Many resources are needed to download a project. Please understand that we have to compensate our server costs. Thank you in advance. Project price only 1 $
You can buy this project and download/modify it how often you want.
####################################
# Cache Simulator Config File #
####################################
# This is the reference config file that contains all the default settings.
# Make your edits/overrides in your application.conf.
# See https://doc.akka.io//docs/akka/current/general/configuration-reference.html
akka.loglevel = WARNING
caffeine.simulator {
# https://doc.akka.io/docs/akka/current/typed/mailboxes.html
mailbox {
mailbox-capacity = 10
mailbox-push-timeout-time = -1
mailbox-type = akka.dispatch.BoundedMailbox
}
report {
# Formats: table, csv
format = table
# Columns: policy, hit rate, hits, misses, evictions, admit rate, steps, time
sort-by = policy
ascending = true
# The output destination, either the console or a file path
output = console
}
# The seed for randomized operations
random-seed = 1033096058
# The number of events to send per actor message
batch-size = 1000
# The maximum number of entries in the cache
maximum-size = 512
policies = [
# Policies that provide an optimal upper bound
opt.Unbounded,
opt.Clairvoyant,
# Policies based on maintaining a linked-list cross-cutting the hash table
linked.Lru,
linked.Mru,
linked.Lfu,
linked.Mfu,
linked.Fifo,
linked.Clock,
linked.S4Lru,
linked.MultiQueue,
linked.SegmentedLru,
# Policies based on obtaining a random sampling from the hash table
sampled.Lru,
sampled.Mru,
sampled.Lfu,
sampled.Mfu,
sampled.Fifo,
sampled.Random,
sampled.Hyperbolic,
# Policies based on the Greedy-Dual algorithm
greedy-dual.Camp,
greedy-dual.Gdsf,
greedy-dual.GDWheel,
# Policies based on the 2Q algorithm
two-queue.TwoQueue,
two-queue.TuQueue,
# Policies based on a sketch algorithm
sketch.WindowTinyLfu,
sketch.S4WindowTinyLfu,
sketch.LruWindowTinyLfu,
sketch.RandomWindowTinyLfu,
sketch.FullySegmentedWindowTinyLfu,
sketch.HillClimberWindowTinyLfu,
sketch.FeedbackTinyLfu,
sketch.FeedbackWindowTinyLfu,
sketch.TinyCache,
sketch.TinyCache_GhostCache,
sketch.WindowTinyCache,
# Policies based on the LIRS algorithm
irr.Frd,
irr.Lirs,
irr.DClock,
irr.ClockPro,
irr.ClockProPlus,
irr.ClockProSimple,
# Policies based on the ARC algorithm
adaptive.Arc,
adaptive.Car,
adaptive.Cart,
# Caching products
product.OHC,
product.Guava,
product.TCache,
product.Cache2k,
product.Caffeine,
product.Ehcache3,
product.ExpiringMap,
]
# The admission policy (opposite of eviction policy)
admission = [
Always,
TinyLfu,
]
# The membership filter
membership {
# caffeine, guava, fast-filter
filter = caffeine
# The false positive probability
fpp = 0.01
# The multiple of the maximum size indicating the expected number of insertions
expected-insertions-multiplier = 3.0
fast-filter {
# bloom, counting-bloom, succinct-counting-bloom, blocked-bloom, blocked-bloom-v2
# succinct-counting-blocked-bloom, succinct-counting-blocked-bloom-v2,
# succinct-counting-blocked-bloom-ranked-v2,
type = blocked-bloom
# The number of bits per key
bits-per-key = 11
}
}
sampled {
# The random sample size
size = 8
# guess: Chooses items at random until the sample size is reached
# shuffle: http://en.wikipedia.org/wiki/Fisher–Yates_shuffle
# reservoir: http://en.wikipedia.org/wiki/Reservoir_sampling
strategy = guess
}
multi-queue {
# The logical time that an entry can reside idle in a queue before being demoted
lifetime = 75
# The number of queues using a 2^n frequency distribution
num-queues = 8
# The percentage for the OUT queue
percent-out = 0.50
}
segmented-lru {
# The percentage for the PROTECTED queue
percent-protected = 0.80
}
s4lru {
# The number of segments
levels = 4
}
two-queue {
# The percentage for the IN queue
percent-in = 0.20
# The percentage for the OUT queue
percent-out = 0.50
}
tu-queue {
# The percentage for the HOT queue
percent-hot = 0.33
# The percentage for the WARM queue
percent-warm = 0.33
}
tiny-lfu {
# CountMinSketch: count-min-4 (4-bit), count-min-64 (64-bit)
# Table: random-table, tiny-table, perfect-table
sketch = count-min-4
# If increments are conservative by only updating the minimum counters for CountMin sketches
count-min.conservative = false
count-min-64 {
eps = 0.0001
confidence = 0.99
}
count-min-4 {
# periodic: Resets by periodically halving all counters
# adaptive: Resets periodically at an adaptive interval
# incremental: Resets by halving counters in an incremental sweep
reset = periodic
# The multiple of the maximum size determining the number of counters
counters-multiplier = 1.0
incremental {
# The incremental reset interval (the number of additions before halving counters)
interval = 16
}
periodic.doorkeeper {
# When enabled the "counters-multiplier" should be reduced to determine the space savings
enabled = false
}
}
}
feedback-tiny-lfu {
# The maximum emphasis to give newly inserted entries
maximum-insertion-gain = 5
# The maximum size of the sample period
maximum-sample-size = 256
# The false positive probability of the adaptive filter
adaptive-fpp = 0.03
}
window-tiny-lfu {
# The percentage for the MAIN space (PROBATION + PROTECTED)
percent-main = [0.99]
# The percentage for the PROTECTED MAIN queue
percent-main-protected = 0.80
}
lru-window-tiny-lfu {
# The percentage for the MAIN queue
percent-main = [0.99]
}
random-window-tiny-lfu {
# The percentage for the MAIN space
percent-main = [0.99]
}
fully-segmented-window-tiny-lfu {
# The percentage for the MAIN space (PROBATION + PROTECTED)
percent-main = [0.99]
# The percentage for the PROTECTED MAIN queue
percent-main-protected = 0.80
# The percentage for the PROTECTED WINDOW queue
percent-window-protected = 0.80
}
s4-window-tiny-lfu {
# The percentage for the MAIN queue
percent-main = [0.99]
# The number of segments in the MAIN space
levels = 4
}
feedback-window-tiny-lfu {
# The initial percentage for the MAIN space (PROBATION + PROTECTED)
percent-main = [0.99]
# The initial percentage for the PROTECTED MAIN queue
percent-main-protected = 0.80
# The initial percentage of the WINDOW space that can be pivoted
percent-pivot = 0.0
# The amount to increase the window when adapting
pivot-increment = 8
# The amount to decrease the window when adapting
pivot-decrement = 4
# The maximum size of the WINDOW space
maximum-window-size = 256
# The maximum size of the sample period
maximum-sample-size = 1024
# The false positive probability of the adaptive filter
adaptive-fpp = 0.03
}
hill-climber-window-tiny-lfu {
# The initial percentage for the MAIN space (PROBATION + PROTECTED)
percent-main = [0.99]
# The initial percentage for the PROTECTED MAIN queue
percent-main-protected = 0.80
# simple: Moves a fixed amount based on if the current direction had a positive impact
# simulated-annealing: A simple hill climber that cools off, reducing the step size
# stochastic-gradient-descent: Uses the gradient and momentum to walk the curve
# adam, nadam, amsgrad: SGD with adaptive step sizes
# indicator: Computes the best configuration based on a sampled skew
# minisim: Simulates multiple configurations and chooses the best one
strategy = [
simple,
indicator,
]
simple {
# The percent of the total size to adapt the window by
percent-pivot = 0.0625
# The size of the sample period (1.0 = 100%)
percent-sample = 10.0
# The difference in hit rate percentage to tolerate before changing directions
tolerance = 0.0
# The rate to decrease the step size to adapt by
step-decay-rate = 0.98
# The rate to decrease the sampling period
sample-decay-rate = 1.0
# The difference in hit rate percentage to tolerate before restarting the adaption
restart-threshold = 0.05
}
simulated-annealing {
# The percent of the total size to adapt the window by
percent-pivot = 0.0625
# The size of the sample period (1.0 = 100%)
percent-sample = 10.0
# The difference in hit rate to tolerate before cooling down
cool-down-tolerance = 0.0
# The cool down rate
cool-down-rate = 0.9
# The minimum temperature, at which point annealing halts
min-temperature = 0.00001
# The difference in hit rate to tolerate before restarting
restart-tolerance = 0.03
# The chance for a random restart
random-restart = 0.01
}
stochastic-gradient-descent {
# The percent of the total size to adapt the window by
percent-pivot = 0.005
# The size of the sample period (1.0 = 100%)
percent-sample = 0.05
# none, momentum, nesterov
acceleration = momentum
# The force of acceleration
beta = 0.9
}
adam {
# The percent of the total size to adapt the window by
percent-pivot = 0.005
# The size of the sample period (1.0 = 100%)
percent-sample = 0.05
# The decay rate of the momentum
beta1 = 0.9
# The decay rate of the velocity
beta2 = 0.999
# The fuzz factor for stability
epsilon = 1e-8
}
nadam {
# The percent of the total size to adapt the window by
percent-pivot = 0.005
# The size of the sample period (1.0 = 100%)
percent-sample = 0.05
# The decay rate of the momentum
beta1 = 0.9
# The decay rate of the velocity
beta2 = 0.999
# The fuzz factor for stability
epsilon = 1e-8
}
amsgrad {
# The percent of the total size to adapt the window by
percent-pivot = 0.005
# The size of the sample period (1.0 = 100%)
percent-sample = 0.05
# The decay rate of the momentum
beta1 = 0.9
# The decay rate of the velocity
beta2 = 0.999
# The fuzz factor for stability
epsilon = 1e-8
}
minisim {
# The period length of the minisim adaptation
period = 1000000
}
}
indicator {
# Skew estimation is based on the top-k items
k = 70
# The size of the stream summary sketch
ss-size = 1000
}
frd {
