.approx.ftz.f32 %f6,%f7;
// inline asm
ex2.approx.f32 %f17, %f12;
mov.f32 %f18, 0f3F800000;
fma.rn.f32 %f9, %f6, %f17, %f18;
// inline asm
rcp.approx.ftz.f32 %f8,%f9;
// inline asm
mov.f32 %f19, 0fC0000000;
fma.rn.f32 %f20, %f8, %f19, %f18;
mov.b32 %r5, %f20;
setp.ltu.f32 %p3, %f2, 0f42B00000;
selp.b32 %r6, %r5, 1065353216, %p3;
mov.b32 %r7, %f1;
and.b32 %r8, %r7, -2147483648;
or.b32 %r9, %r6, %r8;
mov.b32 %f32, %r9;
BB67_4:
cvta.to.global.u64 %rd8, %rd2;
add.s64 %rd10, %rd8, %rd6;
st.global.f32 [%rd10], %f32;
BB67_5:
ret;
}
// .globl vec_tgammaf
.visible .entry vec_tgammaf(
.param .u64 vec_tgammaf_param_0,
.param .u64 vec_tgammaf_param_1,
.param .u64 vec_tgammaf_param_2
)
{
.reg .pred %p<19>;
.reg .f32 %f<101>;
.reg .b32 %r<7>;
.reg .b64 %rd<11>;
ld.param.u64 %rd4, [vec_tgammaf_param_0];
ld.param.u64 %rd2, [vec_tgammaf_param_1];
ld.param.u64 %rd3, [vec_tgammaf_param_2];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB68_13;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f1, [%rd7];
setp.ltu.f32 %p2, %f1, 0f00000000;
@%p2 bra BB68_6;
bra.uni BB68_2;
BB68_6:
cvt.rmi.f32.f32 %f50, %f1;
setp.eq.f32 %p10, %f50, %f1;
selp.f32 %f51, 0f7FFFFFFF, %f1, %p10;
setp.lt.f32 %p11, %f51, 0fC2246666;
selp.f32 %f14, 0fC2246666, %f51, %p11;
setp.lt.f32 %p12, %f14, 0fC2081EB8;
add.f32 %f52, %f14, 0f40C00000;
selp.f32 %f97, %f52, %f14, %p12;
setp.geu.f32 %p13, %f97, 0fBF000000;
mov.f32 %f96, %f97;
@%p13 bra BB68_9;
mov.f32 %f98, %f97;
mov.f32 %f99, %f97;
BB68_8:
add.f32 %f98, %f98, 0f3F800000;
mul.f32 %f99, %f99, %f98;
setp.lt.f32 %p14, %f98, 0fBF000000;
mov.f32 %f97, %f99;
mov.f32 %f96, %f98;
@%p14 bra BB68_8;
BB68_9:
mov.f32 %f53, 0f3BE86AA4;
mov.f32 %f54, 0fBA8AA19E;
fma.rn.f32 %f55, %f54, %f96, %f53;
mov.f32 %f56, 0fBC1E2998;
fma.rn.f32 %f57, %f55, %f96, %f56;
mov.f32 %f58, 0fBD2CBE4A;
fma.rn.f32 %f59, %f57, %f96, %f58;
mov.f32 %f60, 0f3E2A8A17;
fma.rn.f32 %f61, %f59, %f96, %f60;
mov.f32 %f62, 0fBD2C0CBB;
fma.rn.f32 %f63, %f61, %f96, %f62;
mov.f32 %f64, 0fBF27E7A3;
fma.rn.f32 %f65, %f63, %f96, %f64;
mov.f32 %f66, 0f3F13C468;
fma.rn.f32 %f67, %f65, %f96, %f66;
mov.f32 %f68, 0f3F800000;
fma.rn.f32 %f69, %f67, %f96, %f68;
mul.f32 %f70, %f97, %f69;
rcp.rn.f32 %f100, %f70;
setp.geu.f32 %p15, %f14, 0fC2081EB8;
@%p15 bra BB68_12;
add.f32 %f71, %f14, 0f3F800000;
mul.f32 %f72, %f14, %f71;
add.f32 %f73, %f14, 0f40000000;
mul.f32 %f74, %f73, %f72;
add.f32 %f75, %f14, 0f40400000;
mul.f32 %f76, %f75, %f74;
add.f32 %f77, %f14, 0f40800000;
mul.f32 %f78, %f77, %f76;
add.f32 %f79, %f14, 0f40A00000;
mul.f32 %f80, %f79, %f78;
rcp.rn.f32 %f81, %f80;
mul.f32 %f100, %f100, %f81;
setp.geu.f32 %p16, %f1, 0fC2280000;
@%p16 bra BB68_12;
cvt.rzi.s32.f32 %r5, %f1;
and.b32 %r6, %r5, 1;
setp.eq.b32 %p17, %r6, 1;
not.pred %p18, %p17;
selp.f32 %f100, 0f80000000, %f100, %p18;
bra.uni BB68_12;
BB68_2:
setp.gt.f32 %p3, %f1, 0f42100000;
selp.f32 %f2, 0f42100000, %f1, %p3;
setp.gt.f32 %p4, %f2, 0f42081EB8;
add.f32 %f3, %f2, 0fBF800000;
selp.f32 %f88, %f3, %f2, %p4;
add.f32 %f89, %f88, 0fBF800000;
mov.f32 %f27, 0f3F800000;
mov.f32 %f84, %f27;
setp.leu.f32 %p5, %f88, 0f3FC00000;
mov.f32 %f83, %f27;
@%p5 bra BB68_5;
mov.f32 %f90, %f89;
BB68_4:
mov.f32 %f88, %f90;
mul.f32 %f84, %f88, %f84;
add.f32 %f90, %f88, 0fBF800000;
setp.gt.f32 %p6, %f88, 0f3FC00000;
mov.f32 %f82, %f84;
mov.f32 %f83, %f82;
mov.f32 %f89, %f90;
@%p6 bra BB68_4;
BB68_5:
mov.f32 %f11, %f83;
setp.ltu.f32 %p7, %f2, 0f3F000000;
selp.f32 %f28, %f88, %f89, %p7;
mov.f32 %f29, 0f3BE86AA4;
mov.f32 %f30, 0fBA8AA19E;
fma.rn.f32 %f31, %f30, %f28, %f29;
mov.f32 %f32, 0fBC1E2998;
fma.rn.f32 %f33, %f31, %f28, %f32;
mov.f32 %f34, 0fBD2CBE4A;
fma.rn.f32 %f35, %f33, %f28, %f34;
mov.f32 %f36, 0f3E2A8A17;
fma.rn.f32 %f37, %f35, %f28, %f36;
mov.f32 %f38, 0fBD2C0CBB;
fma.rn.f32 %f39, %f37, %f28, %f38;
mov.f32 %f40, 0fBF27E7A3;
fma.rn.f32 %f41, %f39, %f28, %f40;
mov.f32 %f42, 0f3F13C468;
fma.rn.f32 %f43, %f41, %f28, %f42;
fma.rn.f32 %f45, %f43, %f28, %f27;
mul.f32 %f46, %f2, %f45;
setp.lt.f32 %p8, %f2, 0f3F000000;
selp.f32 %f47, %f46, %f45, %p8;
div.approx.f32 %f48, %f11, %f47;
mul.f32 %f49, %f3, %f48;
selp.f32 %f100, %f49, %f48, %p4;
BB68_12:
cvta.to.global.u64 %rd8, %rd2;
add.s64 %rd10, %rd8, %rd6;
st.global.f32 [%rd10], %f100;
BB68_13:
ret;
}
// .globl vec_truncf
.visible .entry vec_truncf(
.param .u64 vec_truncf_param_0,
.param .u64 vec_truncf_param_1,
.param .u64 vec_truncf_param_2
)
{
.reg .pred %p<2>;
.reg .f32 %f<3>;
.reg .b32 %r<5>;
.reg .b64 %rd<10>;
ld.param.u64 %rd4, [vec_truncf_param_0];
ld.param.u64 %rd2, [vec_truncf_param_1];
ld.param.u64 %rd3, [vec_truncf_param_2];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd4;
@%p1 bra BB69_2;
cvta.to.global.u64 %rd5, %rd3;
shl.b64 %rd6, %rd1, 2;
add.s64 %rd7, %rd5, %rd6;
ld.global.f32 %f1, [%rd7];
cvt.rzi.f32.f32 %f2, %f1;
cvta.to.global.u64 %rd8, %rd2;
add.s64 %rd9, %rd8, %rd6;
st.global.f32 [%rd9], %f2;
BB69_2:
ret;
}
// .globl vec_y0f
.visible .entry vec_y0f(
.param .u64 vec_y0f_param_0,
.param .u64 vec_y0f_param_1,
.param .u64 vec_y0f_param_2
)
{
.local .align 4 .b8 __local_depot70[28];
.reg .b64 %SP;
.reg .b64 %SPL;
.reg .pred %p<57>;
.reg .f32 %f<336>;
.reg .b32 %r<375>;
.reg .b64 %rd<61>;
mov.u64 %rd60, __local_depot70;
cvta.local.u64 %SP, %rd60;
ld.param.u64 %rd26, [vec_y0f_param_0];
ld.param.u64 %rd24, [vec_y0f_param_1];
ld.param.u64 %rd25, [vec_y0f_param_2];
add.u64 %rd27, %SP, 0;
cvta.to.local.u64 %rd57, %rd27;
mov.u32 %r143, %ntid.x;
mov.u32 %r144, %ctaid.x;
mov.u32 %r145, %tid.x;
mad.lo.s32 %r1, %r143, %r144, %r145;
cvt.s64.s32 %rd28, %r1;
setp.ge.u64 %p1, %rd28, %rd26;
@%p1 bra BB70_85;
cvta.to.global.u64 %rd29, %rd25;
mul.wide.s32 %rd30, %r1, 4;
add.s64 %rd31, %rd29, %rd30;
ld.global.f32 %f1, [%rd31];
abs.f32 %f2, %f1;
setp.gtu.f32 %p2, %f2, 0f3EE4C176;
@%p2 bra BB70_42;
bra.uni BB70_2;
BB70_42:
setp.gtu.f32 %p31, %f2, 0f3FF67AF8;
@%p31 bra BB70_44;
bra.uni BB70_43;
BB70_44:
setp.gtu.f32 %p32, %f2, 0f40B0B31E;
@%p32 bra BB70_46;
bra.uni BB70_45;
BB70_46:
setp.gtu.f32 %p33, %f2, 0f410A7798;
@%p33 bra BB70_48;
bra.uni BB70_47;
BB70_48:
abs.f32 %f265, %f2;
mov.f32 %f335, 0f00000000;
setp.eq.f32 %p34, %f265, 0f7F800000;
@%p34 bra BB70_82;
// inline asm
rcp.approx.ftz.f32 %f266,%f2;
// inline asm
mul.f32 %f268, %f266, %f266;
mov.f32 %f269, 0f3DD0D5F0;
mov.f32 %f270, 0fBECC69F3;
fma.rn.f32 %f271, %f270, %f268, %f269;
mov.f32 %f272, 0fBD7FF855;
fma.rn.f32 %f273, %f271, %f268, %f272;
mov.f32 %f274, 0f3F800000;
fma.rn.f32 %f275, %f273, %f268, %f274;
mov.f32 %f276, 0fBE50D31C;
mov.f32 %f277, 0f3F8CCD61;
fma.rn.f32 %f278, %f277, %f268, %f276;
mov.f32 %f279, 0f3D854783;
fma.rn.f32 %f280, %f278, %f268, %f279;
mov.f32 %f281, 0fBDFFFFFB;
fma.rn.f32 %f282, %f280, %f268, %f281;
fma.rn.f32 %f38, %f282, %f266, %f2;
rsqrt.approx.f32 %f283, %f2;
mul.f32 %f284, %f283, 0f3F4C422A;
mul.f32 %f39, %f275, %f284;
mul.f32 %f285, %f38, 0f3F22F983;
cvt.rni.s32.f32 %r364, %f285;
cvt.rn.f32.s32 %f286, %r364;
neg.f32 %f287, %f286;
mov.f32 %f288, 0f3FC90FDA;
fma.rn.f32 %f289, %f287, %f288, %f38;
mov.f32 %f290, 0f33A22168;
fma.rn.f32 %f291, %f287, %f290, %f289;
mov.f32 %f292, 0f27C234C5;
fma.rn.f32 %f329, %f287, %f292, %f291;
abs.f32 %f293, %f38;
setp.leu.f32 %p35, %f293, 0f47CE4780;
@%p35 bra BB70_59;
mov.b32 %r73, %f38;
bfe.u32 %r243, %r73, 23, 8;
add.s32 %r244, %r243, -128;
shl.b32 %r245, %r73, 8;
or.b32 %r74, %r245, -2147483648;
shr.u32 %r75, %r244, 5;
mov.u32 %r356, 0;
mov.u64 %rd56, __cudart_i2opi_f;
mov.u32 %r355, -6;
BB70_51:
.pragma "nounroll";
ld.const.u32 %r248, [%rd56];
// inline asm
{
mad.lo.cc.u32 %r246, %r248, %r74, %r356;
madc.hi.u32 %r356, %r248, %r74, 0;
}
// inline asm
st.local.u32 [%rd57], %r246;
add.s64 %rd57, %rd57, 4;
add.s64 %rd56, %rd56, 4;
add.s32 %r355, %r355, 1;
setp.ne.s32 %p36, %r355, 0;
@%p36 bra BB70_51;
and.b32 %r80, %r73, -2147483648;
cvta.to.local.u64 %rd43, %rd27;
st.local.u32 [%rd43+24], %r356;
bfe.u32 %r81, %r73, 23, 5;
mov.u32 %r251, 6;
sub.s32 %r252, %r251, %r75;
mul.wide.s32 %rd44, %r252, 4;
add.s64 %rd17, %rd43, %rd44;
ld.local.u32 %r357, [%rd17];
ld.local.u32 %r358, [%rd17+-4];
setp.eq.s32 %p37, %r81, 0;
@%p37 bra BB70_54;
mov.u32 %r253, 32;
sub.s32 %r254, %r253, %r81;
shr.u32 %r255, %r358, %r254;
shl.b32 %r256, %r357, %r81;
add.s32 %r357, %r255, %r256;
ld.local.u32 %r257, [%rd17+-8];
shr.u32 %r258, %r257, %r254;
shl.b32 %r259, %r358, %r81;
add.s32 %r358, %r258, %r259;
BB70_54:
shr.u32 %r260, %r358, 30;
shl.b32 %r261, %r357, 2;
add.s32 %r359, %r260, %r261;
shl.b32 %r89, %r358, 2;
shr.u32 %r262, %r359, 31;
shr.u32 %r263, %r357, 30;
add.s32 %r90, %r262, %r263;
setp.eq.s32 %p38, %r262, 0;
mov.u32 %r360, %r80;
mov.u32 %r361, %r89;
@%p38 bra BB70_56;
not.b32 %r264, %r359;
neg.s32 %r91, %r89;
setp.eq.s32 %p39, %r89, 0;
