Skip to content

[NOMERG] PPO-Myo #1514

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.

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

Draft
wants to merge 18 commits into
base: main
Choose a base branch
from
Draft

[NOMERG] PPO-Myo #1514

wants to merge 18 commits into from

Conversation

vmoens
Copy link
Contributor

@vmoens vmoens commented Sep 8, 2023

No description provided.

@facebook-github-bot facebook-github-bot added the CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. label Sep 8, 2023
@vmoens vmoens added the new algo New algorithm request or PR label Sep 8, 2023
@github-actions
Copy link

github-actions bot commented Sep 8, 2023

$\color{#D29922}\textsf{\Large⚠\kern{0.2cm}\normalsize Warning}$ Result of CPU Benchmark Tests

Total Benchmarks: 89. Improved: $\large\color{#35bf28}8$. Worsened: $\large\color{#d91a1a}3$.

Expand to view detailed results
Name Max Mean Ops Ops on Repo HEAD Change
test_single 72.7640ms 72.3151ms 13.8284 Ops/s 13.3273 Ops/s $\color{#35bf28}+3.76\%$
test_sync 0.1204s 42.7865ms 23.3718 Ops/s 23.2197 Ops/s $\color{#35bf28}+0.66\%$
test_async 0.1529s 38.8231ms 25.7579 Ops/s 25.6939 Ops/s $\color{#35bf28}+0.25\%$
test_simple 0.6884s 0.6193s 1.6147 Ops/s 1.5960 Ops/s $\color{#35bf28}+1.18\%$
test_transformed 0.8450s 0.7871s 1.2705 Ops/s 1.2494 Ops/s $\color{#35bf28}+1.69\%$
test_serial 1.7843s 1.7181s 0.5820 Ops/s 0.5790 Ops/s $\color{#35bf28}+0.53\%$
test_parallel 1.5155s 1.4441s 0.6925 Ops/s 0.6229 Ops/s $\textbf{\color{#35bf28}+11.18\%}$
test_step_mdp_speed[True-True-True-True-True] 0.2206ms 44.4157μs 22.5145 KOps/s 22.1110 KOps/s $\color{#35bf28}+1.82\%$
test_step_mdp_speed[True-True-True-True-False] 79.4020μs 24.9594μs 40.0651 KOps/s 38.6793 KOps/s $\color{#35bf28}+3.58\%$
test_step_mdp_speed[True-True-True-False-True] 0.1177ms 31.3267μs 31.9216 KOps/s 30.9931 KOps/s $\color{#35bf28}+3.00\%$
test_step_mdp_speed[True-True-True-False-False] 47.3010μs 17.5477μs 56.9876 KOps/s 55.8424 KOps/s $\color{#35bf28}+2.05\%$
test_step_mdp_speed[True-True-False-True-True] 84.2020μs 46.0769μs 21.7029 KOps/s 21.3553 KOps/s $\color{#35bf28}+1.63\%$
test_step_mdp_speed[True-True-False-True-False] 0.1042ms 26.9800μs 37.0645 KOps/s 36.1759 KOps/s $\color{#35bf28}+2.46\%$
test_step_mdp_speed[True-True-False-False-True] 0.2967ms 33.4385μs 29.9056 KOps/s 29.3158 KOps/s $\color{#35bf28}+2.01\%$
test_step_mdp_speed[True-True-False-False-False] 85.3020μs 19.6101μs 50.9941 KOps/s 50.3779 KOps/s $\color{#35bf28}+1.22\%$
test_step_mdp_speed[True-False-True-True-True] 78.2020μs 47.9491μs 20.8555 KOps/s 20.2794 KOps/s $\color{#35bf28}+2.84\%$
test_step_mdp_speed[True-False-True-True-False] 0.1155ms 28.8032μs 34.7183 KOps/s 33.7963 KOps/s $\color{#35bf28}+2.73\%$
test_step_mdp_speed[True-False-True-False-True] 99.7020μs 33.3367μs 29.9969 KOps/s 29.4982 KOps/s $\color{#35bf28}+1.69\%$
test_step_mdp_speed[True-False-True-False-False] 0.1076ms 19.5343μs 51.1921 KOps/s 49.2505 KOps/s $\color{#35bf28}+3.94\%$
test_step_mdp_speed[True-False-False-True-True] 0.1720ms 49.2671μs 20.2975 KOps/s 19.6799 KOps/s $\color{#35bf28}+3.14\%$