# The percentage for the MAIN queue
percent-main = 0.90
# The period length of the indicator adaptation
period = 50000
}
lirs {
# The percentage for the HOT queue
percent-hot = 0.99
# The multiple of the maximum size dedicated to non-resident entries
non-resident-multiplier = 2.0
}
clockpro {
# The percentage for the minimum resident cold entries
percent-min-resident-cold = 0.01
# The percentage for the maximum resident cold entries
percent-max-resident-cold = 0.99
# The lower bound for the number of resident cold entries
lower-bound-resident-cold = 2
# The multiple of the maximum size dedicated to non-resident entries
non-resident-multiplier = 2.0
}
clockproplus {
# The percentage for the minimum resident cold entries
percent-min-resident-cold = 0.01
# The percentage for the maximum resident cold entries
percent-max-resident-cold = 0.5
# The lower bound for the number of resident cold entries
lower-bound-resident-cold = 2
# The multiple of the maximum size dedicated to non-resident entries
non-resident-multiplier = 1.0
}
dclock {
# The percentage for the ACTIVE queue
percent-active = [ 0.5, 0.99 ]
}
expiring-map {
# Policies: Fifo, Lru
policy = lru
}
ohc {
# Policies: Lru, W-TinyLfu
policy = [lru, w-tinylfu]
# The percentage for the EDEN space (admission window)
percent-eden = 0.20
}
tcache {
# Policies: Lru, Lfu
policy = lfu
}
gd-wheel {
# The number of wheels used in the policy
wheels = 2
# The number of queues for each wheel
queues = 256
}
camp {
# Precision parameter
precision = 5
}
trace {
# files: reads from the trace file(s)
# synthetic: reads from a synthetic generator
source = files
# The number of events to skip
skip = 0
# The number of events to process or null if unbounded
limit = null
}
files {
# The paths to the trace files or the file names if in the format's package. To use a mix of
# formats, specify the entry in the form "{format}:{path}", e.g. "lirs:loop.trace.gz".
paths = [ multi1.trace.gz ]
# arc: format from the authors of the ARC algorithm
# adapt-size: format from the authors of the AdaptSize algorithm
# address: format of UCSD program address traces
# address-penalties: format of UCSD program address traces with hit & miss penalties
# cache2k: format from the author of the Cache2k library
# camelab: format of the Camelab storage traces
# cloud-physics: format of the Cloud Physics traces
# corda: format of Corda traces
# gradle: format from the authors of the Gradle build tool
# lirs: format from the authors of the LIRS algorithm
# lrb: format from the authors of the LRB algorithm
# outbrain: format of Outbrain's trace provided on Kaggle
# scarab: format of Scarab Research traces
# snia-cambridge: format of the SNIA MSR Cambridge traces
# snia-k5cloud: format of the SNIA K5cloud traces
# snia-object-store: format of the SNIA IBM ObjectStore traces
# snia-systor: format of the SNIA SYSTOR '17 traces
# snia-tencent-block: format of the SNIA Tencent Block traces
# snia-tencent-photo: format of the SNIA Tencent Photo traces
# tragen: format of the Tragen synthetic trace generator
# twitter: format of the Twitter Cache Cluster traces
# umass-storage: format of the University of Massachusetts storage traces
# umass-youtube: format of the University of Massachusetts youtube traces
# wikipedia: format of the WikiBench request traces
format = lirs
}
synthetic {
# The number of events to generate
events = 10000
# counter, repeating, uniform, exponential, hotspot,
# zipfian, scrambled-zipfian, or skewed-zipfian-latest
distribution = scrambled-zipfian
# A sequence of unique integers starting from...
counter.start = 1
# A sequence of unique integers that repeats
repeating.items = 5000
# A sequence that is generated from the specified set uniformly randomly
uniform {
lower-bound = 1
upper-bound = 1000
}
# A sequence based on an exponential distribution with a mean arrival rate of gamma
exponential.mean = 1.0
# A sequence resembling a hotspot distribution where x% of operations access y% of data items
hotspot {
# The lower bound of the distribution
lower-bound = 1
# The upper bound of the distribution
upper-bound = 1000
# The percentage of the of the interval which comprises the hot set
hotset-fraction = 0.25
# The percentage of operations that access the hot set
hot-opn-fraction = 0.25
}
# A sequence where some items are more popular than others, according to a zipfian distribution
zipfian {
# The number of items
items = 5000
# The skewness factor
constant = 0.99
# A zipfian sequence that scatters the "popular" items across the item space. Use if you don't
# want the head of the distribution (the popular items) clustered together.
scrambled {}
# A zipfian sequence with a popularity distribution of items, skewed to favor recent items
# significantly more than older items
skewed-zipfian-latest {}
}
}
}
