From 955648ca1b015c384296f21ad0a6e61b73cad077 Mon Sep 17 00:00:00 2001 From: Igor Date: Tue, 8 Oct 2024 10:01:22 -0700 Subject: [PATCH] [single node perf] Recalibrate and improve regression perf test Recalibrate for RG change Update limits to be based on min_ratio / max_ratio of many runs update module working set to 100 --- .github/workflows/execution-performance.yaml | 2 +- .../workflow-run-execution-performance.yaml | 53 ++-- testsuite/forge_test.py | 3 +- testsuite/single_node_performance.py | 245 +++++++++++------- 4 files changed, 194 insertions(+), 109 deletions(-) diff --git a/.github/workflows/execution-performance.yaml b/.github/workflows/execution-performance.yaml index 37064dfba35caf..1597d9ecbc8632 100644 --- a/.github/workflows/execution-performance.yaml +++ b/.github/workflows/execution-performance.yaml @@ -23,6 +23,6 @@ jobs: GIT_SHA: ${{ github.event.pull_request.head.sha || github.sha }} RUNNER_NAME: executor-benchmark-runner # Run all tests only on the scheduled cadence, or explicitly requested - IS_FULL_RUN: ${{ github.event_name == 'schedule' || contains(github.event.pull_request.labels.*.name, 'CICD:run-execution-performance-full-test') }} + FLOW: ${{ (github.event_name == 'schedule' || contains(github.event.pull_request.labels.*.name, 'CICD:run-execution-performance-full-test')) && 'CONTINUOUS' || 'LAND_BLOCKING' }} # Ignore target determination if on the scheduled cadence, or explicitly requested IGNORE_TARGET_DETERMINATION: ${{ github.event_name == 'schedule' || contains(github.event.pull_request.labels.*.name, 'CICD:run-execution-performance-test') || contains(github.event.pull_request.labels.*.name, 'CICD:run-execution-performance-full-test') }} diff --git a/.github/workflows/workflow-run-execution-performance.yaml b/.github/workflows/workflow-run-execution-performance.yaml index 086b59b700d3db..7bc43b1e9b7de1 100644 --- a/.github/workflows/workflow-run-execution-performance.yaml +++ b/.github/workflows/workflow-run-execution-performance.yaml @@ -12,16 +12,20 @@ on: required: false default: executor-benchmark-runner type: string - IS_FULL_RUN: + FLOW: required: false - default: false - type: boolean - description: Run complete version of the tests + default: CONTINUOUS + type: string + description: Which set of tests to run. IGNORE_TARGET_DETERMINATION: required: false default: false type: boolean description: Ignore target determination and run the tests + SOURCE: + required: false + default: CI + type: string # This allows the workflow to be triggered manually from the Github UI or CLI # NOTE: because the "number" type is not supported, we default to 720 minute timeout workflow_dispatch: @@ -36,18 +40,34 @@ on: type: choice options: - executor-benchmark-runner - description: The name of the runner to use for the test. - IS_FULL_RUN: + - benchmark-t2d-32 + - benchmark-t2d-60 + - benchmark-c3d-30 + - benchmark-n4-32 + - benchmark-c4-32 + description: The name of the runner to use for the test. (which decides machine specs) + FLOW: required: false - default: false - type: boolean - description: Run complete version of the tests + default: LAND_BLOCKING + options: + - LAND_BLOCKING + - CONTINUOUS + - MAINNET + - MAINNET_LARGE_DB + type: choice + description: Which set of tests to run. MAINNET/MAINNET_LARGE_DB are for performance validation of mainnet nodes. IGNORE_TARGET_DETERMINATION: required: false - default: false + default: true type: boolean description: Ignore target determination and run the tests - + SOURCE: + required: false + default: ADHOC + options: + - ADHOC + type: choice + description: Test source (always adhoc from here) jobs: # This job determines which tests to run test-target-determinator: @@ -63,7 +83,7 @@ jobs: # Run single node execution performance tests single-node-performance: needs: test-target-determinator - timeout-minutes: 60 + timeout-minutes: 120 runs-on: ${{ inputs.RUNNER_NAME }} steps: - uses: actions/checkout@v4 @@ -78,13 +98,8 @@ jobs: - name: Run single node execution benchmark in performance build mode shell: bash - run: TABULATE_INSTALL=lib-only pip install tabulate && testsuite/single_node_performance.py - if: ${{ !inputs.IS_FULL_RUN && (inputs.IGNORE_TARGET_DETERMINATION || needs.test-target-determinator.outputs.run_execution_performance_test == 'true') }} - - - name: Run full version of the single node execution benchmark in performance build mode - shell: bash - run: TABULATE_INSTALL=lib-only pip install tabulate && FLOW=CONTINUOUS testsuite/single_node_performance.py - if: ${{ inputs.IS_FULL_RUN && (inputs.IGNORE_TARGET_DETERMINATION || needs.test-target-determinator.outputs.run_execution_performance_test == 'true') }} + run: TABULATE_INSTALL=lib-only pip install tabulate && FLOW=${{ inputs.FLOW }} SOURCE=${{ inputs.SOURCE }} testsuite/single_node_performance.py + if: ${{ (inputs.IGNORE_TARGET_DETERMINATION || needs.test-target-determinator.outputs.run_execution_performance_test == 'true') }} - run: echo "Skipping single node execution performance! Unrelated changes detected." if: ${{ !inputs.IGNORE_TARGET_DETERMINATION && needs.test-target-determinator.outputs.run_execution_performance_test != 'true' }} diff --git a/testsuite/forge_test.py b/testsuite/forge_test.py index 5e464c58936425..76d993a32cd173 100644 --- a/testsuite/forge_test.py +++ b/testsuite/forge_test.py @@ -64,7 +64,8 @@ class HasAssertMultiLineEqual(Protocol): - def assertMultiLineEqual(self, first: str, second: str, msg: Any = ...) -> None: ... + def assertMultiLineEqual(self, first: str, second: str, msg: Any = ...) -> None: + ... def get_cwd() -> Path: diff --git a/testsuite/single_node_performance.py b/testsuite/single_node_performance.py index 4c6ec760ac5c13..437dfb105cb511 100755 --- a/testsuite/single_node_performance.py +++ b/testsuite/single_node_performance.py @@ -37,7 +37,13 @@ class Flow(Flag): LAND_BLOCKING_AND_C = Flow.LAND_BLOCKING | Flow.CONTINUOUS SELECTED_FLOW = Flow[os.environ.get("FLOW", default="LAND_BLOCKING")] + +print(f"Executing flow: {SELECTED_FLOW}") IS_MAINNET = SELECTED_FLOW in [Flow.MAINNET, Flow.MAINNET_LARGE_DB] +SOURCE = os.environ.get("SOURCE", default="LOCAL") +if SOURCE not in ["ADHOC", "CI", "LOCAL"]: + print(f"Unrecogznied source {SOURCE}") + exit(1) DEFAULT_NUM_INIT_ACCOUNTS = ( "100000000" if SELECTED_FLOW == Flow.MAINNET_LARGE_DB else "2000000" @@ -143,58 +149,60 @@ class RunGroupConfig: # 0-indexed CALIBRATED_TPS_INDEX = -1 +CALIBRATED_COUNT_INDEX = -4 +CALIBRATED_MIN_RATIO_INDEX = -3 +CALIBRATED_MAX_RATIO_INDEX = -2 CALIBRATION_SEPARATOR = " " -# transaction_type module_working_set_size executor_type min_ratio max_ratio median -# (or if from log: -# transaction_type module_working_set_size executor_type block_size expected_tps tps -# ) +# transaction_type module_working_set_size executor_type count min_ratio max_ratio median CALIBRATION = """ -no-op 1 VM 0.822 1.047 37975.3 -no-op 1000 VM 0.775 1.033 22963.8 -apt-fa-transfer 1 VM 0.770 1.059 27299.5 -account-generation 1 VM 0.735 1.026 22663.8 -account-resource32-b 1 VM 0.718 1.049 33440.0 -modify-global-resource 1 VM 0.868 1.019 2819.9 -modify-global-resource 10 VM 0.877 1.018 17562.1 -publish-package 1 VM 0.944 1.037 143.9 -mix_publish_transfer 1 VM 0.953 1.124 2131.6 -batch100-transfer 1 VM 0.768 1.027 770.7 -vector-picture30k 1 VM 0.944 