selp.u32 %r265, 1, 0, %p39;
add.s32 %r359, %r265, %r264;
xor.b32 %r93, %r80, -2147483648;
mov.u32 %r360, %r93;
mov.u32 %r361, %r91;
BB70_56:
mov.u32 %r95, %r360;
neg.s32 %r266, %r90;
setp.eq.s32 %p40, %r80, 0;
selp.b32 %r364, %r90, %r266, %p40;
clz.b32 %r363, %r359;
setp.eq.s32 %p41, %r363, 0;
shl.b32 %r267, %r359, %r363;
mov.u32 %r268, 32;
sub.s32 %r269, %r268, %r363;
shr.u32 %r270, %r361, %r269;
add.s32 %r271, %r270, %r267;
selp.b32 %r99, %r359, %r271, %p41;
mov.u32 %r272, -921707870;
mul.hi.u32 %r362, %r99, %r272;
setp.lt.s32 %p42, %r362, 1;
@%p42 bra BB70_58;
mul.lo.s32 %r273, %r99, -921707870;
shr.u32 %r274, %r273, 31;
shl.b32 %r275, %r362, 1;
add.s32 %r362, %r274, %r275;
add.s32 %r363, %r363, 1;
BB70_58:
mov.u32 %r276, 126;
sub.s32 %r277, %r276, %r363;
shl.b32 %r278, %r277, 23;
add.s32 %r279, %r362, 1;
shr.u32 %r280, %r279, 7;
add.s32 %r281, %r280, 1;
shr.u32 %r282, %r281, 1;
add.s32 %r283, %r282, %r278;
or.b32 %r284, %r283, %r95;
mov.b32 %f329, %r284;
BB70_59:
and.b32 %r285, %r364, 3;
cvt.rn.f32.s32 %f294, %r285;
add.f32 %f295, %f329, 0fC016CBE4;
fma.rn.f32 %f330, %f294, 0f3FC90FDB, %f295;
abs.f32 %f296, %f330;
setp.neu.f32 %p43, %f296, 0f7F800000;
@%p43 bra BB70_61;
mov.f32 %f297, 0f00000000;
mul.rn.f32 %f330, %f330, %f297;
BB70_61:
mul.f32 %f298, %f330, 0f3F22F983;
cvt.rni.s32.f32 %r374, %f298;
cvt.rn.f32.s32 %f299, %r374;
neg.f32 %f300, %f299;
fma.rn.f32 %f302, %f300, %f288, %f330;
fma.rn.f32 %f304, %f300, %f290, %f302;
fma.rn.f32 %f331, %f300, %f292, %f304;
abs.f32 %f306, %f330;
setp.leu.f32 %p44, %f306, 0f47CE4780;
@%p44 bra BB70_71;
mov.b32 %r107, %f330;
shr.u32 %r108, %r107, 23;
bfe.u32 %r288, %r107, 23, 8;
add.s32 %r289, %r288, -128;
shl.b32 %r290, %r107, 8;
or.b32 %r109, %r290, -2147483648;
shr.u32 %r110, %r289, 5;
cvta.to.local.u64 %rd59, %rd27;
mov.u32 %r366, 0;
mov.u64 %rd58, __cudart_i2opi_f;
mov.u32 %r365, -6;
BB70_63:
.pragma "nounroll";
ld.const.u32 %r293, [%rd58];
// inline asm
{
mad.lo.cc.u32 %r291, %r293, %r109, %r366;
madc.hi.u32 %r366, %r293, %r109, 0;
}
// inline asm
st.local.u32 [%rd59], %r291;
add.s64 %rd59, %rd59, 4;
add.s64 %rd58, %rd58, 4;
add.s32 %r365, %r365, 1;
setp.ne.s32 %p45, %r365, 0;
@%p45 bra BB70_63;
and.b32 %r115, %r107, -2147483648;
cvta.to.local.u64 %rd48, %rd27;
st.local.u32 [%rd48+24], %r366;
mov.u32 %r296, 6;
sub.s32 %r297, %r296, %r110;
mul.wide.s32 %rd49, %r297, 4;
add.s64 %rd23, %rd48, %rd49;
ld.local.u32 %r367, [%rd23];
ld.local.u32 %r368, [%rd23+-4];
and.b32 %r118, %r108, 31;
setp.eq.s32 %p46, %r118, 0;
@%p46 bra BB70_66;
mov.u32 %r298, 32;
sub.s32 %r299, %r298, %r118;
shr.u32 %r300, %r368, %r299;
shl.b32 %r301, %r367, %r118;
add.s32 %r367, %r300, %r301;
ld.local.u32 %r302, [%rd23+-8];
shr.u32 %r303, %r302, %r299;
shl.b32 %r304, %r368, %r118;
add.s32 %r368, %r303, %r304;
BB70_66:
shr.u32 %r305, %r368, 30;
shl.b32 %r306, %r367, 2;
add.s32 %r369, %r305, %r306;
shl.b32 %r124, %r368, 2;
shr.u32 %r307, %r369, 31;
shr.u32 %r308, %r367, 30;
add.s32 %r125, %r307, %r308;
setp.eq.s32 %p47, %r307, 0;
mov.u32 %r370, %r115;
mov.u32 %r371, %r124;
@%p47 bra BB70_68;
not.b32 %r309, %r369;
neg.s32 %r126, %r124;
setp.eq.s32 %p48, %r124, 0;
selp.u32 %r310, 1, 0, %p48;
add.s32 %r369, %r310, %r309;
xor.b32 %r128, %r115, -2147483648;
mov.u32 %r370, %r128;
mov.u32 %r371, %r126;
BB70_68:
mov.u32 %r130, %r370;
neg.s32 %r311, %r125;
setp.eq.s32 %p49, %r115, 0;
selp.b32 %r374, %r125, %r311, %p49;
clz.b32 %r373, %r369;
setp.eq.s32 %p50, %r373, 0;
shl.b32 %r312, %r369, %r373;
mov.u32 %r313, 32;
sub.s32 %r314, %r313, %r373;
shr.u32 %r315, %r371, %r314;
add.s32 %r316, %r315, %r312;
selp.b32 %r134, %r369, %r316, %p50;
mov.u32 %r317, -921707870;
mul.hi.u32 %r372, %r134, %r317;
setp.lt.s32 %p51, %r372, 1;
@%p51 bra BB70_70;
mul.lo.s32 %r318, %r134, -921707870;
shr.u32 %r319, %r318, 31;
shl.b32 %r320, %r372, 1;
add.s32 %r372, %r319, %r320;
add.s32 %r373, %r373, 1;
BB70_70:
mov.u32 %r321, 126;
sub.s32 %r322, %r321, %r373;
shl.b32 %r323, %r322, 23;
add.s32 %r324, %r372, 1;
shr.u32 %r325, %r324, 7;
add.s32 %r326, %r325, 1;
shr.u32 %r327, %r326, 1;
add.s32 %r328, %r327, %r323;
or.b32 %r329, %r328, %r130;
mov.b32 %f331, %r329;
BB70_71:
mul.rn.f32 %f49, %f331, %f331;
add.s32 %r141, %r374, 1;
and.b32 %r142, %r141, 1;
setp.eq.s32 %p52, %r142, 0;
@%p52 bra BB70_73;
mov.f32 %f307, 0fBAB6061A;
mov.f32 %f308, 0f37CCF5CE;
fma.rn.f32 %f332, %f308, %f49, %f307;
bra.uni BB70_74;
BB70_2:
mul.f32 %f65, %f2, %f2;
mov.f32 %f66, 0fB71F49B6;
mov.f32 %f67, 0f33DBE5AC;
fma.rn.f32 %f68, %f67, %f65, %f66;
mov.f32 %f69, 0f3A0D3100;
fma.rn.f32 %f70, %f68, %f65, %f69;
mov.f32 %f71, 0fBC83AD8E;
fma.rn.f32 %f72, %f70, %f65, %f71;
mov.f32 %f73, 0f3E35DE5A;
fma.rn.f32 %f74, %f72, %f65, %f73;
mov.f32 %f75, 0fBD9726B5;
fma.rn.f32 %f3, %f74, %f65, %f75;
setp.lt.f32 %p3, %f2, 0f7F800000;
setp.gt.f32 %p4, %f2, 0f00000000;
and.pred %p5, %p4, %p3;
@%p5 bra BB70_4;
bra.uni BB70_3;
BB70_4:
setp.lt.f32 %p6, %f2, 0f00800000;
mul.f32 %f78, %f2, 0f4B800000;
selp.f32 %f79, %f78, %f2, %p6;
selp.f32 %f80, 0fC3170000, 0fC2FE0000, %p6;
mov.b32 %r146, %f79;
and.b32 %r147, %r146, 8388607;
or.b32 %r148, %r147, 1065353216;
mov.b32 %f81, %r148;
shr.u32 %r149, %r146, 23;
cvt.rn.f32.u32 %f82, %r149;
add.f32 %f83, %f80, %f82;
setp.gt.f32 %p7, %f81, 0f3FAE147B;
mul.f32 %f84, %f81, 0f3F000000;
add.f32 %f85, %f83, 0f3F800000;
selp.f32 %f86, %f84, %f81, %p7;
selp.f32 %f87, %f85, %f83, %p7;
add.f32 %f77, %f86, 0f3F800000;
add.f32 %f88, %f86, 0fBF800000;
// inline asm
rcp.approx.ftz.f32 %f76,%f77;
// inline asm
mul.f32 %f89, %f88, %f88;
neg.f32 %f90, %f89;
mul.rn.f32 %f91, %f76, %f90;
add.rn.f32 %f92, %f88, %f91;
mul.f32 %f93, %f92, %f92;
mov.f32 %f94, 0f3C4C6A36;
mov.f32 %f95, 0f3B1E94E6;
fma.rn.f32 %f96, %f95, %f93, %f94;
mov.f32 %f97, 0f3DAAAB1A;
fma.rn.f32 %f98, %f96, %f93, %f97;
mul.f32 %f99, %f93, %f98;
fma.rn.f32 %f100, %f99, %f92, %f91;
add.f32 %f101, %f88, %f100;
mov.f32 %f102, 0f3F317218;
fma.rn.f32 %f321, %f87, %f102, %f101;
bra.uni BB70_5;
BB70_43:
add.f32 %f197, %f2, 0fBF64C176;
add.f32 %f198, %f197, 0f32657D03;
mov.f32 %f199, 0fBE02574C;
mov.f32 %f200, 0f3CDDC8B3;
fma.rn.f32 %f201, %f200, %f198, %f199;
mov.f32 %f202, 0f3E7F2CC9;
fma.rn.f32 %f203, %f201, %f198, %f202;
mov.f32 %f204, 0fBE8BF29B;
fma.rn.f32 %f205, %f203, %f198, %f204;
mov.f32 %f206, 0f3E5BCE93;
fma.rn.f32 %f207, %f205, %f198, %f206;
mov.f32 %f208, 0fBE38C4FF;
fma.rn.f32 %f209, %f207, %f198, %f208;
mov.f32 %f210, 0f3E42774D;
fma.rn.f32 %f211, %f209, %f198, %f210;
mov.f32 %f212, 0fBE525CB2;
fma.rn.f32 %f213, %f211, %f198, %f212;
mov.f32 %f214, 0f3E60F43D;
fma.rn.f32 %f215, %f213, %f198, %f214;
mov.f32 %f216, 0fBE679145;
fma.rn.f32 %f217, %f215, %f198, %f216;
mov.f32 %f218, 0f3E61D24A;
fma.rn.f32 %f219, %f217, %f198, %f218;
mov.f32 %f220, 0fBEFBF1AD;
fma.rn.f32 %f221, %f219, %f198, %f220;
mov.f32 %f222, 0f3F6121BB;
fma.rn.f32 %f223, %f221, %f198, %f222;
mul.f32 %f335, %f198, %f223;
bra.uni BB70_82;
BB70_3:
lg2.approx.f32 %f321, %f2;
BB70_5:
abs.f32 %f7, %f2;
setp.gtu.f32 %p8, %f7, 0f41000000;
@%p8 bra BB70_7;
bra.uni BB70_6;
BB70_7:
abs.f32 %f139, %f7;
mov.f32 %f328, 0f00000000;
setp.eq.f32 %p9, %f139, 0f7F800000;
@%p9 bra BB70_41;
// inline asm
rcp.approx.ftz.f32 %f140,%f7;
// inline asm
mul.f32 %f142, %f140, %f140;
mov.f32 %f143, 0fBF03B7C2;
mov.f32 %f144, 0f4056FE93;
fma.rn.f32 %f145, %f144, %f142, %f143;
mov.f32 %f146, 0f3DD3B3F3;
fma.rn.f32 %f147, %f145, %f142, %f146;
mov.f32 %f148, 0fBD7FFFB6;
fma.rn.f32 %f149, %f147, %f142, %f148;
mov.f32 %f150, 0f3F800000;
fma.rn.f32 %f151, %f149, %f142, %f150;
mov.f32 %f152, 0fBE52412D;
mov.f32 %f153, 0f3F91E009;
fma.rn.f32 %f154, %f153, %f142, %f152;
mov.f32 %f155, 0f3D854ED1;
fma.rn.f32 %f156, %f154, %f142, %f155;
mov.f32 %f157, 0fBDFFFFFF;
fma.rn.f32 %f158, %f156, %f142, %f157;
fma.rn.f32 %f9, %f158, %f140, %f7;
rsqrt.approx.f32 %f159, %f7;
mul.f32 %f160, %f159, 0f3F4C422A;
mul.f32 %f10, %f151, %f160;
mul.f32 %f161, %f9, 0f3F22F983;
cvt.rni.s32.f32 %r344, %f161;
cvt.rn.f32.s32 %f162, %r344;
neg.f32 %f163, %f162;
mov.f32 %f164, 0f3FC90FDA;
fma.rn.f32 %f165, %f163, %f164, %f9;
mov.f32 %f166, 0f33A22168;
fma.rn.f32 %f167, %f163, %f166, %f165;
mov.f32 %f168, 0f27C234C5;
fma.rn.f32 %f322, %f163, %f168, %f167;
abs.f32 %f169, %f9;
setp.leu.f32 %p10, %f169, 0f47CE4780;
@%p10 bra BB70_18;
mov.b32 %r3, %f9;
shl.b32 %r152, %r3, 8;
or.b32 %r4, %r152, -2147483648;
mov.u32 %r336, 0;
mov.u64 %rd53, __cudart_i2opi_f;
mov.u32 %r335, -6;
BB70_10:
.pragma "nounroll";
ld.const.u32 %r155, [%rd53];
// inline asm
{
mad.lo.cc.u32 %r153, %r155, %r4, %r336;
madc.hi.u32 %r336, %r155, %r4, 0;
}
// inline asm
st.local.u32 [%rd57], %r153;
add.s64 %rd57, %rd57, 4;
add.s64 %rd53, %rd53, 4;
add.s32 %r335, %r335, 1;
setp.ne.s32 %p11, %r335, 0;
@%p11 bra BB70_10;