test_step_mdp_speed[True-False-False-True-False] 46.2010μs 30.1361μs 33.1828 KOps/s 31.7860 KOps/s $\color{#35bf28}+4.39\%$
test_step_mdp_speed[True-False-False-False-True] 0.1092ms 35.0559μs 28.5259 KOps/s 28.0495 KOps/s $\color{#35bf28}+1.70\%$
test_step_mdp_speed[True-False-False-False-False] 48.9010μs 21.0179μs 47.5785 KOps/s 46.2244 KOps/s $\color{#35bf28}+2.93\%$
test_step_mdp_speed[False-True-True-True-True] 0.1035ms 48.0221μs 20.8237 KOps/s 20.5497 KOps/s $\color{#35bf28}+1.33\%$
test_step_mdp_speed[False-True-True-True-False] 0.1036ms 28.6096μs 34.9533 KOps/s 33.6958 KOps/s $\color{#35bf28}+3.73\%$
test_step_mdp_speed[False-True-True-False-True] 75.3010μs 37.9141μs 26.3754 KOps/s 26.2477 KOps/s $\color{#35bf28}+0.49\%$
test_step_mdp_speed[False-True-True-False-False] 3.6521ms 21.5229μs 46.4620 KOps/s 44.7483 KOps/s $\color{#35bf28}+3.83\%$
test_step_mdp_speed[False-True-False-True-True] 0.1596ms 49.8177μs 20.0732 KOps/s 19.8802 KOps/s $\color{#35bf28}+0.97\%$
test_step_mdp_speed[False-True-False-True-False] 0.1131ms 29.9880μs 33.3467 KOps/s 32.0666 KOps/s $\color{#35bf28}+3.99\%$
test_step_mdp_speed[False-True-False-False-True] 91.6020μs 39.1163μs 25.5648 KOps/s 25.3716 KOps/s $\color{#35bf28}+0.76\%$
test_step_mdp_speed[False-True-False-False-False] 0.1869ms 22.9934μs 43.4907 KOps/s 41.8121 KOps/s $\color{#35bf28}+4.01\%$
test_step_mdp_speed[False-False-True-True-True] 0.1576ms 51.6318μs 19.3679 KOps/s 19.0153 KOps/s $\color{#35bf28}+1.85\%$
test_step_mdp_speed[False-False-True-True-False] 80.4020μs 32.2601μs 30.9980 KOps/s 30.0644 KOps/s $\color{#35bf28}+3.11\%$
test_step_mdp_speed[False-False-True-False-True] 0.1375ms 38.7720μs 25.7918 KOps/s 25.2603 KOps/s $\color{#35bf28}+2.10\%$
test_step_mdp_speed[False-False-True-False-False] 51.8010μs 23.0100μs 43.4593 KOps/s 42.1491 KOps/s $\color{#35bf28}+3.11\%$
test_step_mdp_speed[False-False-False-True-True] 87.3020μs 51.9167μs 19.2616 KOps/s 18.5749 KOps/s $\color{#35bf28}+3.70\%$
test_step_mdp_speed[False-False-False-True-False] 0.1275ms 33.6199μs 29.7443 KOps/s 28.8481 KOps/s $\color{#35bf28}+3.11\%$
test_step_mdp_speed[False-False-False-False-True] 66.3010μs 39.8967μs 25.0647 KOps/s 24.4657 KOps/s $\color{#35bf28}+2.45\%$
test_step_mdp_speed[False-False-False-False-False] 0.1097ms 24.9241μs 40.1217 KOps/s 39.2583 KOps/s $\color{#35bf28}+2.20\%$
test_values[generalized_advantage_estimate-True-True] 16.7070ms 13.5575ms 73.7599 Ops/s 73.2491 Ops/s $\color{#35bf28}+0.70\%$
test_values[vec_generalized_advantage_estimate-True-True] 48.9111ms 42.7875ms 23.3713 Ops/s 23.5115 Ops/s $\color{#d91a1a}-0.60\%$
test_values[td0_return_estimate-False-False] 0.3433ms 0.2019ms 4.9522 KOps/s 4.0887 KOps/s $\textbf{\color{#35bf28}+21.12\%}$
test_values[td1_return_estimate-False-False] 15.3796ms 13.1280ms 76.1731 Ops/s 77.2620 Ops/s $\color{#d91a1a}-1.41\%$
test_values[vec_td1_return_estimate-False-False] 51.9904ms 42.3253ms 23.6265 Ops/s 23.6342 Ops/s $\color{#d91a1a}-0.03\%$
test_values[td_lambda_return_estimate-True-False] 32.3351ms 31.7939ms 31.4525 Ops/s 31.4174 Ops/s $\color{#35bf28}+0.11\%$
test_values[vec_td_lambda_return_estimate-True-False] 49.0336ms 42.5638ms 23.4941 Ops/s 23.2504 Ops/s $\color{#35bf28}+1.05\%$
test_gae_speed[generalized_advantage_estimate-False-1-512] 11.8066ms 11.6969ms 85.4925 Ops/s 84.8605 Ops/s $\color{#35bf28}+0.74\%$
test_gae_speed[vec_generalized_advantage_estimate-True-1-512] 4.7869ms 3.3566ms 297.9200 Ops/s 305.4508 Ops/s $\color{#d91a1a}-2.47\%$
test_gae_speed[vec_generalized_advantage_estimate-False-1-512] 7.3430ms 0.4780ms 2.0919 KOps/s 2.1540 KOps/s $\color{#d91a1a}-2.88\%$
test_gae_speed[vec_generalized_advantage_estimate-True-32-512] 64.6158ms 59.1275ms 16.9126 Ops/s 17.5589 Ops/s $\color{#d91a1a}-3.68\%$
test_gae_speed[vec_generalized_advantage_estimate-False-32-512] 3.5839ms 2.8718ms 348.2182 Ops/s 360.0475 Ops/s $\color{#d91a1a}-3.29\%$
test_dqn_speed 6.9093ms 1.8437ms 542.3838 Ops/s 537.1148 Ops/s $\color{#35bf28}+0.98\%$
test_ddpg_speed 4.5150ms 2.7549ms 362.9865 Ops/s 358.4791 Ops/s $\color{#35bf28}+1.26\%$
test_sac_speed 8.5867ms 8.0525ms 124.1857 Ops/s 120.9773 Ops/s $\color{#35bf28}+2.65\%$
test_redq_speed 23.7935ms 16.1654ms 61.8606 Ops/s 62.0208 Ops/s $\color{#d91a1a}-0.26\%$
test_redq_deprec_speed 19.3356ms 12.7558ms 78.3956 Ops/s 77.1815 Ops/s $\color{#35bf28}+1.57\%$
test_td3_speed 11.6498ms 10.3043ms 97.0466 Ops/s 96.9623 Ops/s $\color{#35bf28}+0.09\%$
test_cql_speed 34.9622ms 27.3671ms 36.5402 Ops/s 28.4003 Ops/s $\textbf{\color{#35bf28}+28.66\%}$
test_a2c_speed 14.7332ms 5.4842ms 182.3417 Ops/s 181.9087 Ops/s $\color{#35bf28}+0.24\%$
test_ppo_speed 14.0231ms 5.8067ms 172.2157 Ops/s 171.5450 Ops/s $\color{#35bf28}+0.39\%$
test_reinforce_speed 13.6578ms 4.1008ms 243.8566 Ops/s 226.7879 Ops/s $\textbf{\color{#35bf28}+7.53\%}$
test_iql_speed 29.2737ms 21.3970ms 46.7356 Ops/s 43.4684 Ops/s $\textbf{\color{#35bf28}+7.52\%}$
test_sample_rb[TensorDictReplayBuffer-ListStorage-RandomSampler-4000] 3.6979ms 2.8063ms 356.3422 Ops/s 365.9753 Ops/s $\color{#d91a1a}-2.63\%$
test_sample_rb[TensorDictReplayBuffer-LazyMemmapStorage-RandomSampler-10000] 5.6071ms 2.9022ms 344.5637 Ops/s 351.0458 Ops/s $\color{#d91a1a}-1.85\%$
test_sample_rb[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-10000] 5.5247ms 2.9314ms 341.1284 Ops/s 347.2892 Ops/s $\color{#d91a1a}-1.77\%$
test_sample_rb[TensorDictReplayBuffer-ListStorage-SamplerWithoutReplacement-4000] 3.7781ms 2.7858ms 358.9697 Ops/s 363.1905 Ops/s $\color{#d91a1a}-1.16\%$
test_sample_rb[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-10000] 5.7768ms 2.9030ms 344.4752 Ops/s 353.6548 Ops/s $\color{#d91a1a}-2.60\%$
test_sample_rb[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-10000] 5.7699ms 2.8832ms 346.8318 Ops/s 338.2923 Ops/s $\color{#35bf28}+2.52\%$
test_sample_rb[TensorDictPrioritizedReplayBuffer-ListStorage-None-4000] 3.5083ms 2.7162ms 368.1640 Ops/s 363.4539 Ops/s $\color{#35bf28}+1.30\%$
test_sample_rb[TensorDictPrioritizedReplayBuffer-LazyMemmapStorage-None-10000] 5.6279ms 2.8930ms 345.6570 Ops/s 307.1883 Ops/s $\textbf{\color{#35bf28}+12.52\%}$
test_sample_rb[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-10000] 4.9263ms 2.9439ms 339.6876 Ops/s 355.4380 Ops/s $\color{#d91a1a}-4.43\%$
test_iterate_rb[TensorDictReplayBuffer-ListStorage-RandomSampler-4000] 3.6260ms 2.7359ms 365.5097 Ops/s 303.5252 Ops/s $\textbf{\color{#35bf28}+20.42\%}$
test_iterate_rb[TensorDictReplayBuffer-LazyMemmapStorage-RandomSampler-10000] 5.9309ms 2.8773ms 347.5520 Ops/s 351.9930 Ops/s $\color{#d91a1a}-1.26\%$
test_iterate_rb[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-10000] 6.5159ms 2.9395ms 340.1957 Ops/s 354.8387 Ops/s $\color{#d91a1a}-4.13\%$
test_iterate_rb[TensorDictReplayBuffer-ListStorage-SamplerWithoutReplacement-4000] 0.2703s 3.5107ms 284.8417 Ops/s 374.5835 Ops/s $\textbf{\color{#d91a1a}-23.96\%}$
test_iterate_rb[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-10000] 5.0426ms 2.9402ms 340.1082 Ops/s 353.6203 Ops/s $\color{#d91a1a}-3.82\%$
test_iterate_rb[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-10000] 5.3715ms 2.9209ms 342.3554 Ops/s 345.9683 Ops/s $\color{#d91a1a}-1.04\%$
test_iterate_rb[TensorDictPrioritizedReplayBuffer-ListStorage-None-4000] 3.5769ms 2.7219ms 367.3964 Ops/s 378.8805 Ops/s $\color{#d91a1a}-3.03\%$
test_iterate_rb[TensorDictPrioritizedReplayBuffer-LazyMemmapStorage-None-10000] 4.8301ms 2.8726ms 348.1194 Ops/s 356.4187 Ops/s $\color{#d91a1a}-2.33\%$
test_iterate_rb[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-10000] 5.1591ms 2.9206ms 342.3981 Ops/s 353.6106 Ops/s $\color{#d91a1a}-3.17\%$
test_populate_rb[TensorDictReplayBuffer-ListStorage-RandomSampler-400] 0.2843s 30.5081ms 32.7782 Ops/s 32.1042 Ops/s $\color{#35bf28}+2.10\%$
test_populate_rb[TensorDictReplayBuffer-LazyMemmapStorage-RandomSampler-400] 0.1524s 30.2357ms 33.0735 Ops/s 40.9964 Ops/s $\textbf{\color{#d91a1a}-19.33\%}$
test_populate_rb[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-400] 0.1523s 27.6991ms 36.1022 Ops/s 34.1429 Ops/s $\textbf{\color{#35bf28}+5.74\%}$
test_populate_rb[TensorDictReplayBuffer-ListStorage-SamplerWithoutReplacement-400] 0.1518s 30.1984ms 33.1143 Ops/s 34.1508 Ops/s $\color{#d91a1a}-3.04\%$
test_populate_rb[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-400] 0.1508s 27.4780ms 36.3928 Ops/s 37.3257 Ops/s $\color{#d91a1a}-2.50\%$
test_populate_rb[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-400] 0.1550s 30.4410ms 32.8504 Ops/s 34.1468 Ops/s $\color{#d91a1a}-3.80\%$
test_populate_rb[TensorDictPrioritizedReplayBuffer-ListStorage-None-400] 0.1554s 27.6488ms 36.1680 Ops/s 37.0671 Ops/s $\color{#d91a1a}-2.43\%$
test_populate_rb[TensorDictPrioritizedReplayBuffer-LazyMemmapStorage-None-400] 0.1732s 31.0172ms 32.2402 Ops/s 33.8570 Ops/s $\color{#d91a1a}-4.78\%$
test_populate_rb[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-400] 0.1669s 31.3266ms 31.9218 Ops/s 36.6287 Ops/s $\textbf{\color{#d91a1a}-12.85\%}$

Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment
Labels
CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. new algo New algorithm request or PR
Projects
None yet
Development

Successfully merging this pull request may close these issues.

3 participants