1.036 112.2 -vector-picture30k 20 VM 0.835 1.020 1140.7 -smart-table-picture30-k-with200-change 1 VM 0.955 1.051 21.8 -smart-table-picture30-k-with200-change 20 VM 0.926 1.065 185.9 -modify-global-resource-agg-v2 1 VM 0.792 1.060 32740.0 -modify-global-flag-agg-v2 1 VM 0.921 1.014 5199.3 -modify-global-bounded-agg-v2 1 VM 0.906 1.103 8866.4 -modify-global-milestone-agg-v2 1 VM 0.804 1.033 27699.5 -resource-groups-global-write-tag1-kb 1 VM 0.915 1.074 9039.0 -resource-groups-global-write-and-read-tag1-kb 1 VM 0.938 1.016 6221.0 -resource-groups-sender-write-tag1-kb 1 VM 0.835 1.134 19680.6 -resource-groups-sender-multi-change1-kb 1 VM 0.896 1.071 16553.6 -token-v1ft-mint-and-transfer 1 VM 0.894 1.029 1276.2 -token-v1ft-mint-and-transfer 20 VM 0.897 1.024 11901.1 -token-v1nft-mint-and-transfer-sequential 1 VM 0.923 1.025 798.6 -token-v1nft-mint-and-transfer-sequential 20 VM 0.873 1.024 7732.8 -coin-init-and-mint 1 VM 0.779 1.055 29251.9 -coin-init-and-mint 20 VM 0.827 1.077 24185.0 -fungible-asset-mint 1 VM 0.773 1.023 23274.5 -fungible-asset-mint 20 VM 0.803 1.047 21567.9 -no-op5-signers 1 VM 0.854 1.078 37561.3 -token-v2-ambassador-mint 1 VM 0.848 1.022 15753.6 -token-v2-ambassador-mint 20 VM 0.811 1.044 16228.8 -liquidity-pool-swap 1 VM 0.922 1.027 975.7 -liquidity-pool-swap 20 VM 0.881 1.014 8359.6 -liquidity-pool-swap-stable 1 VM 0.890 1.013 957.5 -liquidity-pool-swap-stable 20 VM 0.916 1.019 8035.3 -deserialize-u256 1 VM 0.842 1.060 37561.3 -no-op-fee-payer 1 VM 0.908 1.029 2131.6 -no-op-fee-payer 50 VM 0.890 1.038 27205.9 +no-op 1 VM 36 0.827 1.118 36723.0 +no-op 1000 VM 36 0.803 1.030 22352.6 +apt-fa-transfer 1 VM 36 0.858 1.060 28198.5 +account-generation 1 VM 36 0.863 1.046 22960.6 +account-resource32-b 1 VM 36 0.852 1.087 34327.5 +modify-global-resource 1 VM 36 0.890 1.023 2799.1 +modify-global-resource 100 VM 36 0.871 1.019 34327.5 +publish-package 1 VM 36 0.967 1.074 142.9 +mix_publish_transfer 1 VM 36 0.957 1.134 2145.5 +batch100-transfer 1 VM 36 0.862 1.024 743.6 +vector-picture30k 1 VM 36 0.973 1.018 112.2 +vector-picture30k 100 VM 36 0.826 1.026 1862.3 +smart-table-picture30-k-with200-change 1 VM 36 0.972 1.078 21.5 +smart-table-picture30-k-with200-change 100 VM 36 0.955 1.064 368.5 +modify-global-resource-agg-v2 1 VM 36 0.906 1.107 35479.7 +modify-global-flag-agg-v2 1 VM 36 0.969 1.023 5508.5 +modify-global-bounded-agg-v2 1 VM 36 0.909 1.085 9876.8 +modify-global-milestone-agg-v2 1 VM 36 0.872 1.037 28612.4 +resource-groups-global-write-tag1-kb 1 VM 36 0.889 1.044 9215.7 +resource-groups-global-write-and-read-tag1-kb 1 VM 36 0.917 1.018 6196.8 +resource-groups-sender-write-tag1-kb 1 VM 36 0.898 1.118 19644.1 +resource-groups-sender-multi-change1-kb 1 VM 36 0.912 1.083 16047.2 +token-v1ft-mint-and-transfer 1 VM 36 0.888 1.040 1264.5 +token-v1ft-mint-and-transfer 100 VM 36 0.897 1.024 17774 +token-v1nft-mint-and-transfer-sequential 1 VM 36 0.893 1.019 798.4 +token-v1nft-mint-and-transfer-sequential 100 VM 36 0.885 1.022 12796.9 +coin-init-and-mint 1 VM 36 0.788 1.071 28664.7 +coin-init-and-mint 100 VM 36 0.787 1.094 24092 +fungible-asset-mint 1 VM 36 0.775 1.034 26523.6 +fungible-asset-mint 100 VM 36 0.780 1.063 21446.3 +no-op5-signers 1 VM 36 0.813 1.105 38063.3 +token-v2-ambassador-mint 1 VM 36 0.780 1.037 17637.4 +token-v2-ambassador-mint 100 VM 36 0.778 1.045 16466.1 +liquidity-pool-swap 1 VM 36 0.852 1.017 966.8 +liquidity-pool-swap 100 VM 36 0.874 1.021 10977.4 +liquidity-pool-swap-stable 1 VM 36 0.908 1.019 938.1 +liquidity-pool-swap-stable 100 VM 36 0.916 1.016 10761.9 +deserialize-u256 1 VM 36 0.842 1.081 37424.8 +no-op-fee-payer 1 VM 36 0.869 1.018 2116.2 +no-op-fee-payer 100 VM 36 0.824 1.026 27295.8 """ # when adding a new test, add estimated expected_tps to it, as well as waived=True. # And then after a day or two - add calibration result for it above, removing expected_tps/waived fields. +DEFAULT_MODULE_WORKING_SET_SIZE = 100 + TESTS = [ RunGroupConfig(key=RunGroupKey("no-op"), included_in=LAND_BLOCKING_AND_C), RunGroupConfig(key=RunGroupKey("no-op", module_working_set_size=1000), included_in=LAND_BLOCKING_AND_C), @@ -204,7 +212,7 @@ class RunGroupConfig: RunGroupConfig(key=RunGroupKey("account-generation", executor_type="native"), included_in=Flow.CONTINUOUS), RunGroupConfig(key=RunGroupKey("account-resource32-b"), included_in=Flow.CONTINUOUS), RunGroupConfig(key=RunGroupKey("modify-global-resource"), included_in=LAND_BLOCKING_AND_C | Flow.REPRESENTATIVE), - RunGroupConfig(key=RunGroupKey("modify-global-resource", module_working_set_size=10), included_in=Flow.CONTINUOUS), + RunGroupConfig(key=RunGroupKey("modify-global-resource", module_working_set_size=DEFAULT_MODULE_WORKING_SET_SIZE), included_in=Flow.CONTINUOUS), RunGroupConfig(key=RunGroupKey("publish-package"), included_in=LAND_BLOCKING_AND_C | Flow.REPRESENTATIVE), RunGroupConfig(key=RunGroupKey("mix_publish_transfer"), key_extra=RunGroupKeyExtra( transaction_type_override="publish-package apt-fa-transfer", @@ -214,42 +222,42 @@ class RunGroupConfig: RunGroupConfig(key=RunGroupKey("batch100-transfer", executor_type="native"), included_in=Flow.CONTINUOUS), RunGroupConfig(expected_tps=100, key=RunGroupKey("vector-picture40"), included_in=Flow(0), waived=True), - RunGroupConfig(expected_tps=1000, key=RunGroupKey("vector-picture40", module_working_set_size=20), included_in=Flow(0), waived=True), + RunGroupConfig(expected_tps=1000, key=RunGroupKey("vector-picture40", module_working_set_size=DEFAULT_MODULE_WORKING_SET_SIZE), included_in=Flow(0), waived=True), RunGroupConfig(key=RunGroupKey("vector-picture30k"), included_in=LAND_BLOCKING_AND_C), - RunGroupConfig(key=RunGroupKey("vector-picture30k", module_working_set_size=20), included_in=Flow.CONTINUOUS), + RunGroupConfig(key=RunGroupKey("vector-picture30k", module_working_set_size=DEFAULT_MODULE_WORKING_SET_SIZE), included_in=Flow.CONTINUOUS), RunGroupConfig(key=RunGroupKey("smart-table-picture30-k-with200-change"), included_in=LAND_BLOCKING_AND_C), - RunGroupConfig(key=RunGroupKey("smart-table-picture30-k-with200-change", module_working_set_size=20), included_in=Flow.CONTINUOUS), + RunGroupConfig(key=RunGroupKey("smart-table-picture30-k-with200-change", module_working_set_size=DEFAULT_MODULE_WORKING_SET_SIZE), included_in=Flow.CONTINUOUS), # RunGroupConfig(expected_tps=10, key=RunGroupKey("smart-table-picture1-m-with256-change"), included_in=LAND_BLOCKING_AND_C), # RunGroupConfig(expected_tps=40, key=RunGroupKey("smart-table-picture1-m-with256-change", module_working_set_size=20), included_in=Flow.CONTINUOUS), RunGroupConfig(key=RunGroupKey("modify-global-resource-agg-v2"), included_in=Flow.AGG_V2 | LAND_BLOCKING_AND_C), - RunGroupConfig(expected_tps=10000, key=RunGroupKey("modify-global-resource-agg-v2", module_working_set_size=50), included_in=Flow.AGG_V2, waived=True), + RunGroupConfig(expected_tps=10000, key=RunGroupKey("modify-global-resource-agg-v2", module_working_set_size=DEFAULT_MODULE_WORKING_SET_SIZE), included_in=Flow.AGG_V2, waived=True), RunGroupConfig(key=RunGroupKey("modify-global-flag-agg-v2"), included_in=Flow.AGG_V2 | Flow.CONTINUOUS), - RunGroupConfig(expected_tps=10000, key=RunGroupKey("modify-global-flag-agg-v2", module_working_set_size=50), included_in=Flow.AGG_V2, waived=True), + RunGroupConfig(expected_tps=10000, key=RunGroupKey("modify-global-flag-agg-v2", module_working_set_size=DEFAULT_MODULE_WORKING_SET_SIZE), included_in=Flow.AGG_V2, waived=True), RunGroupConfig(key=RunGroupKey("modify-global-bounded-agg-v2"), included_in=Flow.AGG_V2 | Flow.CONTINUOUS), - RunGroupConfig(expected_tps=10000, key=RunGroupKey("modify-global-bounded-agg-v2", module_working_set_size=50), included_in=Flow.AGG_V2, waived=True), + RunGroupConfig(expected_tps=10000, key=RunGroupKey("modify-global-bounded-agg-v2", module_working_set_size=DEFAULT_MODULE_WORKING_SET_SIZE), included_in=Flow.AGG_V2, waived=True), RunGroupConfig(key=RunGroupKey("modify-global-milestone-agg-v2"), included_in=Flow.AGG_V2 | Flow.CONTINUOUS), RunGroupConfig(key=RunGroupKey("resource-groups-global-write-tag1-kb"), included_in=LAND_BLOCKING_AND_C | Flow.RESOURCE_GROUPS), - RunGroupConfig(expected_tps=8000, key=RunGroupKey("resource-groups-global-write-tag1-kb", module_working_set_size=20), included_in=Flow.RESOURCE_GROUPS, waived=True), + RunGroupConfig(expected_tps=8000, key=RunGroupKey("resource-groups-global-write-tag1-kb", module_working_set_size=DEFAULT_MODULE_WORKING_SET_SIZE), included_in=Flow.RESOURCE_GROUPS, waived=True), RunGroupConfig(key=RunGroupKey("resource-groups-global-write-and-read-tag1-kb"), included_in=Flow.CONTINUOUS | Flow.RESOURCE_GROUPS), - RunGroupConfig(expected_tps=8000, key=RunGroupKey("resource-groups-global-write-and-read-tag1-kb", module_working_set_size=20), included_in=Flow.RESOURCE_GROUPS, waived=True), + RunGroupConfig(expected_tps=8000, key=RunGroupKey("resource-groups-global-write-and-read-tag1-kb", module_working_set_size=DEFAULT_MODULE_WORKING_SET_SIZE), included_in=Flow.RESOURCE_GROUPS, waived=True), RunGroupConfig(key=RunGroupKey("resource-groups-sender-write-tag1-kb"), included_in=Flow.CONTINUOUS | Flow.RESOURCE_GROUPS), - RunGroupConfig(expected_tps=8000, key=RunGroupKey("resource-groups-sender-write-tag1-kb", module_working_set_size=20), included_in=Flow.RESOURCE_GROUPS, waived=True), + RunGroupConfig(expected_tps=8000, key=RunGroupKey("resource-groups-sender-write-tag1-kb", module_working_set_size=DEFAULT_MODULE_WORKING_SET_SIZE), included_in=Flow.RESOURCE_GROUPS, waived=True), RunGroupConfig(key=RunGroupKey("resource-groups-sender-multi-change1-kb"), included_in=LAND_BLOCKING_AND_C | Flow.RESOURCE_GROUPS), - RunGroupConfig(expected_tps=8000, key=RunGroupKey("resource-groups-sender-multi-change1-kb", module_working_set_size=20), included_in=Flow.RESOURCE_GROUPS, waived=True), + RunGroupConfig(expected_tps=8000, key=RunGroupKey("resource-groups-sender-multi-change1-kb", module_working_set_size=DEFAULT_MODULE_WORKING_SET_SIZE), included_in=Flow.RESOURCE_GROUPS, waived=True), RunGroupConfig(key=RunGroupKey("token-v1ft-mint-and-transfer"), included_in=Flow.CONTINUOUS), - RunGroupConfig(key=RunGroupKey("token-v1ft-mint-and-transfer", module_working_set_size=20), included_in=Flow.CONTINUOUS), + RunGroupConfig(key=RunGroupKey("token-v1ft-mint-and-transfer", module_working_set_size=DEFAULT_MODULE_WORKING_SET_SIZE), included_in=Flow.CONTINUOUS), RunGroupConfig(key=RunGroupKey("token-v1nft-mint-and-transfer-sequential"), included_in=Flow.CONTINUOUS), - RunGroupConfig(key=RunGroupKey("token-v1nft-mint-and-transfer-sequential", module_working_set_size=20), included_in=Flow.CONTINUOUS), + RunGroupConfig(key=RunGroupKey("token-v1nft-mint-and-transfer-sequential", module_working_set_size=DEFAULT_MODULE_WORKING_SET_SIZE), included_in=Flow.CONTINUOUS), RunGroupConfig(expected_tps=1300, key=RunGroupKey("token-v1nft-mint-and-transfer-parallel"), included_in=Flow(0), waived=True), - RunGroupConfig(expected_tps=5300, key=RunGroupKey("token-v1nft-mint-and-transfer-parallel", module_working_set_size=20), included_in=Flow(0), waived=True), + RunGroupConfig(expected_tps=5300, key=RunGroupKey("token-v1nft-mint-and-transfer-parallel", module_working_set_size=DEFAULT_MODULE_WORKING_SET_SIZE), included_in=Flow(0), waived=True), RunGroupConfig(key=RunGroupKey("coin-init-and-mint", module_working_set_size=1), included_in=Flow.CONTINUOUS), - RunGroupConfig(key=RunGroupKey("coin-init-and-mint", module_working_set_size=20), included_in=Flow.CONTINUOUS), + RunGroupConfig(key=RunGroupKey("coin-init-and-mint", module_working_set_size=DEFAULT_MODULE_WORKING_SET_SIZE), included_in=Flow.CONTINUOUS), RunGroupConfig(key=RunGroupKey("fungible-asset-mint", module_working_set_size=1), included_in=LAND_BLOCKING_AND_C), - RunGroupConfig(key=RunGroupKey("fungible-asset-mint", module_working_set_size=20), included_in=Flow.CONTINUOUS), + RunGroupConfig(key=RunGroupKey("fungible-asset-mint", module_working_set_size=DEFAULT_MODULE_WORKING_SET_SIZE), included_in=Flow.CONTINUOUS), # RunGroupConfig(expected_tps=1000, key=RunGroupKey("token-v1ft-mint-and-store"), included_in=Flow(0)), # RunGroupConfig(expected_tps=1000, key=RunGroupKey("token-v1nft-mint-and-store-sequential"), included_in=Flow(0)), @@ -258,19 +266,19 @@ class RunGroupConfig: RunGroupConfig(key=RunGroupKey("no-op5-signers"), included_in=Flow.CONTINUOUS), RunGroupConfig(key=RunGroupKey("token-v2-ambassador-mint"), included_in=LAND_BLOCKING_AND_C | Flow.REPRESENTATIVE), - RunGroupConfig(key=RunGroupKey("token-v2-ambassador-mint", module_working_set_size=20), included_in=Flow.CONTINUOUS), + RunGroupConfig(key=RunGroupKey("token-v2-ambassador-mint", module_working_set_size=DEFAULT_MODULE_WORKING_SET_SIZE), included_in=Flow.CONTINUOUS), RunGroupConfig(key=RunGroupKey("liquidity-pool-swap"), included_in=LAND_BLOCKING_AND_C | Flow.REPRESENTATIVE), - RunGroupConfig(key=RunGroupKey("liquidity-pool-swap", module_working_set_size=20), included_in=Flow.CONTINUOUS), + RunGroupConfig(key=RunGroupKey("liquidity-pool-swap", module_working_set_size=DEFAULT_MODULE_WORKING_SET_SIZE), included_in=Flow.CONTINUOUS), RunGroupConfig(key=RunGroupKey("liquidity-pool-swap-stable"), included_in=Flow.CONTINUOUS), - RunGroupConfig(key=RunGroupKey("liquidity-pool-swap-stable", module_working_set_size=20), included_in=Flow.CONTINUOUS), + RunGroupConfig(key=RunGroupKey("liquidity-pool-swap-stable", module_working_set_size=DEFAULT_MODULE_WORKING_SET_SIZE), included_in=Flow.CONTINUOUS), RunGroupConfig(key=RunGroupKey("deserialize-u256"), included_in=Flow.CONTINUOUS), # fee payer sequentializes transactions today. in these tests module publisher is the fee payer, so larger number of modules tests throughput with multiple fee payers RunGroupConfig(key=RunGroupKey("no-op-fee-payer"), included_in=LAND_BLOCKING_AND_C), - RunGroupConfig(key=RunGroupKey("no-op-fee-payer", module_working_set_size=50), included_in=Flow.CONTINUOUS), + RunGroupConfig(key=RunGroupKey("no-op-fee-payer", module_working_set_size=DEFAULT_MODULE_WORKING_SET_SIZE), included_in=Flow.CONTINUOUS), RunGroupConfig(expected_tps=50000, key=RunGroupKey("coin_transfer_connected_components", executor_type="sharded"), key_extra=RunGroupKeyExtra(sharding_traffic_flags="--connected-tx-grps 5000", transaction_type_override=""), included_in=Flow.REPRESENTATIVE, waived=True), RunGroupConfig(expected_tps=50000, key=RunGroupKey("coin_transfer_hotspot", executor_type="sharded"), key_extra=RunGroupKeyExtra(sharding_traffic_flags="--hotspot-probability 0.8", transaction_type_override=""), included_in=Flow.REPRESENTATIVE, waived=True), @@ -355,6 +363,23 @@ class RunGroupInstance: expected_tps: float +@dataclass +class CalibrationData: + expected_tps: float + count: int + min_ratio: float + max_ratio: float + + +@dataclass +class Criteria: + expected_tps: float + min_tps: float + min_warn_tps: float + max_tps: float + max_warn_tps: float + + def get_only(values): assert len(values) == 1, "Multiple values parsed: " + str(values) return values[0] @@ -534,7 +559,12 @@ def print_table( transaction_type=parts[0], module_working_set_size=int(parts[1]), executor_type=parts[2], - ): float(parts[CALIBRATED_TPS_INDEX]) + ): CalibrationData( + expected_tps=float(parts[CALIBRATED_TPS_INDEX]), + count=int(parts[CALIBRATED_COUNT_INDEX]), + min_ratio=float(parts[CALIBRATED_MIN_RATIO_INDEX]), + max_ratio=float(parts[CALIBRATED_MAX_RATIO_INDEX]), + ) for line in CALIBRATION.split("\n") if len( parts := [ @@ -574,11 +604,40 @@ def print_table( if test.expected_tps is not None: print(f"WARNING: using uncalibrated TPS for {test.key}") - expected_tps = test.expected_tps + criteria = Criteria( + expected_tps=test.expected_tps, + min_tps=test.expected_tps * NOISE_LOWER_LIMIT, + min_warn_tps=test.expected_tps * NOISE_LOWER_LIMIT_WARN, + max_tps=test.expected_tps * NOISE_UPPER_LIMIT, + max_warn_tps=test.expected_tps * NOISE_UPPER_LIMIT_WARN, + ) else: assert test.key in calibrated_expected_tps, test - expected_tps = calibrated_expected_tps[test.key] - cur_block_size = int(min([expected_tps, MAX_BLOCK_SIZE])) + cur_calibration = calibrated_expected_tps[test.key] + if cur_calibration.count > 20: + criteria = Criteria( + expected_tps=cur_calibration.expected_tps, + min_tps=cur_calibration.expected_tps * cur_calibration.min_ratio, + min_warn_tps=cur_calibration.expected_tps + * pow(cur_calibration.min_ratio, 0.5), + max_tps=cur_calibration.expected_tps * cur_calibration.max_ratio, + max_warn_tps=cur_calibration.expected_tps + * pow(cur_calibration.max_ratio, 0.5), + ) + else: + criteria = Criteria( + expected_tps=cur_calibration.expected_tps, + min_tps=cur_calibration.expected_tps + * (cur_calibration.min_ratio - 0.1), + min_warn_tps=cur_calibration.expected_tps + * min(cur_calibration.min_ratio, 0.95), + max_tps=cur_calibration.expected_tps + * (cur_calibration.max_ratio + 0.1), + max_warn_tps=cur_calibration.expected_tps + * max(cur_calibration.max_ratio, 1.05), + ) + + cur_block_size = int(min([criteria.expected_tps, MAX_BLOCK_SIZE])) print(f"Testing {test.key}") if test.key_extra.transaction_type_override == "": @@ -641,7 +700,7 @@ def print_table( single_node_result=single_node_result, number_of_threads_results=number_of_threads_results, block_size=cur_block_size, - expected_tps=expected_tps, + expected_tps=criteria.expected_tps, ) ) @@ -657,7 +716,7 @@ def print_table( single_node_result=stage_node_result, number_of_threads_results=number_of_threads_results, block_size=cur_block_size, - expected_tps=expected_tps, + expected_tps=criteria.expected_tps, ) ) @@ -665,18 +724,27 @@ def print_table( print( json.dumps( { - "grep": "grep_json_single_node_perf", + "grep": "grep_json_single_node_perf" + if SOURCE == "CI" + else ( + "grep_json_single_node_perf_adhoc" + if SOURCE == "ADHOC" + else "grep_json_single_node_perf_local" + ), "transaction_type": test.key.transaction_type, "module_working_set_size": test.key.module_working_set_size, "executor_type": test.key.executor_type, "block_size": cur_block_size, "execution_threads": NUMBER_OF_EXECUTION_THREADS, - "expected_tps": expected_tps, + "expected_tps": criteria.expected_tps, + "expected_min_tps": criteria.min_tps, + "expected_max_tps": criteria.max_tps, "waived": test.waived, "tps": single_node_result.tps, "gps": single_node_result.gps, "gpt": single_node_result.gpt, "code_perf_version": CODE_PERF_VERSION, + "flow": str(SELECTED_FLOW), "test_index": test_index, } ) @@ -721,39 +789,31 @@ def print_table( ) print_table(results, by_levels=False, single_field=None) - # if expected TPS is not set, skip performance checks - if expected_tps is None: - continue - - if ( - NOISE_LOWER_LIMIT is not None - and single_node_result.tps < expected_tps * NOISE_LOWER_LIMIT - ): - text = f"regression detected {single_node_result.tps} < {expected_tps * NOISE_LOWER_LIMIT} = {expected_tps} * {NOISE_LOWER_LIMIT}, {test.key} didn't meet TPS requirements" + if NOISE_LOWER_LIMIT is not None and single_node_result.tps < criteria.min_tps: + text = f"regression detected {single_node_result.tps} < {criteria.min_tps} (expected median {criteria.expected_tps}), {test.key} didn't meet TPS requirements" if not test.waived: errors.append(text) else: warnings.append(text) elif ( NOISE_LOWER_LIMIT_WARN is not None - and single_node_result.tps < expected_tps * NOISE_LOWER_LIMIT_WARN + and single_node_result.tps < criteria.min_warn_tps ): - text = f"potential (but within normal noise) regression detected {single_node_result.tps} < {expected_tps * NOISE_LOWER_LIMIT_WARN} = {expected_tps} * {NOISE_LOWER_LIMIT_WARN}, {test.key} didn't meet TPS requirements" + text = f"potential (but within normal noise) regression detected {single_node_result.tps} < {criteria.min_warn_tps} (expected median {criteria.expected_tps}), {test.key} didn't meet TPS requirements" warnings.append(text) elif ( - NOISE_UPPER_LIMIT is not None - and single_node_result.tps > expected_tps * NOISE_UPPER_LIMIT + NOISE_UPPER_LIMIT is not None and single_node_result.tps > criteria.max_tps ): - text = f"perf improvement detected {single_node_result.tps} > {expected_tps * NOISE_UPPER_LIMIT} = {expected_tps} * {NOISE_UPPER_LIMIT}, {test.key} exceeded TPS requirements, increase TPS requirements to match new baseline" + text = f"perf improvement detected {single_node_result.tps} > {criteria.max_tps} (expected median {criteria.expected_tps}), {test.key} exceeded TPS requirements, increase TPS requirements to match new baseline" if not test.waived: errors.append(text) else: warnings.append(text) elif ( NOISE_UPPER_LIMIT_WARN is not None - and single_node_result.tps > expected_tps * NOISE_UPPER_LIMIT_WARN + and single_node_result.tps > criteria.max_warn_tps ): - text = f"potential (but within normal noise) perf improvement detected {single_node_result.tps} > {expected_tps * NOISE_UPPER_LIMIT_WARN} = {expected_tps} * {NOISE_UPPER_LIMIT_WARN}, {test.key} exceeded TPS requirements, increase TPS requirements to match new baseline" + text = f"potential (but within normal noise) perf improvement detected {single_node_result.tps} > {criteria.max_warn_tps} (expected median {criteria.expected_tps}), {test.key} exceeded TPS requirements, increase TPS requirements to match new baseline" warnings.append(text) if HIDE_OUTPUT: @@ -766,6 +826,15 @@ def print_table( if errors: print("Errors: ") print("\n".join(errors)) + print( + """If you expect your PR to change the performance, you need to recalibrate the values. +To do so, you should run the test on your branch 6 times +(https://github.com/aptos-labs/aptos-core/actions/workflows/workflow-run-execution-performance.yaml). +Then go to Humio calibration link (https://gist.github.com/igor-aptos/7b12ca28de03894cddda8e415f37889e), +update it to your branch, and export values as CSV, and then open and copy values inside +testsuite/single_node_performance.py testsuite), and add Blockchain oncall as the reviewer. +""" + ) exit(1) if move_e2e_benchmark_failed: