From 25f70099f4b4fde99455f13506be623a1cfb06eb Mon Sep 17 00:00:00 2001 From: Igor Date: Tue, 22 Oct 2024 15:34:55 -0700 Subject: [PATCH] [single node benchmark] recalibrate and make thresholds depend on number samples present --- testsuite/single_node_performance.py | 133 +++++++++++++-------------- 1 file changed, 65 insertions(+), 68 deletions(-) diff --git a/testsuite/single_node_performance.py b/testsuite/single_node_performance.py index af4a94d442e25..0032151853eb6 100755 --- a/testsuite/single_node_performance.py +++ b/testsuite/single_node_performance.py @@ -159,46 +159,46 @@ class RunGroupConfig: # transaction_type module_working_set_size executor_type count min_ratio max_ratio median CALIBRATION = """ -no-op 1 VM 36 0.776 1.077 40464.7 -no-op 1000 VM 36 0.774 1.039 22432.9 -apt-fa-transfer 1 VM 36 0.838 1.059 28928.4 -account-generation 1 VM 36 0.797 1.049 23457.3 -account-resource32-b 1 VM 36 0.809 1.084 35822.6 -modify-global-resource 1 VM 36 0.921 1.015 2799.1 -modify-global-resource 100 VM 36 0.839 1.04 35206.7 -publish-package 1 VM 36 0.886 1.039 147 -mix_publish_transfer 1 VM 36 0.918 1.086 2079.7 -batch100-transfer 1 VM 36 0.727 1.018 740.9 -vector-picture30k 1 VM 36 0.863 1.024 110.6 -vector-picture30k 100 VM 36 0.573 1.027 2025.7 -smart-table-picture30-k-with200-change 1 VM 36 0.962 1.056 21.3 -smart-table-picture30-k-with200-change 100 VM 36 0.902 1.034 404.9 -modify-global-resource-agg-v2 1 VM 36 0.773 1.085 37354.2 -modify-global-flag-agg-v2 1 VM 36 0.923 1.02 5518.4 -modify-global-bounded-agg-v2 1 VM 36 0.888 1.083 9803.9 -modify-global-milestone-agg-v2 1 VM 36 0.811 1.036 28664.7 -resource-groups-global-write-tag1-kb 1 VM 36 0.8 1.047 9180.8 -resource-groups-global-write-and-read-tag1-kb 1 VM 36 0.872 1.021 6218.9 -resource-groups-sender-write-tag1-kb 1 VM 36 0.899 1.198 19644.1 -resource-groups-sender-multi-change1-kb 1 VM 36 0.815 1.084 16531.9 -token-v1ft-mint-and-transfer 1 VM 36 0.835 1.034 1257.6 -token-v1ft-mint-and-transfer 100 VM 36 0.77 1.028 18152.9 -token-v1nft-mint-and-transfer-sequential 1 VM 36 0.872 1.023 792.7 -token-v1nft-mint-and-transfer-sequential 100 VM 36 0.786 1.035 12888.4 -coin-init-and-mint 1 VM 36 0.809 1.077 29520.8 -coin-init-and-mint 100 VM 36 0.776 1.062 24350.5 -fungible-asset-mint 1 VM 36 0.833 1.043 26523.6 -fungible-asset-mint 100 VM 36 0.867 1.045 21562.8 -no-op5-signers 1 VM 36 0.841 1.085 40094.5 -token-v2-ambassador-mint 1 VM 36 0.862 1.05 17603.3 -token-v2-ambassador-mint 100 VM 36 0.835 1.042 16110.4 -liquidity-pool-swap 1 VM 36 0.86 1.019 961.3 -liquidity-pool-swap 100 VM 36 0.781 1.029 11256.7 -liquidity-pool-swap-stable 1 VM 36 0.908 1.026 936.3 -liquidity-pool-swap-stable 100 VM 36 0.838 1.031 10977.4 -deserialize-u256 1 VM 36 0.771 1.065 39507 -no-op-fee-payer 1 VM 36 0.933 1.038 2095.4 -no-op-fee-payer 100 VM 36 0.82 1.036 27792.6 +no-op 1 VM 57 0.758 1.079 40390.5 +no-op 1000 VM 57 0.740 1.040 22473.1 +apt-fa-transfer 1 VM 57 0.762 1.070 28769.8 +account-generation 1 VM 57 0.774 1.055 23332.3 +account-resource32-b 1 VM 57 0.799 1.084 35822.6 +modify-global-resource 1 VM 57 0.810 1.022 2789.1 +modify-global-resource 100 VM 57 0.757 1.040 35206.7 +publish-package 1 VM 57 0.899 1.055 144.8 +mix_publish_transfer 1 VM 57 0.921 1.094 2141.3 +batch100-transfer 1 VM 58 0.727 1.022 740.9 +vector-picture30k 1 VM 58 0.858 1.030 111.0 +vector-picture30k 100 VM 58 0.546 1.041 2021.7 +smart-table-picture30-k-with200-change 1 VM 58 0.944 1.056 21.3 +smart-table-picture30-k-with200-change 100 VM 58 0.895 1.039 402.8 +modify-global-resource-agg-v2 1 VM 58 0.773 1.085 37354.2 +modify-global-flag-agg-v2 1 VM 58 0.866 1.022 5508.5 +modify-global-bounded-agg-v2 1 VM 58 0.872 1.091 9731.3 +modify-global-milestone-agg-v2 1 VM 58 0.813 1.037 28612.4 +resource-groups-global-write-tag1-kb 1 VM 58 0.800 1.048 9180.8 +resource-groups-global-write-and-read-tag1-kb 1 VM 58 0.849 1.025 6196.8 +resource-groups-sender-write-tag1-kb 1 VM 58 0.886 1.180 19936.8 +resource-groups-sender-multi-change1-kb 1 VM 58 0.819 1.088 16466.1 +token-v1ft-mint-and-transfer 1 VM 58 0.810 1.039 1262.2 +token-v1ft-mint-and-transfer 100 VM 58 0.772 1.032 18083.7 +token-v1nft-mint-and-transfer-sequential 1 VM 58 0.809 1.023 795.5 +token-v1nft-mint-and-transfer-sequential 100 VM 58 0.759 1.035 12888.4 +coin-init-and-mint 1 VM 58 0.814 1.083 29357.5 +coin-init-and-mint 100 VM 58 0.777 1.064 24307.2 +fungible-asset-mint 1 VM 58 0.799 1.037 26666.4 +fungible-asset-mint 100 VM 58 0.810 1.045 21562.8 +no-op5-signers 1 VM 58 0.811 1.085 40094.5 +token-v2-ambassador-mint 1 VM 58 0.784 1.050 17603.3 +token-v2-ambassador-mint 100 VM 58 0.796 1.042 16110.4 +liquidity-pool-swap 1 VM 58 0.810 1.027 961.3 +liquidity-pool-swap 100 VM 58 0.770 1.029 11256.7 +liquidity-pool-swap-stable 1 VM 58 0.794 1.026 936.3 +liquidity-pool-swap-stable 100 VM 58 0.783 1.031 10977.4 +deserialize-u256 1 VM 58 0.775 1.071 39288.2 +no-op-fee-payer 1 VM 58 0.823 1.038 2095.4 +no-op-fee-payer 100 VM 58 0.799 1.038 27842.9 """ # when adding a new test, add estimated expected_tps to it, as well as waived=True. @@ -617,30 +617,27 @@ def print_table( else: assert test.key in calibrated_expected_tps, test 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 - 0.01), - min_warn_tps=cur_calibration.expected_tps - * pow(cur_calibration.min_ratio, 0.5), - max_tps=cur_calibration.expected_tps - * (cur_calibration.max_ratio + 0.01), - 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), - ) + criteria = Criteria( + expected_tps=cur_calibration.expected_tps, + min_tps=cur_calibration.expected_tps + * ( + 1 + - (1 - cur_calibration.min_ratio) + * (1 + 10.0 / cur_calibration.count) + - 1.0 / cur_calibration.count + ), + min_warn_tps=cur_calibration.expected_tps + * pow(cur_calibration.min_ratio, 0.8), + max_tps=cur_calibration.expected_tps + * ( + 1 + + (cur_calibration.max_ratio - 1) + * (1 + 10.0 / cur_calibration.count) + + 1.0 / cur_calibration.count + ), + max_warn_tps=cur_calibration.expected_tps + * pow(cur_calibration.max_ratio, 0.8), + ) cur_block_size = int(min([criteria.expected_tps, MAX_BLOCK_SIZE])) @@ -788,19 +785,19 @@ def print_table( print_table(results, by_levels=False, single_field=None) if 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" + text = f"regression detected {single_node_result.tps}, expected median {criteria.expected_tps}, threshold: {criteria.min_tps}), {test.key} didn't meet TPS requirements" if not test.waived: errors.append(text) else: warnings.append(text) elif single_node_result.tps < criteria.min_warn_tps: - 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" + text = f"potential (but within normal noise) regression detected {single_node_result.tps}, expected median {criteria.expected_tps}, threshold: {criteria.min_warn_tps}), {test.key} didn't meet TPS requirements" warnings.append(text) elif ( not SKIP_PERF_IMPROVEMENT_NOTICE and single_node_result.tps > criteria.max_tps ): - 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" + text = f"perf improvement detected {single_node_result.tps}, expected median {criteria.expected_tps}, threshold: {criteria.max_tps}), {test.key} exceeded TPS requirements, increase TPS requirements to match new baseline" if not test.waived: errors.append(text) else: @@ -809,7 +806,7 @@ def print_table( not SKIP_PERF_IMPROVEMENT_NOTICE and single_node_result.tps > criteria.max_warn_tps ): - 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" + text = f"potential (but within normal noise) perf improvement detected {single_node_result.tps}, expected median {criteria.expected_tps}, threshold: {criteria.max_warn_tps}), {test.key} exceeded TPS requirements, increase TPS requirements to match new baseline" warnings.append(text) if HIDE_OUTPUT: