Skip to content
New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

[PH] update calc threads min to give to plugins to the plugin default. #761

Merged
merged 1 commit into from
Mar 1, 2023
Merged

[PH] update calc threads min to give to plugins to the plugin default. #761

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
18 changes: 9 additions & 9 deletions tests/performance_tests/README.md
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -323,19 +323,19 @@ Performance Harness:
* `--skip-tps-test` Determines whether to skip the max TPS measurement tests (default: False)
* `--calc-producer-threads {none,lmax,full}`
Determines whether to calculate number of worker threads to use in producer thread pool ("none", "lmax", or "full").
In "none" mode, the default, no calculation will be attempted and default configured --producer-threads value will be used.
In "lmax" mode, producer threads will incrementally be tested until the performance rate ceases to increase with the addition of additional threads.
In "full" mode producer threads will incrementally be tested from 2..num logical processors, recording each performance and choosing the local max performance (same value as would be discovered in "lmax" mode). Useful for graphing the full performance impact of each available thread. (default: none)
In "none" mode, the default, no calculation will be attempted and the configured --producer-threads value will be used.
In "lmax" mode, producer threads will incrementally be tested, starting at plugin default, until the performance rate ceases to increase with the addition of additional threads.
In "full" mode producer threads will incrementally be tested from plugin default..num logical processors, recording each performance and choosing the local max performance (same value as would be discovered in "lmax" mode). Useful for graphing the full performance impact of each available thread. (default: none)
* `--calc-chain-threads {none,lmax,full}`
Determines whether to calculate number of worker threads to use in chain thread pool ("none", "lmax", or "full").
In "none" mode, the default, no calculation will be attempted and default configured --chain-threads value will be used.
In "lmax" mode, producer threads will incrementally be tested until the performance rate ceases to increase with the addition of additional threads.
In "full" mode producer threads will incrementally be tested from 2..num logical processors, recording each performance and choosing the local max performance (same value as would be discovered in "lmax" mode). Useful for graphing the full performance impact of each available thread. (default: none)
In "none" mode, the default, no calculation will be attempted and the configured --chain-threads value will be used.
In "lmax" mode, producer threads will incrementally be tested, starting at plugin default, until the performance rate ceases to increase with the addition of additional threads.
In "full" mode producer threads will incrementally be tested from plugin default..num logical processors, recording each performance and choosing the local max performance (same value as would be discovered in "lmax" mode). Useful for graphing the full performance impact of each available thread. (default: none)
* `--calc-net-threads {none,lmax,full}`
Determines whether to calculate number of worker threads to use in net thread pool ("none", "lmax", or "full").
In "none" mode, the default, no calculation will be attempted and default configured --net-threads value will be used.
In "lmax" mode, producer threads will incrementally be tested until the performance rate ceases to increase with the addition of additional threads.
In "full" mode producer threads will incrementally be tested from 2..num logical processors, recording each performance and choosing the local max performance (same value as would be discovered in "lmax" mode). Useful for graphing the full performance impact of each available thread. (default: none)
In "none" mode, the default, no calculation will be attempted and the configured --net-threads value will be used.
In "lmax" mode, producer threads will incrementally be tested, starting at plugin default, until the performance rate ceases to increase with the addition of additional threads.
In "full" mode producer threads will incrementally be tested from plugin default..num logical processors, recording each performance and choosing the local max performance (same value as would be discovered in "lmax" mode). Useful for graphing the full performance impact of each available thread. (default: none)
* `--del-test-report` Whether to save json reports from each test scenario. (default: False)

Performance Harness - TPS Test Config:
Expand Down
27 changes: 15 additions & 12 deletions tests/performance_tests/performance_test.py
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -391,7 +391,8 @@ def runTest(self):
optType = PerformanceTest.PluginThreadOptRunType.FULL
else:
optType = PerformanceTest.PluginThreadOptRunType.LOCAL_MAX
prodResults = self.optimizePluginThreadCount(optPlugin=PerformanceTest.PluginThreadOpt.PRODUCER, optType=optType)
prodResults = self.optimizePluginThreadCount(optPlugin=PerformanceTest.PluginThreadOpt.PRODUCER, optType=optType,
minThreadCount=self.clusterConfig.extraNodeosArgs.producerPluginArgs._producerThreadsNodeosDefault)
print(f"Producer Thread Optimization results: {prodResults}")
self.clusterConfig.extraNodeosArgs.producerPluginArgs.threads = prodResults.recommendedThreadCount

Expand All @@ -402,7 +403,8 @@ def runTest(self):
optType = PerformanceTest.PluginThreadOptRunType.FULL
else:
optType = PerformanceTest.PluginThreadOptRunType.LOCAL_MAX
chainResults = self.optimizePluginThreadCount(optPlugin=PerformanceTest.PluginThreadOpt.CHAIN, optType=optType)
chainResults = self.optimizePluginThreadCount(optPlugin=PerformanceTest.PluginThreadOpt.CHAIN, optType=optType,
minThreadCount=self.clusterConfig.extraNodeosArgs.chainPluginArgs._chainThreadsNodeosDefault)
print(f"Chain Thread Optimization results: {chainResults}")
self.clusterConfig.extraNodeosArgs.chainPluginArgs.threads = chainResults.recommendedThreadCount

Expand All @@ -413,7 +415,8 @@ def runTest(self):
optType = PerformanceTest.PluginThreadOptRunType.FULL
else:
optType = PerformanceTest.PluginThreadOptRunType.LOCAL_MAX
netResults = self.optimizePluginThreadCount(optPlugin=PerformanceTest.PluginThreadOpt.NET, optType=optType)
netResults = self.optimizePluginThreadCount(optPlugin=PerformanceTest.PluginThreadOpt.NET, optType=optType,
minThreadCount=self.clusterConfig.extraNodeosArgs.netPluginArgs._netThreadsNodeosDefault)
print(f"Net Thread Optimization results: {netResults}")
self.clusterConfig.extraNodeosArgs.netPluginArgs.threads = netResults.recommendedThreadCount

Expand Down Expand Up @@ -452,21 +455,21 @@ def createArgumentParser():
ptParserGroup = ptParser.add_argument_group(title=ptGrpTitle, description=ptGrpDescription)
ptParserGroup.add_argument("--skip-tps-test", help="Determines whether to skip the max TPS measurement tests", action='store_true')
ptParserGroup.add_argument("--calc-producer-threads", type=str, help="Determines whether to calculate number of worker threads to use in producer thread pool (\"none\", \"lmax\", or \"full\"). \
In \"none\" mode, the default, no calculation will be attempted and default configured --producer-threads value will be used. \
In \"lmax\" mode, producer threads will incrementally be tested until the performance rate ceases to increase with the addition of additional threads. \
In \"full\" mode producer threads will incrementally be tested from 2..num logical processors, recording each performance and choosing the local max performance (same value as would be discovered in \"lmax\" mode). \
In \"none\" mode, the default, no calculation will be attempted and the configured --producer-threads value will be used. \
In \"lmax\" mode, producer threads will incrementally be tested, starting at plugin default, until the performance rate ceases to increase with the addition of additional threads. \
In \"full\" mode producer threads will incrementally be tested from plugin default..num logical processors, recording each performance and choosing the local max performance (same value as would be discovered in \"lmax\" mode). \
Useful for graphing the full performance impact of each available thread.",
choices=["none", "lmax", "full"], default="none")
ptParserGroup.add_argument("--calc-chain-threads", type=str, help="Determines whether to calculate number of worker threads to use in chain thread pool (\"none\", \"lmax\", or \"full\"). \
In \"none\" mode, the default, no calculation will be attempted and default configured --chain-threads value will be used. \
In \"lmax\" mode, producer threads will incrementally be tested until the performance rate ceases to increase with the addition of additional threads. \
In \"full\" mode producer threads will incrementally be tested from 2..num logical processors, recording each performance and choosing the local max performance (same value as would be discovered in \"lmax\" mode). \
In \"none\" mode, the default, no calculation will be attempted and the configured --chain-threads value will be used. \
In \"lmax\" mode, producer threads will incrementally be tested, starting at plugin default, until the performance rate ceases to increase with the addition of additional threads. \
In \"full\" mode producer threads will incrementally be tested from plugin default..num logical processors, recording each performance and choosing the local max performance (same value as would be discovered in \"lmax\" mode). \
Useful for graphing the full performance impact of each available thread.",
choices=["none", "lmax", "full"], default="none")
ptParserGroup.add_argument("--calc-net-threads", type=str, help="Determines whether to calculate number of worker threads to use in net thread pool (\"none\", \"lmax\", or \"full\"). \
In \"none\" mode, the default, no calculation will be attempted and default configured --net-threads value will be used. \
In \"lmax\" mode, producer threads will incrementally be tested until the performance rate ceases to increase with the addition of additional threads. \
In \"full\" mode producer threads will incrementally be tested from 2..num logical processors, recording each performance and choosing the local max performance (same value as would be discovered in \"lmax\" mode). \
In \"none\" mode, the default, no calculation will be attempted and the configured --net-threads value will be used. \
In \"lmax\" mode, producer threads will incrementally be tested, starting at plugin default, until the performance rate ceases to increase with the addition of additional threads. \
In \"full\" mode producer threads will incrementally be tested from plugin default..num logical processors, recording each performance and choosing the local max performance (same value as would be discovered in \"lmax\" mode). \
Useful for graphing the full performance impact of each available thread.",
choices=["none", "lmax", "full"], default="none")
ptParserGroup.add_argument("--del-test-report", help="Whether to save json reports from each test scenario.", action='store_true')
Expand Down