and.b32 %r9, %r3, -2147483648;
bfe.u32 %r158, %r3, 23, 8;
add.s32 %r159, %r158, -128;
shr.u32 %r160, %r159, 5;
cvta.to.local.u64 %rd34, %rd27;
st.local.u32 [%rd34+24], %r336;
bfe.u32 %r10, %r3, 23, 5;
mov.u32 %r161, 6;
sub.s32 %r162, %r161, %r160;
mul.wide.s32 %rd35, %r162, 4;
add.s64 %rd6, %rd34, %rd35;
ld.local.u32 %r337, [%rd6];
ld.local.u32 %r338, [%rd6+-4];
setp.eq.s32 %p12, %r10, 0;
@%p12 bra BB70_13;
mov.u32 %r163, 32;
sub.s32 %r164, %r163, %r10;
shr.u32 %r165, %r338, %r164;
shl.b32 %r166, %r337, %r10;
add.s32 %r337, %r165, %r166;
ld.local.u32 %r167, [%rd6+-8];
shr.u32 %r168, %r167, %r164;
shl.b32 %r169, %r338, %r10;
add.s32 %r338, %r168, %r169;
BB70_13:
shr.u32 %r170, %r338, 30;
shl.b32 %r171, %r337, 2;
add.s32 %r339, %r170, %r171;
shl.b32 %r18, %r338, 2;
shr.u32 %r172, %r339, 31;
shr.u32 %r173, %r337, 30;
add.s32 %r19, %r172, %r173;
setp.eq.s32 %p13, %r172, 0;
mov.u32 %r340, %r9;
mov.u32 %r341, %r18;
@%p13 bra BB70_15;
not.b32 %r174, %r339;
neg.s32 %r20, %r18;
setp.eq.s32 %p14, %r18, 0;
selp.u32 %r175, 1, 0, %p14;
add.s32 %r339, %r175, %r174;
xor.b32 %r22, %r9, -2147483648;
mov.u32 %r340, %r22;
mov.u32 %r341, %r20;
BB70_15:
mov.u32 %r24, %r340;
neg.s32 %r176, %r19;
setp.eq.s32 %p15, %r9, 0;
selp.b32 %r344, %r19, %r176, %p15;
clz.b32 %r343, %r339;
setp.eq.s32 %p16, %r343, 0;
shl.b32 %r177, %r339, %r343;
mov.u32 %r178, 32;
sub.s32 %r179, %r178, %r343;
shr.u32 %r180, %r341, %r179;
add.s32 %r181, %r180, %r177;
selp.b32 %r28, %r339, %r181, %p16;
mov.u32 %r182, -921707870;
mul.hi.u32 %r342, %r28, %r182;
setp.lt.s32 %p17, %r342, 1;
@%p17 bra BB70_17;
mul.lo.s32 %r183, %r28, -921707870;
shr.u32 %r184, %r183, 31;
shl.b32 %r185, %r342, 1;
add.s32 %r342, %r184, %r185;
add.s32 %r343, %r343, 1;
BB70_17:
mov.u32 %r186, 126;
sub.s32 %r187, %r186, %r343;
shl.b32 %r188, %r187, 23;
add.s32 %r189, %r342, 1;
shr.u32 %r190, %r189, 7;
add.s32 %r191, %r190, 1;
shr.u32 %r192, %r191, 1;
add.s32 %r193, %r192, %r188;
or.b32 %r194, %r193, %r24;
mov.b32 %f322, %r194;
BB70_18:
and.b32 %r195, %r344, 3;
cvt.rn.f32.s32 %f170, %r195;
add.f32 %f171, %f322, 0fBF490FDB;
fma.rn.f32 %f323, %f170, 0f3FC90FDB, %f171;
abs.f32 %f172, %f323;
setp.neu.f32 %p18, %f172, 0f7F800000;
@%p18 bra BB70_20;
mov.f32 %f173, 0f00000000;
mul.rn.f32 %f323, %f323, %f173;
BB70_20:
mul.f32 %f174, %f323, 0f3F22F983;
cvt.rni.s32.f32 %r354, %f174;
cvt.rn.f32.s32 %f175, %r354;
neg.f32 %f176, %f175;
fma.rn.f32 %f178, %f176, %f164, %f323;
fma.rn.f32 %f180, %f176, %f166, %f178;
fma.rn.f32 %f324, %f176, %f168, %f180;
abs.f32 %f182, %f323;
setp.leu.f32 %p19, %f182, 0f47CE4780;
@%p19 bra BB70_30;
mov.b32 %r36, %f323;
shr.u32 %r37, %r36, 23;
bfe.u32 %r198, %r36, 23, 8;
add.s32 %r199, %r198, -128;
shl.b32 %r200, %r36, 8;
or.b32 %r38, %r200, -2147483648;
shr.u32 %r39, %r199, 5;
cvta.to.local.u64 %rd55, %rd27;
mov.u32 %r346, 0;
mov.u64 %rd54, __cudart_i2opi_f;
mov.u32 %r345, -6;
BB70_22:
.pragma "nounroll";
ld.const.u32 %r203, [%rd54];
// inline asm
{
mad.lo.cc.u32 %r201, %r203, %r38, %r346;
madc.hi.u32 %r346, %r203, %r38, 0;
}
// inline asm
st.local.u32 [%rd55], %r201;
add.s64 %rd55, %rd55, 4;
add.s64 %rd54, %rd54, 4;
add.s32 %r345, %r345, 1;
setp.ne.s32 %p20, %r345, 0;
@%p20 bra BB70_22;
and.b32 %r44, %r36, -2147483648;
cvta.to.local.u64 %rd39, %rd27;
st.local.u32 [%rd39+24], %r346;
mov.u32 %r206, 6;
sub.s32 %r207, %r206, %r39;
mul.wide.s32 %rd40, %r207, 4;
add.s64 %rd12, %rd39, %rd40;
ld.local.u32 %r347, [%rd12];
ld.local.u32 %r348, [%rd12+-4];
and.b32 %r47, %r37, 31;
setp.eq.s32 %p21, %r47, 0;
@%p21 bra BB70_25;
mov.u32 %r208, 32;
sub.s32 %r209, %r208, %r47;
shr.u32 %r210, %r348, %r209;
shl.b32 %r211, %r347, %r47;
add.s32 %r347, %r210, %r211;
ld.local.u32 %r212, [%rd12+-8];
shr.u32 %r213, %r212, %r209;
shl.b32 %r214, %r348, %r47;
add.s32 %r348, %r213, %r214;
BB70_25:
shr.u32 %r215, %r348, 30;
shl.b32 %r216, %r347, 2;
add.s32 %r349, %r215, %r216;
shl.b32 %r53, %r348, 2;
shr.u32 %r217, %r349, 31;
shr.u32 %r218, %r347, 30;
add.s32 %r54, %r217, %r218;
setp.eq.s32 %p22, %r217, 0;
mov.u32 %r350, %r44;
mov.u32 %r351, %r53;
@%p22 bra BB70_27;
not.b32 %r219, %r349;
neg.s32 %r55, %r53;
setp.eq.s32 %p23, %r53, 0;
selp.u32 %r220, 1, 0, %p23;
add.s32 %r349, %r220, %r219;
xor.b32 %r57, %r44, -2147483648;
mov.u32 %r350, %r57;
mov.u32 %r351, %r55;
BB70_27:
mov.u32 %r59, %r350;
neg.s32 %r221, %r54;
setp.eq.s32 %p24, %r44, 0;
selp.b32 %r354, %r54, %r221, %p24;
clz.b32 %r353, %r349;
setp.eq.s32 %p25, %r353, 0;
shl.b32 %r222, %r349, %r353;
mov.u32 %r223, 32;
sub.s32 %r224, %r223, %r353;
shr.u32 %r225, %r351, %r224;
add.s32 %r226, %r225, %r222;
selp.b32 %r63, %r349, %r226, %p25;
mov.u32 %r227, -921707870;
mul.hi.u32 %r352, %r63, %r227;
setp.lt.s32 %p26, %r352, 1;
@%p26 bra BB70_29;
mul.lo.s32 %r228, %r63, -921707870;
shr.u32 %r229, %r228, 31;
shl.b32 %r230, %r352, 1;
add.s32 %r352, %r229, %r230;
add.s32 %r353, %r353, 1;
BB70_29:
mov.u32 %r231, 126;
sub.s32 %r232, %r231, %r353;
shl.b32 %r233, %r232, 23;
add.s32 %r234, %r352, 1;
shr.u32 %r235, %r234, 7;
add.s32 %r236, %r235, 1;
shr.u32 %r237, %r236, 1;
add.s32 %r238, %r237, %r233;
or.b32 %r239, %r238, %r59;
mov.b32 %f324, %r239;
BB70_30:
mul.rn.f32 %f20, %f324, %f324;
add.s32 %r70, %r354, 1;
and.b32 %r71, %r70, 1;
setp.eq.s32 %p27, %r71, 0;
@%p27 bra BB70_32;
mov.f32 %f183, 0fBAB6061A;
mov.f32 %f184, 0f37CCF5CE;
fma.rn.f32 %f325, %f184, %f20, %f183;
bra.uni BB70_33;
BB70_6:
add.f32 %f103, %f7, 0fC019E8A9;
add.f32 %f104, %f103, 0fB3E971B3;
mov.f32 %f105, 0fA9ACA9B3;
mov.f32 %f106, 0fA6B3B8E7;
fma.rn.f32 %f107, %f106, %f104, %f105;
mov.f32 %f108, 0f2C3F0E18;
fma.rn.f32 %f109, %f107, %f104, %f108;
mov.f32 %f110, 0fACD41781;
fma.rn.f32 %f111, %f109, %f104, %f110;
mov.f32 %f112, 0fAFE90F38;
fma.rn.f32 %f113, %f111, %f104, %f112;
mov.f32 %f114, 0f3020305B;
fma.rn.f32 %f115, %f113, %f104, %f114;
mov.f32 %f116, 0f33797143;
fma.rn.f32 %f117, %f115, %f104, %f116;
mov.f32 %f118, 0f30F76F85;
fma.rn.f32 %f119, %f117, %f104, %f118;
mov.f32 %f120, 0fB6B6DFC6;
fma.rn.f32 %f121, %f119, %f104, %f120;
mov.f32 %f122, 0fB6F665C9;
fma.rn.f32 %f123, %f121, %f104, %f122;
mov.f32 %f124, 0f399E2DEB;
fma.rn.f32 %f125, %f123, %f104, %f124;
mov.f32 %f126, 0f3A4AE334;
fma.rn.f32 %f127, %f125, %f104, %f126;
mov.f32 %f128, 0fBBEEAA1B;
fma.rn.f32 %f129, %f127, %f104, %f128;
mov.f32 %f130, 0fBCDA7747;
fma.rn.f32 %f131, %f129, %f104, %f130;
mul.f32 %f132, %f104, %f131;
add.f32 %f133, %f7, 0fC0B0A47B;
add.f32 %f134, %f133, 0f339A7A37;
mul.f32 %f135, %f134, %f132;
add.f32 %f136, %f7, 0fC10A75AB;
add.f32 %f137, %f136, 0fB4CCCDED;
mul.f32 %f328, %f137, %f135;
bra.uni BB70_41;
BB70_45:
add.f32 %f224, %f2, 0fC07D4A9A;
add.f32 %f225, %f224, 0fB3D9856A;
mov.f32 %f226, 0fB45E2607;
mov.f32 %f227, 0fB449DD3F;
fma.rn.f32 %f228, %f227, %f225, %f226;
mov.f32 %f229, 0fB6857064;
fma.rn.f32 %f230, %f228, %f225, %f229;
mov.f32 %f231, 0f38554610;
fma.rn.f32 %f232, %f230, %f225, %f231;
mov.f32 %f233, 0f394ACED7;
fma.rn.f32 %f234, %f232, %f225, %f233;
mov.f32 %f235, 0fBB0F1A0C;
fma.rn.f32 %f236, %f234, %f225, %f235;
mov.f32 %f237, 0fBBE07F2E;
fma.rn.f32 %f238, %f236, %f225, %f237;
mov.f32 %f239, 0f3D6FB6B5;
fma.rn.f32 %f240, %f238, %f225, %f239;
mov.f32 %f241, 0f3D504DF1;
fma.rn.f32 %f242, %f240, %f225, %f241;
mov.f32 %f243, 0fBECE1A13;
fma.rn.f32 %f244, %f242, %f225, %f243;
mul.f32 %f335, %f225, %f244;
bra.uni BB70_82;
BB70_47:
add.f32 %f245, %f2, 0fC0E2C0EE;
add.f32 %f246, %f245, 0fB39CE420;
mov.f32 %f247, 0f3629DA6C;
mov.f32 %f248, 0f3510CEBE;
fma.rn.f32 %f249, %f248, %f246, %f247;
mov.f32 %f250, 0fB84054C0;
fma.rn.f32 %f251, %f249, %f246, %f250;
mov.f32 %f252, 0fB91318AB;
fma.rn.f32 %f253, %f251, %f246, %f252;
mov.f32 %f254, 0f3B0E9921;
fma.rn.f32 %f255, %f253, %f246, %f254;
mov.f32 %f256, 0f3B5974D5;
fma.rn.f32 %f257, %f255, %f246, %f256;
mov.f32 %f258, 0fBD44B4D7;
fma.rn.f32 %f259, %f257, %f246, %f258;
mov.f32 %f260, 0fBCAD7799;
fma.rn.f32 %f261, %f259, %f246, %f260;
mov.f32 %f262, 0f3E99A665;
fma.rn.f32 %f263, %f261, %f246, %f262;
mul.f32 %f335, %f246, %f263;
bra.uni BB70_82;
BB70_32:
mov.f32 %f185, 0f3C08839E;
mov.f32 %f186, 0fB94CA1F9;
fma.rn.f32 %f325, %f186, %f20, %f185;
BB70_33:
@%p27 bra BB70_35;
mov.f32 %f187, 0f3D2AAAA5;
fma.rn.f32 %f188, %f325, %f20, %f187;
mov.f32 %f189, 0fBF000000;
fma.rn.f32 %f326, %f188, %f20, %f189;
bra.uni BB70_36;
BB70_35:
mov.f32 %f190, 0fBE2AAAA3;
fma.rn.f32 %f191, %f325, %f20, %f190;
mov.f32 %f192, 0f00000000;
fma.rn.f32 %f326, %f191, %f20, %f192;
BB70_36:
fma.rn.f32 %f327, %f326, %f324, %f324;
@%p27 bra BB70_38;
fma.rn.f32 %f327, %f326, %f20, %f150;
BB70_38:
and.b32 %r240, %r70, 2;
setp.eq.s32 %p30, %r240, 0;
@%p30 bra BB70_40;
mov.f32 %f194, 0f00000000;
mov.f32 %f195, 0fBF800000;
fma.rn.f32 %f327, %f327, %f195, %f194;
BB70_40:
mul.f32 %f328, %f10, %f327;
BB70_41:
mul.f32 %f196, %f321, 0f3F22F983;
fma.rn.f32 %f335, %f196, %f328, %f3;
BB70_82:
setp.geu.f32 %p56, %f1, 0f00000000;
@%p56 bra BB70_84;
mov.f32 %f320, 0fBF800000;
sqrt.rn.f32 %f335, %f320;
BB70_84:
cvta.to.global.u64 %rd50, %rd24;
add.s64 %rd52, %rd50, %rd30;
st.global.f32 [%rd52], %f335;
BB70_85:
ret;
BB70_73:
mov.f32 %f309, 0f3C08839E;
mov.f32 %f310, 0fB94CA1F9;
fma.rn.f32 %f332, %f310, %f49, %f309;
BB70_74:
@%p52 bra BB70_76;
mov.f32 %f311, 0f3D2AAAA5;
fma.rn.f32 %f312, %f332, %f49, %f311;
mov.f32 %f313, 0fBF000000;
fma.rn.f32 %f333, %f312, %f49, %f313;
bra.uni BB70_77;
BB70_76:
mov.f32 %f314, 0fBE2AAAA3;
fma.rn.f32 %f315, %f332, %f49, %f314;
mov.f32 %f316, 0f00000000;
fma.rn.f32 %f333, %f315, %f49, %f316;
BB70_77:
fma.rn.f32 %f334, %f333, %f331, %f331;
@%p52 bra BB70_79;
fma.rn.f32 %f334, %f333, %f49, %f274;
BB70_79:
and.b32 %r330, %r141, 2;
setp.eq.s32 %p55, %r330, 0;
@%p55 bra BB70_81;
mov.f32 %f318, 0f00000000;
mov.f32 %f319, 0fBF800000;
fma.rn.f32 %f334, %f334, %f319, %f318;
BB70_81:
mul.f32 %f335, %f39, %f334;
bra.uni BB70_82;
}
// .globl vec_y1f
.visible .entry vec_y1f(
.param .u64 vec_y1f_param_0,
.param .u64 vec_y1f_param_1,
.param .u64 vec_y1f_param_2
)
{
.local .align 4 .b8 __local_depot71[28];
.reg .b64 %SP;
.reg .b64 %SPL;
.reg .pred %p<59>;
.reg .f32 %f<332>;
.reg .b32 %r<381>;
.reg .b64 %rd<59>;
mov.u64 %rd58, __local_depot71;
cvta.local.u64 %SP, %rd58;
ld.param.u64 %rd26, [vec_y1f_param_0];
ld.param.u64 %rd24, [vec_y1f_param_1];
ld.param.u64 %rd25, [vec_y1f_param_2];
add.u64 %rd27, %SP, 0;
cvta.to.local.u64 %rd57, %rd27;
mov.u32 %r144, %ntid.x;
mov.u32 %r145, %ctaid.x;
mov.u32 %r146, %tid.x;
mad.lo.s32 %r1, %r144, %r145, %r146;
cvt.s64.s32 %rd28, %r1;
setp.ge.u64 %p1, %rd28, %rd26;
@%p1 bra BB71_87;
cvta.to.global.u64 %rd29, %rd25;
mul.wide.s32 %rd30, %r1, 4;
add.s64 %rd31, %rd29, %rd30;
ld.global.f32 %f1, [%rd31];
abs.f32 %f2, %f1;
setp.lt.f32 %p2, %f2, 0f00800000;
@%p2 bra BB71_83;
bra.uni BB71_2;
BB71_83:
mov.f32 %f315, 0fBF22F983;
div.rn.f32 %f331, %f315, %f2;
bra.uni BB71_84;
BB71_2:
setp.gtu.f32 %p3, %f2, 0f3FD96AC4;
@%p3 bra BB71_43;
bra.uni BB71_3;
BB71_43:
setp.gtu.f32 %p33, %f2, 0f40740EEE;
@%p33 bra BB71_45;
bra.uni BB71_44;
BB71_45:
setp.gtu.f32 %p34, %f2, 0f40E06937;
@%p34 bra BB71_47;
bra.uni BB71_46;
BB71_47:
setp.gtu.f32 %p35, %f2, 0f4122C2E3;
@%p35 bra BB71_49;
bra.uni BB71_48;
BB71_49:
abs.f32 %f260, %f2;
mov.f32 %f331, 0f00000000;
setp.eq.f32 %p36, %f260, 0f7F800000;
@%p36 bra BB71_84;
add.s64 %rd13, %rd57, 24;
// inline asm
rcp.approx.ftz.f32 %f261,%f2;
// inline asm
mul.f32 %f263, %f261, %f261;
mov.f32 %f264, 0fBE44AB90;
mov.f32 %f265, 0f3F267F60;
fma.rn.f32 %f266, %f265, %f263, %f264;
mov.f32 %f267, 0f3E3FFEBF;
fma.rn.f32 %f268, %f266, %f263, %f267;
mov.f32 %f269, 0f3F800000;
fma.rn.f32 %f270, %f268, %f263, %f269;
mov.f32 %f271, 0f3EBB73AB;
mov.f32 %f272, 0fBFE4E1AB;
fma.rn.f32 %f273, %f272, %f263, %f271;
mov.f32 %f274, 0fBE27FB6E;
fma.rn.f32 %f275, %f273, %f263, %f274;
mov.f32 %f276, 0f3EBFFFFF;
fma.rn.f32 %f277, %f275, %f263, %f276;
fma.rn.f32 %f38, %f277, %f261, %f2;
rsqrt.approx.f32 %f278, %f2;
mul.f32 %f279, %f278, 0f3F4C422A;
mul.f32 %f39, %f270, %f279;
mul.f32 %f280, %f38, 0f3F22F983;
cvt.rni.s32.f32 %r370, %f280;
cvt.rn.f32.s32 %f281, %r370;
neg.f32 %f282, %f281;
mov.f32 %f283, 0f3FC90FDA;
fma.rn.f32 %f284, %f282, %f283, %f38;
mov.f32 %f285, 0f33A22168;
fma.rn.f32 %f286, %f282, %f285, %f284;
mov.f32 %f287, 0f27C234C5;
fma.rn.f32 %f325, %f282, %f287, %f286;
abs.f32 %f288, %f38;
setp.leu.f32 %p37, %f288, 0f47CE4780;
@%p37 bra BB71_60;
mov.b32 %r73, %f38;
shr.u32 %r74, %r73, 23;
bfe.u32 %r249, %r73, 23, 8;
add.s32 %r250, %r249, -128;
shl.b32 %r251, %r73, 8;
or.b32 %r75, %r251, -2147483648;
shr.u32 %r76, %r250, 5;
mov.u32 %r362, 0;
mov.u64 %rd51, __cudart_i2opi_f;
mov.u32 %r361, -6;
mov.u64 %rd56, %rd57;
BB71_52:
.pragma "nounroll";
ld.const.u32 %r254, [%rd51];
// inline asm
{
mad.lo.cc.u32 %r252, %r254, %r75, %r362;
madc.hi.u32 %r362, %r254, %r75, 0;
}
// inline asm
st.local.u32 [%rd56], %r252;
add.s64 %rd56, %rd56, 4;
add.s64 %rd51, %rd51, 4;
add.s32 %r361, %r361, 1;
setp.ne.s32 %p38, %r361, 0;
@%p38 bra BB71_52;
and.b32 %r81, %r73, -2147483648;
st.local.u32 [%rd13], %r362;
mov.u32 %r257, 6;
sub.s32 %r258, %r257, %r76;
mul.wide.s32 %rd42, %r258, 4;
add.s64 %rd18, %rd57, %rd42;
ld.local.u32 %r363, [%rd18];
ld.local.u32 %r364, [%rd18+-4];
and.b32 %r84, %r74, 31;
setp.eq.s32 %p39, %r84, 0;
@%p39 bra BB71_55;
mov.u32 %r259, 32;
sub.s32 %r260, %r259, %r84;
shr.u32 %r261, %r364, %r260;
shl.b32 %r262, %r363, %r84;
add.s32 %r363, %r261, %r262;
ld.local.u32 %r263, [%rd18+-8];
shr.u32 %r264, %r263, %r260;
shl.b32 %r265, %r364, %r84;
add.s32 %r364, %r264, %r265;
BB71_55:
shr.u32 %r266, %r364, 30;
shl.b32 %r267, %r363, 2;
add.s32 %r365, %r266, %r267;
shl.b32 %r90, %r364, 2;
shr.u32 %r268, %r365, 31;
shr.u32 %r269, %r363, 30;
add.s32 %r91, %r268, %r269;
setp.eq.s32 %p40, %r268, 0;
mov.u32 %r366, %r81;
mov.u32 %r367, %r90;
@%p40 bra BB71_57;
not.b32 %r270, %r365;
neg.s32 %r92, %r90;
setp.eq.s32 %p41, %r90, 0;
selp.u32 %r271, 1, 0, %p41;
add.s32 %r365, %r271, %r270;
xor.b32 %r94, %r81, -2147483648;
mov.u32 %r366, %r94;
mov.u32 %r367, %r92;
BB71_57:
mov.u32 %r96, %r366;
neg.s32 %r272, %r91;
setp.eq.s32 %p42, %r81, 0;
selp.b32 %r370, %r91, %r272, %p42;
clz.b32 %r369, %r365;
setp.eq.s32 %p43, %r369, 0;
shl.b32 %r273, %r365, %r369;
mov.u32 %r274, 32;
sub.s32 %r275, %r274, %r369;
shr.u32 %r276, %r367, %r275;
add.s32 %r277, %r276, %r273;
selp.b32 %r100, %r365, %r277, %p43;
mov.u32 %r278, -921707870;
mul.hi.u32 %r368, %r100, %r278;
setp.lt.s32 %p44, %r368, 1;
@%p44 bra BB71_59;
mul.lo.s32 %r279, %r100, -921707870;
shr.u32 %r280, %r279, 31;
shl.b32 %r281, %r368, 1;
add.s32 %r368, %r280, %r281;
add.s32 %r369, %r369, 1;
BB71_59:
mov.u32 %r282, 126;
sub.s32 %r283, %r282, %r369;
shl.b32 %r284, %r283, 23;
add.s32 %r285, %r368, 1;
shr.u32 %r286, %r285, 7;
add.s32 %r287, %r286, 1;
shr.u32 %r288, %r287, 1;
add.s32 %r289, %r288, %r284;
or.b32 %r290, %r289, %r96;
mov.b32 %f325, %r290;
BB71_60:
and.b32 %r291, %r370, 3;
cvt.rn.f32.s32 %f289, %r291;
add.f32 %f290, %f325, 0fC07B53D1;
fma.rn.f32 %f326, %f289, 0f3FC90FDB, %f290;
abs.f32 %f291, %f326;
setp.neu.f32 %p45, %f291, 0f7F800000;
@%p45 bra BB71_62;
mov.f32 %f292, 0f00000000;
mul.rn.f32 %f326, %f326, %f292;
BB71_62:
mul.f32 %f293, %f326, 0f3F22F983;
cvt.rni.s32.f32 %r380, %f293;
cvt.rn.f32.s32 %f294, %r380;
neg.f32 %f295, %f294;
fma.rn.f32 %f297, %f295, %f283, %f326;
fma.rn.f32 %f299, %f295, %f285, %f297;
fma.rn.f32 %f327, %f295, %f287, %f299;
abs.f32 %f301, %f326;
setp.leu.f32 %p46, %f301, 0f47CE4780;
@%p46 bra BB71_72;
mov.b32 %r108, %f326;
shr.u32 %r109, %r108, 23;
bfe.u32 %r294, %r108, 23, 8;
add.s32 %r295, %r294, -128;
shl.b32 %r296, %r108, 8;
or.b32 %r110, %r296, -2147483648;
shr.u32 %r111, %r295, 5;
mov.u32 %r372, 0;
mov.u64 %rd52, __cudart_i2opi_f;
mov.u32 %r371, -6;
mov.u64 %rd55, %rd57;
BB71_64:
.pragma "nounroll";
ld.const.u32 %r299, [%rd52];
// inline asm
{
mad.lo.cc.u32 %r297, %r299, %r110, %r372;
madc.hi.u32 %r372, %r299, %r110, 0;
}
// inline asm
st.local.u32 [%rd55], %r297;
add.s64 %rd55, %rd55, 4;
add.s64 %rd52, %rd52, 4;
add.s32 %r371, %r371, 1;
setp.ne.s32 %p47, %r371, 0;
@%p47 bra BB71_64;
and.b32 %r116, %r108, -2147483648;
st.local.u32 [%rd13], %r372;
mov.u32 %r302, 6;
sub.s32 %r303, %r302, %r111;
mul.wide.s32 %rd44, %r303, 4;
add.s64 %rd23, %rd57, %rd44;
ld.local.u32 %r373, [%rd23];
ld.local.u32 %r374, [%rd23+-4];
and.b32 %r119, %r109, 31;
setp.eq.s32 %p48, %r119, 0;
@%p48 bra BB71_67;
mov.u32 %r304, 32;
sub.s32 %r305, %r304, %r119;
shr.u32 %r306, %r374, %r305;
shl.b32 %r307, %r373, %r119;
add.s32 %r373, %r306, %r307;
ld.local.u32 %r308, [%rd23+-8];
shr.u32 %r309, %r308, %r305;
shl.b32 %r310, %r374, %r119;
add.s32 %r374, %r309, %r310;
BB71_67:
shr.u32 %r311, %r374, 30;
shl.b32 %r312, %r373, 2;
add.s32 %r375, %r311, %r312;
shl.b32 %r125, %r374, 2;
shr.u32 %r313, %r375, 31;
shr.u32 %r314, %r373, 30;
add.s32 %r126, %r313, %r314;
setp.eq.s32 %p49, %r313, 0;
mov.u32 %r376, %r116;
mov.u32 %r377, %r125;
@%p49 bra BB71_69;
not.b32 %r315, %r375;
neg.s32 %r127, %r125;
setp.eq.s32 %p50, %r125, 0;
selp.u32 %r316, 1, 0, %p50;
add.s32 %r375, %r316, %r315;
xor.b32 %r129, %r116, -2147483648;
mov.u32 %r376, %r129;
mov.u32 %r377, %r127;
BB71_69:
mov.u32 %r131, %r376;
neg.s32 %r317, %r126;
setp.eq.s32 %p51, %r116, 0;
selp.b32 %r380, %r126, %r317, %p51;
clz.b32 %r379, %r375;
setp.eq.s32 %p52, %r379, 0;
shl.b32 %r318, %r375, %r379;
mov.u32 %r319, 32;
sub.s32 %r320, %r319, %r379;
shr.u32 %r321, %r377, %r320;
add.s32 %r322, %r321, %r318;
selp.b32 %r135, %r375, %r322, %p52;
mov.u32 %r323, -921707870;
mul.hi.u32 %r378, %r135, %r323;
setp.lt.s32 %p53, %r378, 1;
@%p53 bra BB71_71;
mul.lo.s32 %r324, %r135, -921707870;
shr.u32 %r325, %r324, 31;
shl.b32 %r326, %r378, 1;
add.s32 %r378, %r325, %r326;
add.s32 %r379, %r379, 1;
BB71_71:
mov.u32 %r327, 126;
sub.s32 %r328, %r327, %r379;
shl.b32 %r329, %r328, 23;
add.s32 %r330, %r378, 1;
shr.u32 %r331, %r330, 7;
add.s32 %r332, %r331, 1;
shr.u32 %r333, %r332, 1;
add.s32 %r334, %r333, %r329;
or.b32 %r335, %r334, %r131;
mov.b32 %f327, %r335;
BB71_72:
mul.rn.f32 %f49, %f327, %f327;
add.s32 %r142, %r380, 1;
and.b32 %r143, %r142, 1;
setp.eq.s32 %p54, %r143, 0;
@%p54 bra BB71_74;
mov.f32 %f302, 0fBAB6061A;
mov.f32 %f303, 0f37CCF5CE;
fma.rn.f32 %f328, %f303, %f49, %f302;
bra.uni BB71_75;
BB71_3:
mul.f32 %f66, %f2, %f2;
mov.f32 %f67, 0fB58527DA;
mov.f32 %f68, 0f321462CC;
fma.rn.f32 %f69, %f68, %f66, %f67;
mov.f32 %f70, 0f38963E95;
fma.rn.f32 %f71, %f69, %f66, %f70;
mov.f32 %f72, 0fBB41ADCB;
fma.rn.f32 %f73, %f71, %f66, %f72;
mov.f32 %f74, 0f3D5E9CBB;
fma.rn.f32 %f75, %f73, %f66, %f74;
mov.f32 %f76, 0fBE48C331;
fma.rn.f32 %f3, %f75, %f66, %f76;
setp.lt.f32 %p4, %f2, 0f7F800000;
setp.gt.f32 %p5, %f2, 0f00000000;
and.pred %p6, %p5, %p4;
@%p6 bra BB71_5;
bra.uni BB71_4;
BB71_5:
mov.b32 %r147, %f2;
and.b32 %r148, %r147, 8388607;
or.b32 %r149, %r148, 1065353216;
mov.b32 %f79, %r149;
shr.u32 %r150, %r147, 23;
cvt.rn.f32.u32 %f80, %r150;
add.f32 %f81, %f80, 0fC2FE0000;
setp.gt.f32 %p7, %f79, 0f3FAE147B;
mul.f32 %f82, %f79, 0f3F000000;
add.f32 %f83, %f81, 0f3F800000;
selp.f32 %f84, %f82, %f79, %p7;
selp.f32 %f85, %f83, %f81, %p7;
add.f32 %f78, %f84, 0f3F800000;
add.f32 %f86, %f84, 0fBF800000;
// inline asm
rcp.approx.ftz.f32 %f77,%f78;
// inline asm
mul.f32 %f87, %f86, %f86;
neg.f32 %f88, %f87;
mul.rn.f32 %f89, %f77, %f88;
add.rn.f32 %f90, %f86, %f89;
mul.f32 %f91, %f90, %f90;
mov.f32 %f92, 0f3C4C6A36;
mov.f32 %f93, 0f3B1E94E6;
fma.rn.f32 %f94, %f93, %f91, %f92;
mov.f32 %f95, 0f3DAAAB1A;
fma.rn.f32 %f96, %f94, %f91, %f95;
mul.f32 %f97, %f91, %f96;
fma.rn.f32 %f98, %f97, %f90, %f89;
add.f32 %f99, %f86, %f98;
mov.f32 %f100, 0f3F317218;
fma.rn.f32 %f317, %f85, %f100, %f99;
bra.uni BB71_6;
BB71_44:
add.f32 %f198, %f2, 0fC00C9DF7;
add.f32 %f199, %f198, 0f33B200DC;
mov.f32 %f200, 0f39064A88;
mov.f32 %f201, 0fB789E29D;
fma.rn.f32 %f202, %f201, %f199, %f200;
mov.f32 %f203, 0fB9F0AB0D;
fma.rn.f32 %f204, %f202, %f199, %f203;
mov.f32 %f205, 0f3A8F6102;
fma.rn.f32 %f206, %f204, %f199, %f205;
mov.f32 %f207, 0fBB2C7045;
fma.rn.f32 %f208, %f206, %f199, %f207;
mov.f32 %f209, 0f3BF35DF7;
fma.rn.f32 %f210, %f208, %f199, %f209;
mov.f32 %f211, 0fBB9D097C;
fma.rn.f32 %f212, %f210, %f199, %f211;
mov.f32 %f213, 0fBD06968A;
fma.rn.f32 %f214, %f212, %f199, %f213;
mov.f32 %f215, 0fBDF2B7DF;
fma.rn.f32 %f216, %f214, %f199, %f215;
mov.f32 %f217, 0f3F055242;
fma.rn.f32 %f218, %f216, %f199, %f217;
mul.f32 %f331, %f199, %f218;
bra.uni BB71_84;
BB71_4:
lg2.approx.f32 %f317, %f2;
BB71_6:
abs.f32 %f7, %f2;
setp.gtu.f32 %p8, %f7, 0f40FB3333;
@%p8 bra BB71_8;
bra.uni BB71_7;
BB71_8:
abs.f32 %f133, %f7;
mov.f32 %f324, 0f00000000;
setp.eq.f32 %p9, %f133, 0f7F800000;
@%p9 bra BB71_42;
// inline asm
rcp.approx.ftz.f32 %f134,%f7;
// inline asm
mul.f32 %f136, %f134, %f134;
mov.f32 %f137, 0f3F3FF7E9;
mov.f32 %f138, 0fC082CB37;
fma.rn.f32 %f139, %f138, %f136, %f137;
mov.f32 %f140, 0fBE458BAE;
fma.rn.f32 %f141, %f139, %f136, %f140;
mov.f32 %f142, 0f3E3FFF8B;
fma.rn.f32 %f143, %f141, %f136, %f142;
mov.f32 %f144, 0f3F800000;
fma.rn.f32 %f145, %f143, %f136, %f144;
mov.f32 %f146, 0f3EB914AD;
mov.f32 %f147, 0fBFCA3BA2;
fma.rn.f32 %f148, %f147, %f136, %f146;
mov.f32 %f149, 0fBE27F2EC;
fma.rn.f32 %f150, %f148, %f136, %f149;
mov.f32 %f151, 0f3EBFFFFD;
fma.rn.f32 %f152, %f150, %f136, %f151;
fma.rn.f32 %f9, %f152, %f134, %f7;
rsqrt.approx.f32 %f153, %f7;
mul.f32 %f154, %f153, 0f3F4C422A;
mul.f32 %f10, %f145, %f154;
mul.f32 %f155, %f9, 0f3F22F983;
cvt.rni.s32.f32 %r350, %f155;
cvt.rn.f32.s32 %f156, %r350;
neg.f32 %f157, %f156;
mov.f32 %f158, 0f3FC90FDA;
fma.rn.f32 %f159, %f157, %f158, %f9;
mov.f32 %f160, 0f33A22168;
fma.rn.f32 %f161, %f157, %f160, %f159;
mov.f32 %f162, 0f27C234C5;
fma.rn.f32 %f318, %f157, %f162, %f161;
abs.f32 %f163, %f9;
setp.leu.f32 %p10, %f163, 0f47CE4780;
@%p10 bra BB71_19;
mov.b32 %r3, %f9;
shl.b32 %r153, %r3, 8;
or.b32 %r4, %r153, -2147483648;
mov.u32 %r342, 0;
mov.u64 %rd48, __cudart_i2opi_f;
mov.u32 %r341, -6;
BB71_11:
.pragma "nounroll";
ld.const.u32 %r156, [%rd48];
// inline asm
{
mad.lo.cc.u32 %r154, %r156, %r4, %r342;
madc.hi.u32 %r342, %r156, %r4, 0;
}
// inline asm
st.local.u32 [%rd57], %r154;
add.s64 %rd57, %rd57, 4;
add.s64 %rd48, %rd48, 4;
add.s32 %r341, %r341, 1;
setp.ne.s32 %p11, %r341, 0;
@%p11 bra BB71_11;
and.b32 %r9, %r3, -2147483648;
bfe.u32 %r159, %r3, 23, 8;
add.s32 %r160, %r159, -128;
shr.u32 %r161, %r160, 5;
cvta.to.local.u64 %rd34, %rd27;
st.local.u32 [%rd34+24], %r342;
bfe.u32 %r10, %r3, 23, 5;
mov.u32 %r162, 6;
sub.s32 %r163, %r162, %r161;
mul.wide.s32 %rd35, %r163, 4;
add.s64 %rd6, %rd34, %rd35;
ld.local.u32 %r343, [%rd6];
ld.local.u32 %r344, [%rd6+-4];
setp.eq.s32 %p12, %r10, 0;
@%p12 bra BB71_14;
mov.u32 %r164, 32;
sub.s32 %r165, %r164, %r10;
shr.u32 %r166, %r344, %r165;
shl.b32 %r167, %r343, %r10;
add.s32 %r343, %r166, %r167;
ld.local.u32 %r168, [%rd6+-8];
shr.u32 %r169, %r168, %r165;
shl.b32 %r170, %r344, %r10;
add.s32 %r344, %r169, %r170;
BB71_14:
shr.u32 %r171, %r344, 30;
shl.b32 %r172, %r343, 2;
add.s32 %r345, %r171, %r172;
shl.b32 %r18, %r344, 2;
shr.u32 %r173, %r345, 31;
shr.u32 %r174, %r343, 30;
add.s32 %r19, %r173, %r174;
setp.eq.s32 %p13, %r173, 0;
mov.u32 %r346, %r9;
mov.u32 %r347, %r18;
@%p13 bra BB71_16;
not.b32 %r175, %r345;
neg.s32 %r20, %r18;
setp.eq.s32 %p14, %r18, 0;
selp.u32 %r176, 1, 0, %p14;
add.s32 %r345, %r176, %r175;
xor.b32 %r22, %r9, -2147483648;
mov.u32 %r346, %r22;
mov.u32 %r347, %r20;
BB71_16:
mov.u32 %r24, %r346;
neg.s32 %r177, %r19;
setp.eq.s32 %p15, %r9, 0;
selp.b32 %r350, %r19, %r177, %p15;
clz.b32 %r349, %r345;
setp.eq.s32 %p16, %r349, 0;
shl.b32 %r178, %r345, %r349;
mov.u32 %r179, 32;
sub.s32 %r180, %r179, %r349;
shr.u32 %r181, %r347, %r180;
add.s32 %r182, %r181, %r178;
selp.b32 %r28, %r345, %r182, %p16;
mov.u32 %r183, -921707870;
mul.hi.u32 %r348, %r28, %r183;
setp.lt.s32 %p17, %r348, 1;
@%p17 bra BB71_18;
mul.lo.s32 %r184, %r28, -921707870;
shr.u32 %r185, %r184, 31;
shl.b32 %r186, %r348, 1;
add.s32 %r348, %r185, %r186;
add.s32 %r349, %r349, 1;
BB71_18:
mov.u32 %r187, 126;
sub.s32 %r188, %r187, %r349;
shl.b32 %r189, %r188, 23;
add.s32 %r190, %r348, 1;
shr.u32 %r191, %r190, 7;
add.s32 %r192, %r191, 1;
shr.u32 %r193, %r192, 1;
add.s32 %r194, %r193, %r189;
or.b32 %r195, %r194, %r24;
mov.b32 %f318, %r195;
BB71_19:
and.b32 %r196, %r350, 3;
cvt.rn.f32.s32 %f164, %r196;
add.f32 %f165, %f318, 0fC016CBE4;
fma.rn.f32 %f319, %f164, 0f3FC90FDB, %f165;
abs.f32 %f166, %f319;
setp.neu.f32 %p18, %f166, 0f7F800000;
@%p18 bra BB71_21;
mov.f32 %f167, 0f00000000;
mul.rn.f32 %f319, %f319, %f167;
BB71_21:
mul.f32 %f168, %f319, 0f3F22F983;
cvt.rni.s32.f32 %r360, %f168;
cvt.rn.f32.s32 %f169, %r360;
neg.f32 %f170, %f169;
fma.rn.f32 %f172, %f170, %f158, %f319;
fma.rn.f32 %f174, %f170, %f160, %f172;
fma.rn.f32 %f320, %f170, %f162, %f174;
abs.f32 %f176, %f319;
setp.leu.f32 %p19, %f176, 0f47CE4780;
@%p19 bra BB71_31;
mov.b32 %r36, %f319;
shr.u32 %r37, %r36, 23;
bfe.u32 %r199, %r36, 23, 8;
add.s32 %r200, %r199, -128;
shl.b32 %r201, %r36, 8;
or.b32 %r38, %r201, -2147483648;
shr.u32 %r39, %r200, 5;
cvta.to.local.u64 %rd50, %rd27;
mov.u32 %r352, 0;
mov.u64 %rd49, __cudart_i2opi_f;
mov.u32 %r351, -6;
BB71_23:
.pragma "nounroll";
ld.const.u32 %r204, [%rd49];
// inline asm
{
mad.lo.cc.u32 %r202, %r204, %r38, %r352;
madc.hi.u32 %r352, %r204, %r38, 0;
}
// inline asm
st.local.u32 [%rd50], %r202;
add.s64 %rd50, %rd50, 4;
add.s64 %rd49, %rd49, 4;
add.s32 %r351, %r351, 1;
setp.ne.s32 %p20, %r351, 0;
@%p20 bra BB71_23;
and.b32 %r44, %r36, -2147483648;
cvta.to.local.u64 %rd39, %rd27;
st.local.u32 [%rd39+24], %r352;
mov.u32 %r207, 6;
sub.s32 %r208, %r207, %r39;
mul.wide.s32 %rd40, %r208, 4;
add.s64 %rd12, %rd39, %rd40;
ld.local.u32 %r353, [%rd12];
ld.local.u32 %r354, [%rd12+-4];
and.b32 %r47, %r37, 31;
setp.eq.s32 %p21, %r47, 0;
@%p21 bra BB71_26;
mov.u32 %r209, 32;
sub.s32 %r210, %r209, %r47;
shr.u32 %r211, %r354, %r210;
shl.b32 %r212, %r353, %r47;
add.s32 %r353, %r211, %r212;
ld.local.u32 %r213, [%rd12+-8];
shr.u32 %r214, %r213, %r210;
shl.b32 %r215, %r354, %r47;
add.s32 %r354, %r214, %r215;
BB71_26:
shr.u32 %r216, %r354, 30;
shl.b32 %r217, %r353, 2;
add.s32 %r355, %r216, %r217;
shl.b32 %r53, %r354, 2;
shr.u32 %r218, %r355, 31;
shr.u32 %r219, %r353, 30;
add.s32 %r54, %r218, %r219;
setp.eq.s32 %p22, %r218, 0;
mov.u32 %r356, %r44;
mov.u32 %r357, %r53;
@%p22 bra BB71_28;
not.b32 %r220, %r355;
neg.s32 %r55, %r53;
setp.eq.s32 %p23, %r53, 0;
selp.u32 %r221, 1, 0, %p23;
add.s32 %r355, %r221, %r220;
xor.b32 %r57, %r44, -2147483648;
mov.u32 %r356, %r57;
mov.u32 %r357, %r55;
BB71_28:
mov.u32 %r59, %r356;
neg.s32 %r222, %r54;
setp.eq.s32 %p24, %r44, 0;
selp.b32 %r360, %r54, %r222, %p24;
clz.b32 %r359, %r355;
setp.eq.s32 %p25, %r359, 0;
shl.b32 %r223, %r355, %r359;
mov.u32 %r224, 32;
sub.s32 %r225, %r224, %r359;
shr.u32 %r226, %r357, %r225;
add.s32 %r227, %r226, %r223;
selp.b32 %r63, %r355, %r227, %p25;
mov.u32 %r228, -921707870;
mul.hi.u32 %r358, %r63, %r228;
setp.lt.s32 %p26, %r358, 1;
@%p26 bra BB71_30;
mul.lo.s32 %r229, %r63, -921707870;
shr.u32 %r230, %r229, 31;
shl.b32 %r231, %r358, 1;
add.s32 %r358, %r230, %r231;
add.s32 %r359, %r359, 1;
BB71_30:
mov.u32 %r232, 126;
sub.s32 %r233, %r232, %r359;
shl.b32 %r234, %r233, 23;
add.s32 %r235, %r358, 1;
shr.u32 %r236, %r235, 7;
add.s32 %r237, %r236, 1;
shr.u32 %r238, %r237, 1;
add.s32 %r239, %r238, %r234;
or.b32 %r240, %r239, %r59;
mov.b32 %f320, %r240;
BB71_31:
mul.rn.f32 %f20, %f320, %f320;
add.s32 %r70, %r360, 1;
and.b32 %r71, %r70, 1;
setp.eq.s32 %p27, %r71, 0;
@%p27 bra BB71_33;
mov.f32 %f177, 0fBAB6061A;
mov.f32 %f178, 0f37CCF5CE;
fma.rn.f32 %f321, %f178, %f20, %f177;
bra.uni BB71_34;
BB71_7:
add.f32 %f101, %f7, 0fC0753AAC;
add.f32 %f102, %f101, 0f33A5090F;
mov.f32 %f103, 0f2B81BF42;
mov.f32 %f104, 0f29AF3463;
fma.rn.f32 %f105, %f104, %f102, %f103;
mov.f32 %f106, 0fADE21EC1;
fma.rn.f32 %f107, %f105, %f102, %f106;
mov.f32 %f108, 0fAF5DDEFF;
fma.rn.f32 %f109, %f107, %f102, %f108;
mov.f32 %f110, 0f319B0C9D;
fma.rn.f32 %f111, %f109, %f102, %f110;
mov.f32 %f112, 0f32E81173;
fma.rn.f32 %f113, %f111, %f102, %f112;
mov.f32 %f114, 0fB50F8DC8;
fma.rn.f32 %f115, %f113, %f102, %f114;
mov.f32 %f116, 0fB61E653D;
fma.rn.f32 %f117, %f115, %f102, %f116;
mov.f32 %f118, 0f382CD9C5;
fma.rn.f32 %f119, %f117, %f102, %f118;
mov.f32 %f120, 0f38F9EB10;
fma.rn.f32 %f121, %f119, %f102, %f120;
mov.f32 %f122, 0fBAECEB9C;
fma.rn.f32 %f123, %f121, %f102, %f122;
mov.f32 %f124, 0fBB276FFD;
fma.rn.f32 %f125, %f123, %f102, %f124;
mov.f32 %f126, 0f3D073993;
fma.rn.f32 %f127, %f125, %f102, %f126;
add.f32 %f128, %f7, 0fC0E07FB0;
add.f32 %f129, %f128, 0f3444B8DB;
mul.f32 %f130, %f129, %f127;
mul.f32 %f131, %f102, %f130;
mul.f32 %f324, %f7, %f131;
bra.uni BB71_42;
BB71_46:
add.f32 %f219, %f2, 0fC0ADBFF2;
add.f32 %f220, %f219, 0fB4687B03;
mov.f32 %f221, 0fB508A416;
mov.f32 %f222, 0f32BE57D0;
fma.rn.f32 %f223, %f222, %f220, %f221;
mov.f32 %f224, 0fB63F8A14;
fma.rn.f32 %f225, %f223, %f220, %f224;
mov.f32 %f226, 0f38427E02;
fma.rn.f32 %f227, %f225, %f220, %f226;
mov.f32 %f228, 0f3919BB1C;
fma.rn.f32 %f229, %f227, %f220, %f228;
mov.f32 %f230, 0fBB0DF1FD;
fma.rn.f32 %f231, %f229, %f220, %f230;
mov.f32 %f232, 0fBB885189;
fma.rn.f32 %f233, %f231, %f220, %f232;
mov.f32 %f234, 0f3D50AEC1;
fma.rn.f32 %f235, %f233, %f220, %f234;
mov.f32 %f236, 0f3D005CFC;
fma.rn.f32 %f237, %f235, %f220, %f236;
mov.f32 %f238, 0fBEAE3E2B;
fma.rn.f32 %f239, %f237, %f220, %f238;
mul.f32 %f331, %f220, %f239;
bra.uni BB71_84;
BB71_48:
add.f32 %f240, %f2, 0fC109893D;
add.f32 %f241, %f240, 0fB4E6169B;
mov.f32 %f242, 0f3602902E;
mov.f32 %f243, 0f350CF383;
fma.rn.f32 %f244, %f243, %f241, %f242;
mov.f32 %f245, 0fB8375F71;
fma.rn.f32 %f246, %f244, %f241, %f245;
mov.f32 %f247, 0fB8D9FAA8;
fma.rn.f32 %f248, %f246, %f241, %f247;
mov.f32 %f249, 0f3B03D19A;
fma.rn.f32 %f250, %f248, %f241, %f249;
mov.f32 %f251, 0f3B1E736D;
fma.rn.f32 %f252, %f250, %f241, %f251;
mov.f32 %f253, 0fBD31CAE5;
fma.rn.f32 %f254, %f252, %f241, %f253;
mov.f32 %f255, 0fBC8159B6;
fma.rn.f32 %f256, %f254, %f241, %f255;
mov.f32 %f257, 0f3E8AFCCA;
fma.rn.f32 %f258, %f256, %f241, %f257;
mul.f32 %f331, %f241, %f258;
bra.uni BB71_84;
BB71_33:
mov.f32 %f179, 0f3C08839E;
mov.f32 %f180, 0fB94CA1F9;
fma.rn.f32 %f321, %f180, %f20, %f179;
BB71_34:
@%p27 bra BB71_36;
mov.f32 %f181, 0f3D2AAAA5;
fma.rn.f32 %f182, %f321, %f20, %f181;
mov.f32 %f183, 0fBF000000;
fma.rn.f32 %f322, %f182, %f20, %f183;
bra.uni BB71_37;
BB71_36:
mov.f32 %f184, 0fBE2AAAA3;
fma.rn.f32 %f185, %f321, %f20, %f184;
mov.f32 %f186, 0f00000000;
fma.rn.f32 %f322, %f185, %f20, %f186;
BB71_37:
fma.rn.f32 %f323, %f322, %f320, %f320;
@%p27 bra BB71_39;
fma.rn.f32 %f323, %f322, %f20, %f144;
BB71_39:
and.b32 %r241, %r70, 2;
setp.eq.s32 %p30, %r241, 0;
@%p30 bra BB71_41;
mov.f32 %f188, 0f00000000;
mov.f32 %f189, 0fBF800000;
fma.rn.f32 %f323, %f323, %f189, %f188;
BB71_41:
mul.f32 %f324, %f10, %f323;
BB71_42:
neg.f32 %f190, %f324;
setp.lt.f32 %p31, %f2, 0f00000000;
selp.f32 %f191, %f190, %f324, %p31;
mov.b32 %r242, %f2;
and.b32 %r243, %r242, -2147483648;
mov.b32 %r244, %f191;
and.b32 %r245, %r244, 2147483647;
or.b32 %r246, %r245, %r243;
mov.b32 %f192, %r246;
setp.lt.f32 %p32, %f7, 0f0DA24260;
selp.f32 %f193, %f192, %f191, %p32;
mov.f32 %f194, 0fBF800000;
div.rn.f32 %f195, %f194, %f2;
fma.rn.f32 %f196, %f317, %f193, %f195;
mul.f32 %f197, %f196, 0f3F22F983;
fma.rn.f32 %f331, %f2, %f3, %f197;
BB71_84:
setp.geu.f32 %p58, %f1, 0f00000000;
@%p58 bra BB71_86;
mov.f32 %f316, 0fBF800000;
sqrt.rn.f32 %f331, %f316;
BB71_86:
cvta.to.global.u64 %rd45, %rd24;
add.s64 %rd47, %rd45, %rd30;
st.global.f32 [%rd47], %f331;
BB71_87:
ret;
BB71_74:
mov.f32 %f304, 0f3C08839E;
mov.f32 %f305, 0fB94CA1F9;
fma.rn.f32 %f328, %f305, %f49, %f304;
BB71_75:
@%p54 bra BB71_77;
mov.f32 %f306, 0f3D2AAAA5;
fma.rn.f32 %f307, %f328, %f49, %f306;
mov.f32 %f308, 0fBF000000;
fma.rn.f32 %f329, %f307, %f49, %f308;
bra.uni BB71_78;
BB71_77:
mov.f32 %f309, 0fBE2AAAA3;
fma.rn.f32 %f310, %f328, %f49, %f309;
mov.f32 %f311, 0f00000000;
fma.rn.f32 %f329, %f310, %f49, %f311;
BB71_78:
fma.rn.f32 %f330, %f329, %f327, %f327;
@%p54 bra BB71_80;
fma.rn.f32 %f330, %f329, %f49, %f269;
BB71_80:
and.b32 %r336, %r142, 2;
setp.eq.s32 %p57, %r336, 0;
@%p57 bra BB71_82;
mov.f32 %f313, 0f00000000;
mov.f32 %f314, 0fBF800000;
fma.rn.f32 %f330, %f330, %f314, %f313;
BB71_82:
mul.f32 %f331, %f39, %f330;
bra.uni BB71_84;
}
// .globl vec_copysignf
.visible .entry vec_copysignf(
.param .u64 vec_copysignf_param_0,
.param .u64 vec_copysignf_param_1,
.param .u64 vec_copysignf_param_2,
.param .u64 vec_copysignf_param_3
)
{
.reg .pred %p<2>;
.reg .b32 %r<10>;
.reg .b64 %rd<13>;
ld.param.u64 %rd5, [vec_copysignf_param_0];
ld.param.u64 %rd2, [vec_copysignf_param_1];
ld.param.u64 %rd3, [vec_copysignf_param_2];
ld.param.u64 %rd4, [vec_copysignf_param_3];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd5;
@%p1 bra BB72_2;
cvta.to.global.u64 %rd6, %rd3;
shl.b64 %rd7, %rd1, 2;
add.s64 %rd8, %rd6, %rd7;
cvta.to.global.u64 %rd9, %rd4;
add.s64 %rd10, %rd9, %rd7;
ld.global.u32 %r5, [%rd10];
and.b32 %r6, %r5, -2147483648;
ld.global.u32 %r7, [%rd8];
and.b32 %r8, %r7, 2147483647;
or.b32 %r9, %r6, %r8;
cvta.to.global.u64 %rd11, %rd2;
add.s64 %rd12, %rd11, %rd7;
st.global.u32 [%rd12], %r9;
BB72_2:
ret;
}
// .globl vec_fdimf
.visible .entry vec_fdimf(
.param .u64 vec_fdimf_param_0,
.param .u64 vec_fdimf_param_1,
.param .u64 vec_fdimf_param_2,
.param .u64 vec_fdimf_param_3
)
{
.reg .pred %p<3>;
.reg .f32 %f<5>;
.reg .b32 %r<5>;
.reg .b64 %rd<13>;
ld.param.u64 %rd5, [vec_fdimf_param_0];
ld.param.u64 %rd2, [vec_fdimf_param_1];
ld.param.u64 %rd3, [vec_fdimf_param_2];
ld.param.u64 %rd4, [vec_fdimf_param_3];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd5;
@%p1 bra BB73_2;
cvta.to.global.u64 %rd6, %rd3;
shl.b64 %rd7, %rd1, 2;
add.s64 %rd8, %rd6, %rd7;
cvta.to.global.u64 %rd9, %rd4;
add.s64 %rd10, %rd9, %rd7;
ld.global.f32 %f1, [%rd10];
ld.global.f32 %f2, [%rd8];
sub.f32 %f3, %f2, %f1;
setp.gtu.f32 %p2, %f2, %f1;
selp.f32 %f4, %f3, 0f00000000, %p2;
cvta.to.global.u64 %rd11, %rd2;
add.s64 %rd12, %rd11, %rd7;
st.global.f32 [%rd12], %f4;
BB73_2:
ret;
}
// .globl vec_fdividef
.visible .entry vec_fdividef(
.param .u64 vec_fdividef_param_0,
.param .u64 vec_fdividef_param_1,
.param .u64 vec_fdividef_param_2,
.param .u64 vec_fdividef_param_3
)
{
.reg .pred %p<2>;
.reg .f32 %f<4>;
.reg .b32 %r<5>;
.reg .b64 %rd<13>;
ld.param.u64 %rd5, [vec_fdividef_param_0];
ld.param.u64 %rd2, [vec_fdividef_param_1];
ld.param.u64 %rd3, [vec_fdividef_param_2];
ld.param.u64 %rd4, [vec_fdividef_param_3];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd5;
@%p1 bra BB74_2;
cvta.to.global.u64 %rd6, %rd3;
shl.b64 %rd7, %rd1, 2;
add.s64 %rd8, %rd6, %rd7;
cvta.to.global.u64 %rd9, %rd4;
add.s64 %rd10, %rd9, %rd7;
ld.global.f32 %f1, [%rd10];
ld.global.f32 %f2, [%rd8];
div.rn.f32 %f3, %f2, %f1;
cvta.to.global.u64 %rd11, %rd2;
add.s64 %rd12, %rd11, %rd7;
st.global.f32 [%rd12], %f3;
BB74_2:
ret;
}
// .globl vec_fmaxf
.visible .entry vec_fmaxf(
.param .u64 vec_fmaxf_param_0,
.param .u64 vec_fmaxf_param_1,
.param .u64 vec_fmaxf_param_2,
.param .u64 vec_fmaxf_param_3
)
{
.reg .pred %p<2>;
.reg .f32 %f<4>;
.reg .b32 %r<5>;
.reg .b64 %rd<13>;
ld.param.u64 %rd5, [vec_fmaxf_param_0];
ld.param.u64 %rd2, [vec_fmaxf_param_1];
ld.param.u64 %rd3, [vec_fmaxf_param_2];
ld.param.u64 %rd4, [vec_fmaxf_param_3];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd5;
@%p1 bra BB75_2;
cvta.to.global.u64 %rd6, %rd3;
shl.b64 %rd7, %rd1, 2;
add.s64 %rd8, %rd6, %rd7;
cvta.to.global.u64 %rd9, %rd4;
add.s64 %rd10, %rd9, %rd7;
ld.global.f32 %f1, [%rd10];
ld.global.f32 %f2, [%rd8];
max.f32 %f3, %f2, %f1;
cvta.to.global.u64 %rd11, %rd2;
add.s64 %rd12, %rd11, %rd7;
st.global.f32 [%rd12], %f3;
BB75_2:
ret;
}
// .globl vec_fminf
.visible .entry vec_fminf(
.param .u64 vec_fminf_param_0,
.param .u64 vec_fminf_param_1,
.param .u64 vec_fminf_param_2,
.param .u64 vec_fminf_param_3
)
{
.reg .pred %p<2>;
.reg .f32 %f<4>;
.reg .b32 %r<5>;
.reg .b64 %rd<13>;
ld.param.u64 %rd5, [vec_fminf_param_0];
ld.param.u64 %rd2, [vec_fminf_param_1];
ld.param.u64 %rd3, [vec_fminf_param_2];
ld.param.u64 %rd4, [vec_fminf_param_3];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd5;
@%p1 bra BB76_2;
cvta.to.global.u64 %rd6, %rd3;
shl.b64 %rd7, %rd1, 2;
add.s64 %rd8, %rd6, %rd7;
cvta.to.global.u64 %rd9, %rd4;
add.s64 %rd10, %rd9, %rd7;
ld.global.f32 %f1, [%rd10];
ld.global.f32 %f2, [%rd8];
min.f32 %f3, %f2, %f1;
cvta.to.global.u64 %rd11, %rd2;
add.s64 %rd12, %rd11, %rd7;
st.global.f32 [%rd12], %f3;
BB76_2:
ret;
}
// .globl vec_fmodf
.visible .entry vec_fmodf(
.param .u64 vec_fmodf_param_0,
.param .u64 vec_fmodf_param_1,
.param .u64 vec_fmodf_param_2,
.param .u64 vec_fmodf_param_3
)
{
.reg .pred %p<20>;
.reg .f32 %f<48>;
.reg .b32 %r<22>;
.reg .b64 %rd<14>;
ld.param.u64 %rd5, [vec_fmodf_param_0];
ld.param.u64 %rd2, [vec_fmodf_param_1];
ld.param.u64 %rd3, [vec_fmodf_param_2];
ld.param.u64 %rd4, [vec_fmodf_param_3];
mov.u32 %r3, %tid.x;
mov.u32 %r4, %ntid.x;
mov.u32 %r5, %ctaid.x;
mad.lo.s32 %r6, %r4, %r5, %r3;
cvt.s64.s32 %rd1, %r6;
setp.ge.u64 %p1, %rd1, %rd5;
@%p1 bra BB77_15;
cvta.to.global.u64 %rd6, %rd3;
shl.b64 %rd7, %rd1, 2;
add.s64 %rd8, %rd6, %rd7;
cvta.to.global.u64 %rd9, %rd4;
add.s64 %rd10, %rd9, %rd7;
ld.global.f32 %f1, [%rd8];
abs.f32 %f46, %f1;
ld.global.f32 %f3, [%rd10];
abs.f32 %f4, %f3;
setp.eq.f32 %p2, %f46, 0f7F800000;
setp.eq.f32 %p3, %f4, 0f00000000;
or.pred %p4, %p2, %p3;
mov.f32 %f47, 0f7FFFFFFF;
@%p4 bra BB77_14;
setp.ltu.f32 %p5, %f46, %f4;
@%p5 bra BB77_13;
bra.uni BB77_3;
BB77_13:
setp.gtu.f32 %p18, %f4, 0f7F800000;
add.f32 %f41, %f1, %f3;
selp.f32 %f42, %f41, %f1, %p18;
add.f32 %f43, %f1, %f42;
setp.leu.f32 %p19, %f46, 0f00000000;
selp.f32 %f47, %f43, %f42, %p19;
bra.uni BB77_14;
BB77_3:
lg2.approx.f32 %f21, %f46;
cvt.rzi.s32.f32 %r7, %f21;
lg2.approx.f32 %f22, %f4;
cvt.rzi.s32.f32 %r8, %f22;
sub.s32 %r1, %r7, %r8;
abs.f32 %f5, %f4;
setp.eq.f32 %p6, %f5, 0f00000000;
setp.eq.f32 %p7, %f5, 0f7F800000;
or.pred %p8, %p6, %p7;
setp.eq.s32 %p9, %r7, %r8;
or.pred %p10, %p8, %p9;
@%p10 bra BB77_9;
bra.uni BB77_4;
BB77_9:
setp.leu.f32 %p13, %f5, 0f00000000;
add.f32 %f37, %f4, %f4;
selp.f32 %f44, %f37, %f4, %p13;
bra.uni BB77_10;
BB77_4:
abs.s32 %r2, %r1;
setp.lt.s32 %p11, %r2, 126;
@%p11 bra BB77_8;
bra.uni BB77_5;
BB77_8:
cvt.rn.f32.s32 %f36, %r1;
// inline asm
ex2.approx.ftz.f32 %f35,%f36;
// inline asm
mul.f32 %f44, %f4, %f35;
bra.uni BB77_10;
BB77_5:
setp.lt.s32 %p12, %r2, 252;
@%p12 bra BB77_7;
bra.uni BB77_6;
BB77_7:
shr.u32 %r14, %r1, 31;
add.s32 %r15, %r1, %r14;
shr.s32 %r16, %r15, 1;
cvt.rn.f32.s32 %f31, %r16;
// inline asm
ex2.approx.ftz.f32 %f30,%f31;
// inline asm
mul.f32 %f34, %f4, %f30;
sub.s32 %r17, %r1, %r16;
cvt.rn.f32.s32 %f33, %r17;
// inline asm
ex2.approx.ftz.f32 %f32,%f33;
// inline asm
mul.f32 %f44, %f34, %f32;
bra.uni BB77_10;
BB77_6:
shr.s32 %r9, %r1, 31;
shr.u32 %r10, %r9, 30;
add.s32 %r11, %r1, %r10;
shr.s32 %r12, %r11, 2;
cvt.rn.f32.s32 %f24, %r12;
// inline asm
ex2.approx.ftz.f32 %f23,%f24;
// inline asm
mul.f32 %f27, %f4, %f23;
mul.f32 %f28, %f23, %f27;
mul.f32 %f29, %f23, %f28;
mad.lo.s32 %r13, %r12, -3, %r1;
cvt.rn.f32.s32 %f26, %r13;
// inline asm
ex2.approx.ftz.f32 %f25,%f26;
// inline asm
mul.f32 %f44, %f25, %f29;
BB77_10:
mul.f32 %f38, %f46, 0f3F000000;
setp.gtu.f32 %p14, %f44, %f38;
add.f32 %f39, %f44, %f44;
selp.f32 %f45, %f44, %f39, %p14;
setp.ltu.f32 %p15, %f45, %f4;
@%p15 bra BB77_12;
BB77_11:
sub.f32 %f40, %f46, %f45;
setp.ltu.f32 %p16, %f46, %f45;
selp.f32 %f46, %f46, %f40, %p16;
mul.f32 %f45, %f45, 0f3F000000;
setp.ge.f32 %p17, %f45, %f4;
@%p17 bra BB77_11;
BB77_12:
mov.b32 %r18, %f1;
and.b32 %r19, %r18, -2147483648;
mov.b32 %r20, %f46;
or.b32 %r21, %r20, %r19;
mov.b32 %f47, %r21;
BB77_14:
cvta.to.global.u64 %rd11, %rd2;
add.s64 %rd13, %rd11, %rd7;
st.global.f32 [%rd13], %f47;
BB77_15:
ret;
}
// .globl vec_hypotf
.visible .entry vec_hypotf(
.param .u64 vec_hypotf_param_0,
.param .u64 vec_hypotf_param_1,
.param .u64 vec_hypotf_param_2,
.param .u64 vec_hypotf_param_3
)
{
.reg .pred %p<4>;
.reg .f32 %f<17>;
.reg .b32 %r<13>;
.reg .b64 %rd<13>;
ld.param.u64 %rd5, [vec_hypotf_param_0];
ld.param.u64 %rd2, [vec_hypotf_param_1];
ld.param.u64 %rd3, [vec_hypotf_param_2];
ld.param.u64 %rd4, [vec_hypotf_param_3];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p1, %rd1, %rd5;
@%p1 bra BB78_2;
cvta.to.global.u64 %rd6, %rd3;
shl.b64 %rd7, %rd1, 2;
add.s64 %rd8, %rd6, %rd7;
cvta.to.global.u64 %rd9, %rd4;
add.s64 %rd10, %rd9, %rd7;
ld.global.f32 %f1, [%rd8];
abs.f32 %f2, %f1;
mov.b32 %r5, %f2;
ld.global.f32 %f3, [%rd10];
abs.f32 %f4, %f3;
mov.b32 %r6, %f4;
min.s32 %r7, %r6, %r5;
mov.b32 %f5, %r7;
max.s32 %r8, %r5, %r6;
mov.b32 %f6, %r8;
and.b32 %r9, %r8, -33554432;
mov.u32 %r10, 2122317824;
sub.s32 %r11, %r10, %r9;
mov.b32 %f7, %r11;
mul.f32 %f8, %f5, %f7;
mul.f32 %f9, %f6, %f7;
mul.f32 %f10, %f8, %f8;
fma.rn.f32 %f11, %f9, %f9, %f10;
sqrt.rn.f32 %f12, %f11;
add.s32 %r12, %r9, 8388608;
mov.b32 %f13, %r12;
mul.f32 %f14, %f12, %f13;
setp.eq.f32 %p2, %f5, 0f00000000;
selp.f32 %f15, %f6, %f14, %p2;
setp.eq.f32 %p3, %f5, 0f7F800000;
selp.f32 %f16, 0f7F800000, %f15, %p3;
cvta.to.global.u64 %rd11, %rd2;
add.s64 %rd12, %rd11, %rd7;
st.global.f32 [%rd12], %f16;
BB78_2:
ret;
}
// .globl vec_nextafterf
.visible .entry vec_nextafterf(
.param .u64 vec_nextafterf_param_0,
.param .u64 vec_nextafterf_param_1,
.param .u64 vec_nextafterf_param_2,
.param .u64 vec_nextafterf_param_3
)
{
.reg .pred %p<14>;
.reg .f32 %f<11>;
.reg .b32 %r<18>;
.reg .b64 %rd<14>;
ld.param.u64 %rd5, [vec_nextafterf_param_0];
ld.param.u64 %rd2, [vec_nextafterf_param_1];
ld.param.u64 %rd3, [vec_nextafterf_param_2];
ld.param.u64 %rd4, [vec_nextafterf_param_3];
mov.u32 %r3, %tid.x;
mov.u32 %r4, %ntid.x;
mov.u32 %r5, %ctaid.x;
mad.lo.s32 %r6, %r4, %r5, %r3;
cvt.s64.s32 %rd1, %r6;
setp.ge.u64 %p1, %rd1, %rd5;
@%p1 bra BB79_9;
cvta.to.global.u64 %rd6, %rd3;
shl.b64 %rd7, %rd1, 2;
add.s64 %rd8, %rd6, %rd7;
cvta.to.global.u64 %rd9, %rd4;
add.s64 %rd10, %rd9, %rd7;
ld.global.f32 %f1, [%rd8];
mov.b32 %r1, %f1;
ld.global.f32 %f10, [%rd10];
mov.b32 %r2, %f10;
abs.f32 %f7, %f1;
setp.gtu.f32 %p2, %f7, 0f7F800000;
@%p2 bra BB79_7;
abs.f32 %f8, %f10;
setp.gtu.f32 %p3, %f8, 0f7F800000;
@%p3 bra BB79_7;
bra.uni BB79_3;
BB79_7:
add.f32 %f10, %f1, %f10;
bra.uni BB79_8;
BB79_3:
or.b32 %r7, %r2, %r1;
mov.b32 %f9, %r7;
setp.eq.f32 %p4, %f9, 0f00000000;
@%p4 bra BB79_8;
setp.eq.f32 %p5, %f1, 0f00000000;
@%p5 bra BB79_6;
bra.uni BB79_5;
BB79_6:
and.b32 %r16, %r2, -2147483648;
or.b32 %r17, %r16, 1;
mov.b32 %f10, %r17;
bra.uni BB79_8;
BB79_5:
setp.lt.f32 %p6, %f1, %f10;
setp.lt.f32 %p7, %f1, 0f00000000;
and.pred %p8, %p6, %p7;
selp.s32 %r8, -1, 0, %p8;
add.s32 %r9, %r8, %r1;
setp.gt.f32 %p9, %f1, 0f00000000;
and.pred %p10, %p6, %p9;
selp.u32 %r10, 1, 0, %p10;
add.s32 %r11, %r9, %r10;
setp.gt.f32 %p11, %f1, %f10;
and.pred %p12, %p11, %p7;
selp.u32 %r12, 1, 0, %p12;
add.s32 %r13, %r11, %r12;
and.pred %p13, %p11, %p9;
selp.s32 %r14, -1, 0, %p13;
add.s32 %r15, %r13, %r14;
mov.b32 %f10, %r15;
BB79_8:
cvta.to.global.u64 %rd11, %rd2;
add.s64 %rd13, %rd11, %rd7;
st.global.f32 [%rd13], %f10;
BB79_9:
ret;
}
// .globl vec_powf
.visible .entry vec_powf(
.param .u64 vec_powf_param_0,
.param .u64 vec_powf_param_1,
.param .u64 vec_powf_param_2,
.param .u64 vec_powf_param_3
)
{
.reg .pred %p<30>;
.reg .f32 %f<103>;
.reg .b32 %r<24>;
.reg .b64 %rd<14>;
ld.param.u64 %rd6, [vec_powf_param_0];
ld.param.u64 %rd3, [vec_powf_param_1];
ld.param.u64 %rd4, [vec_powf_param_2];
ld.param.u64 %rd5, [vec_powf_param_3];
mov.u32 %r1, %tid.x;
mov.u32 %r2, %ntid.x;
mov.u32 %r3, %ctaid.x;
mad.lo.s32 %r4, %r2, %r3, %r1;
cvt.s64.s32 %rd1, %r4;
setp.ge.u64 %p2, %rd1, %rd6;
@%p2 bra BB80_15;
cvta.to.global.u64 %rd7, %rd4;
cvta.to.global.u64 %rd2, %rd3;
shl.b64 %rd8, %rd1, 2;
add.s64 %rd9, %rd7, %rd8;
cvta.to.global.u64 %rd10, %rd5;
add.s64 %rd11, %rd10, %rd8;
ld.global.f32 %f1, [%rd11];
mul.f32 %f22, %f1, 0f3F000000;
cvt.rzi.f32.f32 %f23, %f22;
fma.rn.f32 %f24, %f23, 0fC0000000, %f1;
abs.f32 %f2, %f24;
ld.global.f32 %f3, [%rd9];
abs.f32 %f4, %f3;
setp.lt.f32 %p3, %f4, 0f00800000;
mul.f32 %f25, %f4, 0f4B800000;
selp.f32 %f26, 0fC3170000, 0fC2FE0000, %p3;
selp.f32 %f27, %f25, %f4, %p3;
mov.b32 %r5, %f27;
and.b32 %r6, %r5, 8388607;
or.b32 %r7, %r6, 1065353216;
mov.b32 %f28, %r7;
shr.u32 %r8, %r5, 23;
cvt.rn.f32.u32 %f29, %r8;
add.f32 %f30, %f26, %f29;
setp.gt.f32 %p4, %f28, 0f3FB504F3;
mul.f32 %f31, %f28, 0f3F000000;
add.f32 %f32, %f30, 0f3F800000;
selp.f32 %f33, %f31, %f28, %p4;
selp.f32 %f34, %f32, %f30, %p4;
add.f32 %f35, %f33, 0fBF800000;
add.f32 %f19, %f33, 0f3F800000;
// inline asm
rcp.approx.ftz.f32 %f18,%f19;
// inline asm
add.f32 %f36, %f35, %f35;
mul.f32 %f37, %f18, %f36;
mul.f32 %f38, %f37, %f37;
mov.f32 %f39, 0f3C4CAF63;
mov.f32 %f40, 0f3B18F0FE;
fma.rn.f32 %f41, %f40, %f38, %f39;
mov.f32 %f42, 0f3DAAAABD;
fma.rn.f32 %f43, %f41, %f38, %f42;
mul.rn.f32 %f44, %f43, %f38;
mul.rn.f32 %f45, %f44, %f37;
sub.f32 %f46, %f35, %f37;
neg.f32 %f47, %f37;
add.f32 %f48, %f46, %f46;
fma.rn.f32 %f49, %f47, %f35, %f48;
mul.rn.f32 %f50, %f18, %f49;
add.f32 %f51, %f45, %f37;
sub.f32 %f52, %f37, %f51;
add.f32 %f53, %f45, %f52;
add.f32 %f54, %f50, %f53;
add.f32 %f55, %f51, %f54;
sub.f32 %f56, %f51, %f55;
add.f32 %f57, %f54, %f56;
mov.f32 %f58, 0f3F317200;
mul.rn.f32 %f59, %f34, %f58;
mov.f32 %f60, 0f35BFBE8E;
mul.rn.f32 %f61, %f34, %f60;
add.f32 %f62, %f59, %f55;
sub.f32 %f63, %f59, %f62;
add.f32 %f64, %f55, %f63;
add.f32 %f65, %f57, %f64;
add.f32 %f66, %f61, %f65;
add.f32 %f67, %f62, %f66;
sub.f32 %f68, %f62, %f67;
add.f32 %f69, %f66, %f68;
abs.f32 %f5, %f1;
setp.gt.f32 %p5, %f5, 0f77F684DF;
mul.f32 %f70, %f1, 0f39000000;
selp.f32 %f71, %f70, %f1, %p5;
mul.rn.f32 %f72, %f71, %f67;
neg.f32 %f73, %f72;
fma.rn.f32 %f74, %f71, %f67, %f73;
fma.rn.f32 %f75, %f71, %f69, %f74;
mov.f32 %f76, 0f00000000;
fma.rn.f32 %f77, %f76, %f67, %f75;
add.rn.f32 %f78, %f72, %f77;
neg.f32 %f79, %f78;
add.rn.f32 %f80, %f72, %f79;
add.rn.f32 %f81, %f80, %f77;
mov.b32 %r9, %f78;
setp.eq.s32 %p6, %r9, 1118925336;
add.s32 %r10, %r9, -1;
mov.b32 %f82, %r10;
add.f32 %f83, %f81, 0f37000000;
selp.f32 %f84, %f82, %f78, %p6;
selp.f32 %f6, %f83, %f81, %p6;
mul.f32 %f85, %f84, 0f3FB8AA3B;
cvt.rzi.f32.f32 %f86, %f85;
mov.f32 %f87, 0fBF317200;
fma.rn.f32 %f88, %f86, %f87, %f84;
mov.f32 %f89, 0fB5BFBE8E;
fma.rn.f32 %f90, %f86, %f89, %f88;
mul.f32 %f21, %f90, 0f3FB8AA3B;
// inline asm
ex2.approx.ftz.f32 %f20,%f21;
// inline asm
add.f32 %f91, %f86, 0f00000000;
ex2.approx.f32 %f92, %f91;
mul.f32 %f93, %f20, %f92;
setp.lt.f32 %p7, %f84, 0fC2D20000;
selp.f32 %f94, 0f00000000, %f93, %p7;
setp.gt.f32 %p8, %f84, 0f42D20000;
selp.f32 %f101, 0f7F800000, %f94, %p8;
setp.eq.f32 %p9, %f101, 0f7F800000;
@%p9 bra BB80_3;
fma.rn.f32 %f101, %f101, %f6, %f101;
BB80_3:
setp.lt.f32 %p10, %f3, 0f00000000;
setp.eq.f32 %p11, %f2, 0f3F800000;
and.pred %p1, %p10, %p11;
mov.b32 %r11, %f101;
xor.b32 %r12, %r11, -2147483648;
mov.b32 %f95, %r12;
selp.f32 %f102, %f95, %f101, %p1;
setp.eq.f32 %p12, %f3, 0f00000000;
@%p12 bra BB80_6;
bra.uni BB80_4;
BB80_6:
add.f32 %f97, %f3, %f3;
mov.b32 %r13, %f97;
selp.b32 %r14, %r13, 0, %p11;
or.b32 %r15, %r14, 2139095040;
setp.lt.f32 %p16, %f1, 0f00000000;
selp.b32 %r16, %r15, %r14, %p16;
mov.b32 %f102, %r16;
bra.uni BB80_7;
BB80_4:
setp.geu.f32 %p13, %f3, 0f00000000;
@%p13 bra BB80_7;
cvt.rzi.f32.f32 %f96, %f1;
setp.neu.f32 %p14, %f96, %f1;
selp.f32 %f102, 0f7FFFFFFF, %f102, %p14;
BB80_7:
add.f32 %f98, %f4, %f5;
mov.b32 %r17, %f98;
setp.lt.s32 %p17, %r17, 2139095040;
@%p17 bra BB80_14;
setp.gtu.f32 %p18, %f4, 0f7F800000;
setp.gtu.f32 %p19, %f5, 0f7F800000;
or.pred %p20, %p18, %p19;
@%p20 bra BB80_13;
bra.uni BB80_9;
BB80_13:
add.f32 %f102, %f3, %f1;
bra.uni BB80_14;
BB80_9:
setp.eq.f32 %p21, %f5, 0f7F800000;
@%p21 bra BB80_12;
bra.uni BB80_10;
BB80_12:
setp.gt.f32 %p24, %f4, 0f3F800000;
selp.b32 %r21, 2139095040, 0, %p24;
xor.b32 %r22, %r21, 2139095040;
setp.lt.f32 %p25, %f1, 0f00000000;
selp.b32 %r23, %r22, %r21, %p25;
mov.b32 %f99, %r23;
setp.eq.f32 %p26, %f3, 0fBF800000;
selp.f32 %f102, 0f3F800000, %f99, %p26;
bra.uni BB80_14;
BB80_10:
setp.neu.f32 %p22, %f4, 0f7F800000;
@%p22 bra BB80_14;
setp.ltu.f32 %p23, %f1, 0f00000000;
selp.b32 %r18, 0, 2139095040, %p23;
or.b32 %r19, %r18, -2147483648;
selp.b32 %r20, %r19, %r18, %p1;
mov.b32 %f102, %r20;
BB80_14:
setp.eq.f32 %p27, %f1, 0f00000000;
setp.eq.f32 %p28, %f3, 0f3F800000;
or.pred %p29, %p28, %p27;
selp.f32 %f100, 0f3F800000, %f102, %p29;
add.s64 %rd13, %rd2, %rd8;
st.global.f32 [%rd13], %f100;
BB80_15:
ret;
}
// .globl vec_remainderf
.visible .entry vec_remainderf(
.param .u64 vec_remainderf_param_0,
.param .u64 vec_remainderf_param_1,
.param .u64 vec_remainderf_param_2,
.param .u64 vec_remainderf_param_3
)
{
.reg .pred %p<29>;
.reg .f32 %f<52>;
.reg .b32 %r<27>;
.reg .b64 %rd<14>;
ld.param.u64 %rd5, [vec_remainderf_param_0];
ld.param.u64 %rd2, [vec_remainderf_param_1];
ld.param.u64 %rd3, [vec_remainderf_param_2];
ld.param.u64 %rd4, [vec_remainderf_param_3];
mov.u32 %r5, %tid.x;
mov.u32 %r6, %ntid.x;
mov.u32 %r7, %ctaid.x;
mad.lo.s32 %r8, %r6, %r7, %r5;
cvt.s64.s32 %rd1, %r8;
setp.ge.u64 %p3, %rd1, %rd5;
@%p3 bra BB81_20;
cvta.to.global.u64 %rd6, %rd3;
shl.b64 %rd7, %rd1, 2;
add.s64 %rd8, %rd6, %rd7;
cvta.to.global.u64 %rd9, %rd4;
add.s64 %rd10, %rd9, %rd7;
ld.global.f32 %f1, [%rd8];
abs.f32 %f2, %f1;
ld.global.f32 %f3, [%rd10];
abs.f32 %f4, %f3;
setp.gtu.f32 %p4, %f2, 0f7F800000;
setp.gtu.f32 %p5, %f4, 0f7F800000;
or.pred %p6, %p4, %p5;
@%p6 bra BB81_18;
bra.uni BB81_2;
BB81_18:
add.f32 %f51, %f1, %f3;
bra.uni BB81_19;
BB81_2:
setp.eq.f32 %p7, %f2, 0f7F800000;
setp.eq.f32 %p8, %f4, 0f00000000;
or.pred %p9, %p7, %p8;
mov.f32 %f51, 0f7FFFFFFF;
@%p9 bra BB81_19;
setp.ltu.f32 %p10, %f2, %f4;
mov.u32 %r26, 0;
mov.f32 %f49, %f2;
@%p10 bra BB81_15;
lg2.approx.f32 %f22, %f2;
cvt.rzi.s32.f32 %r10, %f22;
lg2.approx.f32 %f23, %f4;
cvt.rzi.s32.f32 %r11, %f23;
sub.s32 %r1, %r10, %r11;
abs.f32 %f5, %f4;
setp.eq.f32 %p11, %f5, 0f00000000;
setp.eq.f32 %p12, %f5, 0f7F800000;
or.pred %p13, %p11, %p12;
setp.eq.s32 %p14, %r10, %r11;
or.pred %p15, %p13, %p14;
@%p15 bra BB81_10;
bra.uni BB81_5;
BB81_10:
setp.leu.f32 %p18, %f5, 0f00000000;
add.f32 %f38, %f4, %f4;
selp.f32 %f44, %f38, %f4, %p18;
bra.uni BB81_11;
BB81_5:
abs.s32 %r2, %r1;
setp.lt.s32 %p16, %r2, 126;
@%p16 bra BB81_9;
bra.uni BB81_6;
BB81_9:
cvt.rn.f32.s32 %f37, %r1;
// inline asm
ex2.approx.ftz.f32 %f36,%f37;
// inline asm
mul.f32 %f44, %f4, %f36;
bra.uni BB81_11;
BB81_6:
setp.lt.s32 %p17, %r2, 252;
@%p17 bra BB81_8;
bra.uni BB81_7;
BB81_8:
shr.u32 %r17, %r1, 31;
add.s32 %r18, %r1, %r17;
shr.s32 %r19, %r18, 1;
cvt.rn.f32.s32 %f32, %r19;
// inline asm
ex2.approx.ftz.f32 %f31,%f32;
// inline asm
mul.f32 %f35, %f4, %f31;
sub.s32 %r20, %r1, %r19;
cvt.rn.f32.s32 %f34, %r20;
// inline asm
ex2.approx.ftz.f32 %f33,%f34;
// inline asm
mul.f32 %f44, %f35, %f33;
bra.uni BB81_11;
BB81_7:
shr.s32 %r12, %r1, 31;
shr.u32 %r13, %r12, 30;
add.s32 %r14, %r1, %r13;
shr.s32 %r15, %r14, 2;
cvt.rn.f32.s32 %f25, %r15;
// inline asm
ex2.approx.ftz.f32 %f24,%f25;
// inline asm
mul.f32 %f28, %f4, %f24;
mul.f32 %f29, %f24, %f28;
mul.f32 %f30, %f24, %f29;
mad.lo.s32 %r16, %r15, -3, %r1;
cvt.rn.f32.s32 %f27, %r16;
// inline asm
ex2.approx.ftz.f32 %f26,%f27;
// inline asm
mul.f32 %f44, %f26, %f30;
BB81_11:
mul.f32 %f39, %f2, 0f3F000000;
setp.gtu.f32 %p19, %f44, %f39;
add.f32 %f40, %f44, %f44;
selp.f32 %f45, %f44, %f40, %p19;
setp.ltu.f32 %p20, %f45, %f4;
mov.f32 %f46, %f2;
mov.f32 %f49, %f46;
@%p20 bra BB81_15;
mov.f32 %f50, %f2;
BB81_13:
mov.f32 %f12, %f50;
mov.f32 %f13, %f45;
sub.f32 %f41, %f12, %f13;
setp.ltu.f32 %p21, %f12, %f13;
selp.f32 %f50, %f12, %f41, %p21;
mul.f32 %f45, %f13, 0f3F000000;
setp.ge.f32 %p22, %f45, %f4;
@%p22 bra BB81_13;
setp.ge.f32 %p23, %f12, %f13;
selp.u32 %r26, 1, 0, %p23;
mov.f32 %f49, %f50;
BB81_15:
add.f32 %f17, %f49, %f49;
setp.gt.f32 %p25, %f17, %f4;
mov.pred %p28, -1;
@%p25 bra BB81_17;
setp.eq.f32 %p26, %f17, %f4;
setp.ne.s32 %p27, %r26, 0;
and.pred %p28, %p26, %p27;
BB81_17:
sub.f32 %f42, %f49, %f4;
selp.f32 %f43, %f42, %f49, %p28;
mov.b32 %r22, %f1;
and.b32 %r23, %r22, -2147483648;
mov.b32 %r24, %f43;
xor.b32 %r25, %r24, %r23;
mov.b32 %f51, %r25;
BB81_19:
cvta.to.global.u64 %rd11, %rd2;
add.s64 %rd13, %rd11, %rd7;
st.global.f32 [%rd13], %f51;
BB81_20:
ret;
}
© 2015 - 2025 Weber Informatics LLC | Privacy Policy