Skip to content

[Feature] Incremental TensorStorageCheckpointer#3344

Open
vmoens wants to merge 1 commit intomainfrom
incremental-ckpt
Open

[Feature] Incremental TensorStorageCheckpointer#3344
vmoens wants to merge 1 commit intomainfrom
incremental-ckpt

Conversation

@vmoens
Copy link
Collaborator

@vmoens vmoens commented Jan 19, 2026

No description provided.

@pytorch-bot
Copy link

pytorch-bot bot commented Jan 19, 2026

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/rl/3344

Note: Links to docs will display an error until the docs builds have been completed.

✅ You can merge normally! (2 Unrelated Failures)

As of commit b561f2f with merge base ae50733 (image):

BROKEN TRUNK - The following jobs failed but were present on the merge base:

👉 Rebase onto the `viable/strict` branch to avoid these failures

This comment was automatically generated by Dr. CI and updates every 15 minutes.

@meta-cla meta-cla 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 Jan 19, 2026
@github-actions
Copy link
Contributor

github-actions bot commented Jan 19, 2026

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

Total Benchmarks: 153. Improved: $\large\color{#35bf28}16$. Worsened: $\large\color{#d91a1a}13$.

Expand to view detailed results
Name Max Mean Ops Ops on Repo HEAD Change
test_tensor_to_bytestream_speed[pickle] 80.7489μs 80.1390μs 12.4783 KOps/s 12.2010 KOps/s $\color{#35bf28}+2.27\%$
test_tensor_to_bytestream_speed[torch.save] 0.1450ms 0.1447ms 6.9090 KOps/s 7.0853 KOps/s $\color{#d91a1a}-2.49\%$
test_tensor_to_bytestream_speed[untyped_storage] 0.1144s 0.1138s 8.7843 Ops/s 8.0752 Ops/s $\textbf{\color{#35bf28}+8.78\%}$
test_tensor_to_bytestream_speed[numpy] 2.5501μs 2.5442μs 393.0522 KOps/s 386.5924 KOps/s $\color{#35bf28}+1.67\%$
test_tensor_to_bytestream_speed[safetensors] 37.1486μs 36.9229μs 27.0835 KOps/s 24.9601 KOps/s $\textbf{\color{#35bf28}+8.51\%}$
test_simple 0.6665s 0.5758s 1.7367 Ops/s 1.7432 Ops/s $\color{#d91a1a}-0.37\%$
test_transformed 1.2490s 1.1574s 0.8640 Ops/s 0.8667 Ops/s $\color{#d91a1a}-0.31\%$
test_serial 1.6903s 1.6865s 0.5930 Ops/s 0.5829 Ops/s $\color{#35bf28}+1.73\%$
test_parallel 1.3188s 1.1833s 0.8451 Ops/s 0.7983 Ops/s $\textbf{\color{#35bf28}+5.86\%}$
test_step_mdp_speed[True-True-True-True-True] 0.1534ms 44.0485μs 22.7023 KOps/s 22.2900 KOps/s $\color{#35bf28}+1.85\%$
test_step_mdp_speed[True-True-True-True-False] 53.2900μs 25.4110μs 39.3531 KOps/s 40.0731 KOps/s $\color{#d91a1a}-1.80\%$
test_step_mdp_speed[True-True-True-False-True] 59.6710μs 25.4702μs 39.2616 KOps/s 40.0270 KOps/s $\color{#d91a1a}-1.91\%$
test_step_mdp_speed[True-True-True-False-False] 40.8000μs 14.0873μs 70.9861 KOps/s 72.6980 KOps/s $\color{#d91a1a}-2.35\%$
test_step_mdp_speed[True-True-False-True-True] 85.5300μs 48.0527μs 20.8105 KOps/s 20.9866 KOps/s $\color{#d91a1a}-0.84\%$
test_step_mdp_speed[True-True-False-True-False] 55.1400μs 27.5725μs 36.2681 KOps/s 36.1482 KOps/s $\color{#35bf28}+0.33\%$
test_step_mdp_speed[True-True-False-False-True] 62.3410μs 27.8202μs 35.9451 KOps/s 35.7216 KOps/s $\color{#35bf28}+0.63\%$
test_step_mdp_speed[True-True-False-False-False] 43.4400μs 16.8489μs 59.3512 KOps/s 60.0392 KOps/s $\color{#d91a1a}-1.15\%$
test_step_mdp_speed[True-False-True-True-True] 78.1810μs 51.9625μs 19.2447 KOps/s 19.4498 KOps/s $\color{#d91a1a}-1.05\%$
test_step_mdp_speed[True-False-True-True-False] 67.1100μs 31.2590μs 31.9908 KOps/s 32.5308 KOps/s $\color{#d91a1a}-1.66\%$
test_step_mdp_speed[True-False-True-False-True] 55.1210μs 28.3347μs 35.2924 KOps/s 35.8359 KOps/s $\color{#d91a1a}-1.52\%$
test_step_mdp_speed[True-False-True-False-False] 43.9210μs 17.1410μs 58.3397 KOps/s 60.4935 KOps/s $\color{#d91a1a}-3.56\%$
test_step_mdp_speed[True-False-False-True-True] 0.1215ms 53.5302μs 18.6810 KOps/s 19.0487 KOps/s $\color{#d91a1a}-1.93\%$
test_step_mdp_speed[True-False-False-True-False] 68.4610μs 32.6067μs 30.6686 KOps/s 30.7292 KOps/s $\color{#d91a1a}-0.20\%$
test_step_mdp_speed[True-False-False-False-True] 65.2800μs 30.6606μs 32.6151 KOps/s 33.3304 KOps/s $\color{#d91a1a}-2.15\%$
test_step_mdp_speed[True-False-False-False-False] 46.0210μs 18.9497μs 52.7714 KOps/s 52.9079 KOps/s $\color{#d91a1a}-0.26\%$
test_step_mdp_speed[False-True-True-True-True] 86.2210μs 51.0775μs 19.5781 KOps/s 19.5089 KOps/s $\color{#35bf28}+0.35\%$
test_step_mdp_speed[False-True-True-True-False] 62.5410μs 31.2788μs 31.9705 KOps/s 32.8931 KOps/s $\color{#d91a1a}-2.80\%$
test_step_mdp_speed[False-True-True-False-True] 58.8710μs 32.2407μs 31.0167 KOps/s 32.0945 KOps/s $\color{#d91a1a}-3.36\%$
test_step_mdp_speed[False-True-True-False-False] 60.9100μs 18.5418μs 53.9321 KOps/s 55.7102 KOps/s $\color{#d91a1a}-3.19\%$
test_step_mdp_speed[False-True-False-True-True] 2.6942ms 54.3459μs 18.4007 KOps/s 18.8904 KOps/s $\color{#d91a1a}-2.59\%$
test_step_mdp_speed[False-True-False-True-False] 67.5200μs 33.6152μs 29.7484 KOps/s 30.4026 KOps/s $\color{#d91a1a}-2.15\%$
test_step_mdp_speed[False-True-False-False-True] 89.7310μs 34.5548μs 28.9395 KOps/s 29.5811 KOps/s $\color{#d91a1a}-2.17\%$
test_step_mdp_speed[False-True-False-False-False] 49.8110μs 21.2922μs 46.9655 KOps/s 47.6348 KOps/s $\color{#d91a1a}-1.41\%$
test_step_mdp_speed[False-False-True-True-True] 83.8010μs 56.4019μs 17.7299 KOps/s 17.8583 KOps/s $\color{#d91a1a}-0.72\%$
test_step_mdp_speed[False-False-True-True-False] 67.2810μs 36.3832μs 27.4852 KOps/s 27.4600 KOps/s $\color{#35bf28}+0.09\%$
test_step_mdp_speed[False-False-True-False-True] 68.4310μs 34.5414μs 28.9508 KOps/s 29.2850 KOps/s $\color{#d91a1a}-1.14\%$
test_step_mdp_speed[False-False-True-False-False] 53.2410μs 21.4252μs 46.6740 KOps/s 48.1422 KOps/s $\color{#d91a1a}-3.05\%$
test_step_mdp_speed[False-False-False-True-True] 95.1810μs 58.5689μs 17.0739 KOps/s 17.4496 KOps/s $\color{#d91a1a}-2.15\%$
test_step_mdp_speed[False-False-False-True-False] 63.3910μs 38.6351μs 25.8832 KOps/s 25.8633 KOps/s $\color{#35bf28}+0.08\%$
test_step_mdp_speed[False-False-False-False-True] 66.8010μs 36.6017μs 27.3211 KOps/s 27.7593 KOps/s $\color{#d91a1a}-1.58\%$
test_step_mdp_speed[False-False-False-False-False] 53.7600μs 23.3743μs 42.7820 KOps/s 42.9831 KOps/s $\color{#d91a1a}-0.47\%$
test_non_tensor_env_rollout_speed[1000-single-True] 0.8655s 0.7716s 1.2960 Ops/s 1.2984 Ops/s $\color{#d91a1a}-0.18\%$
test_non_tensor_env_rollout_speed[1000-single-False] 0.7324s 0.6357s 1.5731 Ops/s 1.5808 Ops/s $\color{#d91a1a}-0.49\%$
test_non_tensor_env_rollout_speed[1000-serial-no-buffers-True] 1.7509s 1.6780s 0.5959 Ops/s 0.5956 Ops/s $\color{#35bf28}+0.07\%$
test_non_tensor_env_rollout_speed[1000-serial-no-buffers-False] 1.5338s 1.4541s 0.6877 Ops/s 0.6899 Ops/s $\color{#d91a1a}-0.32\%$
test_non_tensor_env_rollout_speed[1000-serial-buffers-True] 2.0011s 1.9263s 0.5191 Ops/s 0.5181 Ops/s $\color{#35bf28}+0.21\%$
test_non_tensor_env_rollout_speed[1000-serial-buffers-False] 1.7832s 1.7038s 0.5869 Ops/s 0.5854 Ops/s $\color{#35bf28}+0.26\%$
test_non_tensor_env_rollout_speed[1000-parallel-no-buffers-True] 4.7709s 4.6217s 0.2164 Ops/s 0.2118 Ops/s $\color{#35bf28}+2.16\%$
test_non_tensor_env_rollout_speed[1000-parallel-no-buffers-False] 4.4835s 4.4311s 0.2257 Ops/s 0.2222 Ops/s $\color{#35bf28}+1.57\%$
test_non_tensor_env_rollout_speed[1000-parallel-buffers-True] 2.0393s 1.9534s 0.5119 Ops/s 0.5057 Ops/s $\color{#35bf28}+1.24\%$
test_non_tensor_env_rollout_speed[1000-parallel-buffers-False] 1.7607s 1.6858s 0.5932 Ops/s 0.6006 Ops/s $\color{#d91a1a}-1.24\%$
test_values[generalized_advantage_estimate-True-True] 10.7190ms 10.4693ms 95.5172 Ops/s 91.8566 Ops/s $\color{#35bf28}+3.99\%$
test_values[vec_generalized_advantage_estimate-True-True] 20.0405ms 18.0620ms 55.3647 Ops/s 87.4572 Ops/s $\textbf{\color{#d91a1a}-36.70\%}$
test_values[td0_return_estimate-False-False] 0.2277ms 0.1301ms 7.6880 KOps/s 7.9489 KOps/s $\color{#d91a1a}-3.28\%$
test_values[td1_return_estimate-False-False] 29.6915ms 28.4860ms 35.1049 Ops/s 33.9749 Ops/s $\color{#35bf28}+3.33\%$
test_values[vec_td1_return_estimate-False-False] 18.9109ms 18.1232ms 55.1779 Ops/s 87.6007 Ops/s $\textbf{\color{#d91a1a}-37.01\%}$
test_values[td_lambda_return_estimate-True-False] 44.1577ms 42.1150ms 23.7445 Ops/s 22.7893 Ops/s $\color{#35bf28}+4.19\%$
test_values[vec_td_lambda_return_estimate-True-False] 18.6022ms 18.1377ms 55.1336 Ops/s 87.1915 Ops/s $\textbf{\color{#d91a1a}-36.77\%}$
test_gae_speed[generalized_advantage_estimate-False-1-512] 9.2901ms 9.2018ms 108.6747 Ops/s 102.7717 Ops/s $\textbf{\color{#35bf28}+5.74\%}$
test_gae_speed[vec_generalized_advantage_estimate-True-1-512] 2.0239ms 1.5223ms 656.8930 Ops/s 682.5557 Ops/s $\color{#d91a1a}-3.76\%$
test_gae_speed[vec_generalized_advantage_estimate-False-1-512] 0.4817ms 0.4393ms 2.2763 KOps/s 2.2453 KOps/s $\color{#35bf28}+1.38\%$
test_gae_speed[vec_generalized_advantage_estimate-True-32-512] 34.9026ms 34.2681ms 29.1817 Ops/s 43.5223 Ops/s $\textbf{\color{#d91a1a}-32.95\%}$
test_gae_speed[vec_generalized_advantage_estimate-False-32-512] 1.8818ms 1.7478ms 572.1410 Ops/s 568.3909 Ops/s $\color{#35bf28}+0.66\%$
test_dqn_speed[False-None] 1.7161ms 1.4294ms 699.5717 Ops/s 701.5619 Ops/s $\color{#d91a1a}-0.28\%$
test_dqn_speed[False-backward] 2.0334ms 1.9682ms 508.0889 Ops/s 511.5445 Ops/s $\color{#d91a1a}-0.68\%$
test_dqn_speed[True-None] 0.9574ms 0.5358ms 1.8662 KOps/s 1.7842 KOps/s $\color{#35bf28}+4.60\%$
test_dqn_speed[True-backward] 1.0189ms 0.9924ms 1.0076 KOps/s 835.6571 Ops/s $\textbf{\color{#35bf28}+20.58\%}$
test_dqn_speed[reduce-overhead-None] 0.9067ms 0.5232ms 1.9113 KOps/s 1.8220 KOps/s $\color{#35bf28}+4.90\%$
test_ddpg_speed[False-None] 3.2752ms 2.8753ms 347.7956 Ops/s 339.2930 Ops/s $\color{#35bf28}+2.51\%$
test_ddpg_speed[False-backward] 4.2488ms 4.1268ms 242.3210 Ops/s 238.4365 Ops/s $\color{#35bf28}+1.63\%$
test_ddpg_speed[True-None] 1.7843ms 1.3970ms 715.8348 Ops/s 695.9918 Ops/s $\color{#35bf28}+2.85\%$
test_ddpg_speed[True-backward] 2.5758ms 2.4081ms 415.2722 Ops/s 371.2241 Ops/s $\textbf{\color{#35bf28}+11.87\%}$
test_ddpg_speed[reduce-overhead-None] 1.7697ms 1.3838ms 722.6510 Ops/s 713.2024 Ops/s $\color{#35bf28}+1.32\%$
test_sac_speed[False-None] 8.8088ms 8.1859ms 122.1612 Ops/s 123.0949 Ops/s $\color{#d91a1a}-0.76\%$
test_sac_speed[False-backward] 12.1706ms 11.5005ms 86.9524 Ops/s 87.2403 Ops/s $\color{#d91a1a}-0.33\%$
test_sac_speed[True-None] 2.6133ms 2.1618ms 462.5694 Ops/s 461.0812 Ops/s $\color{#35bf28}+0.32\%$
test_sac_speed[True-backward] 4.1438ms 4.0497ms 246.9309 Ops/s 242.9775 Ops/s $\color{#35bf28}+1.63\%$
test_sac_speed[reduce-overhead-None] 2.5235ms 2.1182ms 472.1062 Ops/s 457.7270 Ops/s $\color{#35bf28}+3.14\%$
test_redq_speed[False-None] 15.1084ms 10.6811ms 93.6233 Ops/s 95.4558 Ops/s $\color{#d91a1a}-1.92\%$
test_redq_speed[False-backward] 18.5365ms 18.0277ms 55.4703 Ops/s 55.4143 Ops/s $\color{#35bf28}+0.10\%$
test_redq_speed[True-None] 4.6906ms 4.4256ms 225.9569 Ops/s 221.3767 Ops/s $\color{#35bf28}+2.07\%$
test_redq_speed[True-backward] 9.9225ms 9.7151ms 102.9327 Ops/s 99.7258 Ops/s $\color{#35bf28}+3.22\%$
test_redq_speed[reduce-overhead-None] 4.5021ms 4.3155ms 231.7202 Ops/s 223.7186 Ops/s $\color{#35bf28}+3.58\%$
test_redq_deprec_speed[False-None] 11.7218ms 11.0602ms 90.4139 Ops/s 89.6617 Ops/s $\color{#35bf28}+0.84\%$
test_redq_deprec_speed[False-backward] 16.2089ms 15.7858ms 63.3481 Ops/s 62.7737 Ops/s $\color{#35bf28}+0.92\%$
test_redq_deprec_speed[True-None] 3.8273ms 3.6325ms 275.2939 Ops/s 274.3240 Ops/s $\color{#35bf28}+0.35\%$
test_redq_deprec_speed[True-backward] 7.8013ms 7.5923ms 131.7133 Ops/s 131.3788 Ops/s $\color{#35bf28}+0.25\%$
test_redq_deprec_speed[reduce-overhead-None] 4.6684ms 3.6505ms 273.9326 Ops/s 282.3118 Ops/s $\color{#d91a1a}-2.97\%$
test_td3_speed[False-None] 8.2189ms 8.0326ms 124.4927 Ops/s 122.7051 Ops/s $\color{#35bf28}+1.46\%$
test_td3_speed[False-backward] 11.3516ms 10.9691ms 91.1653 Ops/s 90.2362 Ops/s $\color{#35bf28}+1.03\%$
test_td3_speed[True-None] 1.8856ms 1.8317ms 545.9393 Ops/s 534.9074 Ops/s $\color{#35bf28}+2.06\%$
test_td3_speed[True-backward] 3.7880ms 3.6521ms 273.8134 Ops/s 248.7474 Ops/s $\textbf{\color{#35bf28}+10.08\%}$
test_td3_speed[reduce-overhead-None] 1.8087ms 1.7861ms 559.8718 Ops/s 555.9813 Ops/s $\color{#35bf28}+0.70\%$
test_cql_speed[False-None] 27.1763ms 25.9078ms 38.5985 Ops/s 38.4514 Ops/s $\color{#35bf28}+0.38\%$
test_cql_speed[False-backward] 38.9741ms 35.6496ms 28.0508 Ops/s 28.0468 Ops/s $\color{#35bf28}+0.01\%$
test_cql_speed[True-None] 12.5284ms 12.1841ms 82.0744 Ops/s 80.1285 Ops/s $\color{#35bf28}+2.43\%$
test_cql_speed[True-backward] 18.6410ms 18.2345ms 54.8410 Ops/s 53.8886 Ops/s $\color{#35bf28}+1.77\%$
test_cql_speed[reduce-overhead-None] 12.5733ms 12.3233ms 81.1472 Ops/s 80.6332 Ops/s $\color{#35bf28}+0.64\%$
test_a2c_speed[False-None] 5.5526ms 5.3646ms 186.4068 Ops/s 184.9240 Ops/s $\color{#35bf28}+0.80\%$
test_a2c_speed[False-backward] 11.9042ms 11.6300ms 85.9845 Ops/s 84.3127 Ops/s $\color{#35bf28}+1.98\%$
test_a2c_speed[True-None] 3.8333ms 3.7087ms 269.6368 Ops/s 262.9101 Ops/s $\color{#35bf28}+2.56\%$
test_a2c_speed[True-backward] 8.9382ms 8.6111ms 116.1286 Ops/s 114.9066 Ops/s $\color{#35bf28}+1.06\%$
test_a2c_speed[reduce-overhead-None] 5.8505ms 3.7491ms 266.7298 Ops/s 269.2290 Ops/s $\color{#d91a1a}-0.93\%$
test_ppo_speed[False-None] 6.1554ms 5.9134ms 169.1060 Ops/s 169.9401 Ops/s $\color{#d91a1a}-0.49\%$
test_ppo_speed[False-backward] 13.1283ms 12.5294ms 79.8124 Ops/s 80.6514 Ops/s $\color{#d91a1a}-1.04\%$
test_ppo_speed[True-None] 3.8247ms 3.6237ms 275.9626 Ops/s 272.6685 Ops/s $\color{#35bf28}+1.21\%$
test_ppo_speed[True-backward] 8.6865ms 8.4016ms 119.0256 Ops/s 119.5303 Ops/s $\color{#d91a1a}-0.42\%$
test_ppo_speed[reduce-overhead-None] 3.7586ms 3.5717ms 279.9783 Ops/s 276.6454 Ops/s $\color{#35bf28}+1.20\%$
test_reinforce_speed[False-None] 4.7089ms 4.5186ms 221.3066 Ops/s 221.4843 Ops/s $\color{#d91a1a}-0.08\%$
test_reinforce_speed[False-backward] 7.6344ms 7.3242ms 136.5334 Ops/s 136.0753 Ops/s $\color{#35bf28}+0.34\%$
test_reinforce_speed[True-None] 2.9567ms 2.8189ms 354.7442 Ops/s 342.6730 Ops/s $\color{#35bf28}+3.52\%$
test_reinforce_speed[True-backward] 7.9838ms 7.7843ms 128.4633 Ops/s 130.9849 Ops/s $\color{#d91a1a}-1.93\%$
test_reinforce_speed[reduce-overhead-None] 2.9592ms 2.8279ms 353.6244 Ops/s 345.1464 Ops/s $\color{#35bf28}+2.46\%$
test_iql_speed[False-None] 20.2655ms 19.7182ms 50.7146 Ops/s 50.2026 Ops/s $\color{#35bf28}+1.02\%$
test_iql_speed[False-backward] 31.2035ms 30.5051ms 32.7814 Ops/s 33.0251 Ops/s $\color{#d91a1a}-0.74\%$
test_iql_speed[True-None] 8.7798ms 8.4768ms 117.9685 Ops/s 117.8907 Ops/s $\color{#35bf28}+0.07\%$
test_iql_speed[True-backward] 16.9381ms 16.6699ms 59.9883 Ops/s 60.3774 Ops/s $\color{#d91a1a}-0.64\%$
test_iql_speed[reduce-overhead-None] 8.7019ms 8.4884ms 117.8084 Ops/s 114.7268 Ops/s $\color{#35bf28}+2.69\%$
test_rb_sample[TensorDictReplayBuffer-ListStorage-RandomSampler-4000] 7.6919ms 6.1461ms 162.7052 Ops/s 164.1202 Ops/s $\color{#d91a1a}-0.86\%$
test_rb_sample[TensorDictReplayBuffer-LazyMemmapStorage-RandomSampler-10000] 0.6045ms 0.3659ms 2.7330 KOps/s 2.9509 KOps/s $\textbf{\color{#d91a1a}-7.38\%}$
test_rb_sample[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-10000] 0.5907ms 0.3645ms 2.7432 KOps/s 3.7258 KOps/s $\textbf{\color{#d91a1a}-26.37\%}$
test_rb_sample[TensorDictReplayBuffer-ListStorage-SamplerWithoutReplacement-4000] 6.0195ms 5.7933ms 172.6121 Ops/s 171.0139 Ops/s $\color{#35bf28}+0.93\%$
test_rb_sample[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-10000] 0.7674ms 0.2781ms 3.5956 KOps/s 3.3512 KOps/s $\textbf{\color{#35bf28}+7.29\%}$
test_rb_sample[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-10000] 0.4718ms 0.2616ms 3.8228 KOps/s 3.4426 KOps/s $\textbf{\color{#35bf28}+11.04\%}$
test_rb_sample[TensorDictReplayBuffer-LazyMemmapStorage-sampler6-10000] 1.6604ms 1.4306ms 699.0059 Ops/s 767.1673 Ops/s $\textbf{\color{#d91a1a}-8.88\%}$
test_rb_sample[TensorDictReplayBuffer-LazyTensorStorage-sampler7-10000] 1.5741ms 1.3353ms 748.9123 Ops/s 811.3955 Ops/s $\textbf{\color{#d91a1a}-7.70\%}$
test_rb_sample[TensorDictPrioritizedReplayBuffer-ListStorage-None-4000] 9.9865ms 6.0915ms 164.1641 Ops/s 166.6615 Ops/s $\color{#d91a1a}-1.50\%$
test_rb_sample[TensorDictPrioritizedReplayBuffer-LazyMemmapStorage-None-10000] 2.2054ms 0.4420ms 2.2622 KOps/s 2.0552 KOps/s $\textbf{\color{#35bf28}+10.07\%}$
test_rb_sample[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-10000] 0.6184ms 0.4144ms 2.4131 KOps/s 1.9635 KOps/s $\textbf{\color{#35bf28}+22.89\%}$
test_rb_iterate[TensorDictReplayBuffer-ListStorage-RandomSampler-4000] 5.9672ms 5.8235ms 171.7187 Ops/s 170.2431 Ops/s $\color{#35bf28}+0.87\%$
test_rb_iterate[TensorDictReplayBuffer-LazyMemmapStorage-RandomSampler-10000] 1.0804ms 0.3688ms 2.7114 KOps/s 2.9432 KOps/s $\textbf{\color{#d91a1a}-7.88\%}$
test_rb_iterate[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-10000] 0.5939ms 0.3145ms 3.1797 KOps/s 3.0969 KOps/s $\color{#35bf28}+2.67\%$
test_rb_iterate[TensorDictReplayBuffer-ListStorage-SamplerWithoutReplacement-4000] 6.1296ms 5.7651ms 173.4585 Ops/s 171.1583 Ops/s $\color{#35bf28}+1.34\%$
test_rb_iterate[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-10000] 0.8748ms 0.3599ms 2.7782 KOps/s 3.0207 KOps/s $\textbf{\color{#d91a1a}-8.03\%}$
test_rb_iterate[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-10000] 0.5798ms 0.3653ms 2.7372 KOps/s 3.3901 KOps/s $\textbf{\color{#d91a1a}-19.26\%}$
test_rb_iterate[TensorDictPrioritizedReplayBuffer-ListStorage-None-4000] 6.4143ms 5.9173ms 168.9953 Ops/s 166.6994 Ops/s $\color{#35bf28}+1.38\%$
test_rb_iterate[TensorDictPrioritizedReplayBuffer-LazyMemmapStorage-None-10000] 2.3334ms 0.4523ms 2.2110 KOps/s 1.8807 KOps/s $\textbf{\color{#35bf28}+17.56\%}$
test_rb_iterate[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-10000] 0.6239ms 0.4173ms 2.3966 KOps/s 1.9085 KOps/s $\textbf{\color{#35bf28}+25.58\%}$
test_rb_populate[TensorDictReplayBuffer-ListStorage-RandomSampler-400] 0.5788s 17.1918ms 58.1673 Ops/s 197.5932 Ops/s $\textbf{\color{#d91a1a}-70.56\%}$
test_rb_populate[TensorDictReplayBuffer-LazyMemmapStorage-RandomSampler-400] 8.0807ms 1.8622ms 536.9997 Ops/s 563.6724 Ops/s $\color{#d91a1a}-4.73\%$
test_rb_populate[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-400] 10.3446ms 1.3027ms 767.6369 Ops/s 1.0597 KOps/s $\textbf{\color{#d91a1a}-27.56\%}$
test_rb_populate[TensorDictReplayBuffer-ListStorage-SamplerWithoutReplacement-400] 8.0812ms 5.0518ms 197.9494 Ops/s 196.3319 Ops/s $\color{#35bf28}+0.82\%$
test_rb_populate[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-400] 9.7886ms 1.8472ms 541.3563 Ops/s 496.6281 Ops/s $\textbf{\color{#35bf28}+9.01\%}$
test_rb_populate[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-400] 1.0434ms 0.8408ms 1.1894 KOps/s 809.6915 Ops/s $\textbf{\color{#35bf28}+46.90\%}$
test_rb_populate[TensorDictPrioritizedReplayBuffer-ListStorage-None-400] 0.5113s 15.4719ms 64.6331 Ops/s 58.1861 Ops/s $\textbf{\color{#35bf28}+11.08\%}$
test_rb_populate[TensorDictPrioritizedReplayBuffer-LazyMemmapStorage-None-400] 12.0066ms 2.0576ms 486.0023 Ops/s 498.9611 Ops/s $\color{#d91a1a}-2.60\%$
test_rb_populate[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-400] 2.2023ms 1.0781ms 927.5349 Ops/s 897.8077 Ops/s $\color{#35bf28}+3.31\%$
test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-10000-10000-100-True] 37.1657ms 34.2771ms 29.1740 Ops/s 28.4485 Ops/s $\color{#35bf28}+2.55\%$
test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-10000-10000-100-False] 19.9016ms 18.0023ms 55.5483 Ops/s 55.0197 Ops/s $\color{#35bf28}+0.96\%$
test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-100000-10000-100-True] 39.3153ms 35.5154ms 28.1568 Ops/s 27.5677 Ops/s $\color{#35bf28}+2.14\%$
test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-100000-10000-100-False] 20.1341ms 18.4386ms 54.2342 Ops/s 53.4840 Ops/s $\color{#35bf28}+1.40\%$
test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-1000000-10000-100-True] 38.7242ms 37.0005ms 27.0267 Ops/s 26.0995 Ops/s $\color{#35bf28}+3.55\%$
test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-1000000-10000-100-False] 21.0627ms 19.7068ms 50.7439 Ops/s 50.4177 Ops/s $\color{#35bf28}+0.65\%$

@github-actions
Copy link
Contributor

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

Total Benchmarks: 148. Improved: $\large\color{#35bf28}17$. Worsened: $\large\color{#d91a1a}9$.

Expand to view detailed results
Name Max Mean Ops Ops on Repo HEAD Change
test_tensor_to_bytestream_speed[pickle] 80.0668μs 79.1464μs 12.6348 KOps/s 12.5983 KOps/s $\color{#35bf28}+0.29\%$
test_tensor_to_bytestream_speed[torch.save] 0.1389ms 0.1381ms 7.2428 KOps/s 7.2993 KOps/s $\color{#d91a1a}-0.77\%$
test_tensor_to_bytestream_speed[untyped_storage] 0.1093s 0.1084s 9.2253 Ops/s 9.1512 Ops/s $\color{#35bf28}+0.81\%$
test_tensor_to_bytestream_speed[numpy] 2.4516μs 2.4475μs 408.5855 KOps/s 412.8004 KOps/s $\color{#d91a1a}-1.02\%$
test_tensor_to_bytestream_speed[safetensors] 36.7234μs 36.4248μs 27.4538 KOps/s 27.5005 KOps/s $\color{#d91a1a}-0.17\%$
test_simple 0.8905s 0.7993s 1.2510 Ops/s 1.2433 Ops/s $\color{#35bf28}+0.62\%$
test_transformed 1.5109s 1.4134s 0.7075 Ops/s 0.7100 Ops/s $\color{#d91a1a}-0.35\%$
test_serial 2.3664s 2.2805s 0.4385 Ops/s 0.4309 Ops/s $\color{#35bf28}+1.77\%$
test_parallel 2.0442s 1.9384s 0.5159 Ops/s 0.5243 Ops/s $\color{#d91a1a}-1.60\%$
test_step_mdp_speed[True-True-True-True-True] 0.3352ms 44.1731μs 22.6382 KOps/s 22.7249 KOps/s $\color{#d91a1a}-0.38\%$
test_step_mdp_speed[True-True-True-True-False] 94.2520μs 23.8445μs 41.9384 KOps/s 41.1969 KOps/s $\color{#35bf28}+1.80\%$
test_step_mdp_speed[True-True-True-False-True] 64.0210μs 24.1600μs 41.3908 KOps/s 40.6803 KOps/s $\color{#35bf28}+1.75\%$
test_step_mdp_speed[True-True-True-False-False] 55.1210μs 13.5068μs 74.0369 KOps/s 75.4021 KOps/s $\color{#d91a1a}-1.81\%$
test_step_mdp_speed[True-True-False-True-True] 85.3310μs 46.8300μs 21.3538 KOps/s 21.1941 KOps/s $\color{#35bf28}+0.75\%$
test_step_mdp_speed[True-True-False-True-False] 64.0010μs 27.2804μs 36.6563 KOps/s 36.9365 KOps/s $\color{#d91a1a}-0.76\%$
test_step_mdp_speed[True-True-False-False-True] 56.6310μs 27.2538μs 36.6922 KOps/s 36.7391 KOps/s $\color{#d91a1a}-0.13\%$
test_step_mdp_speed[True-True-False-False-False] 43.8210μs 16.4270μs 60.8752 KOps/s 61.7624 KOps/s $\color{#d91a1a}-1.44\%$
test_step_mdp_speed[True-False-True-True-True] 87.2310μs 49.8358μs 20.0659 KOps/s 20.1248 KOps/s $\color{#d91a1a}-0.29\%$
test_step_mdp_speed[True-False-True-True-False] 70.1220μs 30.3436μs 32.9559 KOps/s 33.7792 KOps/s $\color{#d91a1a}-2.44\%$
test_step_mdp_speed[True-False-True-False-True] 58.6410μs 27.5414μs 36.3089 KOps/s 36.6481 KOps/s $\color{#d91a1a}-0.93\%$
test_step_mdp_speed[True-False-True-False-False] 53.8110μs 16.4721μs 60.7089 KOps/s 61.8527 KOps/s $\color{#d91a1a}-1.85\%$
test_step_mdp_speed[True-False-False-True-True] 91.4720μs 52.3938μs 19.0862 KOps/s 19.3685 KOps/s $\color{#d91a1a}-1.46\%$
test_step_mdp_speed[True-False-False-True-False] 71.1410μs 32.8235μs 30.4660 KOps/s 30.7395 KOps/s $\color{#d91a1a}-0.89\%$
test_step_mdp_speed[True-False-False-False-True] 64.0910μs 29.9591μs 33.3789 KOps/s 33.5425 KOps/s $\color{#d91a1a}-0.49\%$
test_step_mdp_speed[True-False-False-False-False] 53.4410μs 19.1112μs 52.3254 KOps/s 53.3033 KOps/s $\color{#d91a1a}-1.83\%$
test_step_mdp_speed[False-True-True-True-True] 93.5320μs 50.0081μs 19.9968 KOps/s 20.0702 KOps/s $\color{#d91a1a}-0.37\%$
test_step_mdp_speed[False-True-True-True-False] 55.7110μs 30.2816μs 33.0233 KOps/s 33.6025 KOps/s $\color{#d91a1a}-1.72\%$
test_step_mdp_speed[False-True-True-False-True] 0.1047ms 31.0785μs 32.1766 KOps/s 32.1157 KOps/s $\color{#35bf28}+0.19\%$
test_step_mdp_speed[False-True-True-False-False] 52.9010μs 18.2379μs 54.8309 KOps/s 56.2603 KOps/s $\color{#d91a1a}-2.54\%$
test_step_mdp_speed[False-True-False-True-True] 2.7445ms 52.9222μs 18.8957 KOps/s 19.1719 KOps/s $\color{#d91a1a}-1.44\%$
test_step_mdp_speed[False-True-False-True-False] 61.8310μs 32.7526μs 30.5319 KOps/s 31.0019 KOps/s $\color{#d91a1a}-1.52\%$
test_step_mdp_speed[False-True-False-False-True] 68.9110μs 33.7168μs 29.6588 KOps/s 29.3718 KOps/s $\color{#35bf28}+0.98\%$
test_step_mdp_speed[False-True-False-False-False] 55.8910μs 20.7087μs 48.2890 KOps/s 48.6019 KOps/s $\color{#d91a1a}-0.64\%$
test_step_mdp_speed[False-False-True-True-True] 91.2420μs 54.9841μs 18.1871 KOps/s 18.4314 KOps/s $\color{#d91a1a}-1.33\%$
test_step_mdp_speed[False-False-True-True-False] 71.9710μs 35.5878μs 28.0995 KOps/s 28.6304 KOps/s $\color{#d91a1a}-1.85\%$
test_step_mdp_speed[False-False-True-False-True] 68.6010μs 33.9586μs 29.4476 KOps/s 29.3772 KOps/s $\color{#35bf28}+0.24\%$
test_step_mdp_speed[False-False-True-False-False] 60.2310μs 20.9580μs 47.7146 KOps/s 49.0671 KOps/s $\color{#d91a1a}-2.76\%$
test_step_mdp_speed[False-False-False-True-True] 91.1510μs 57.2310μs 17.4730 KOps/s 17.8294 KOps/s $\color{#d91a1a}-2.00\%$
test_step_mdp_speed[False-False-False-True-False] 89.8720μs 36.9578μs 27.0579 KOps/s 26.8854 KOps/s $\color{#35bf28}+0.64\%$
test_step_mdp_speed[False-False-False-False-True] 70.8910μs 35.8875μs 27.8649 KOps/s 27.8126 KOps/s $\color{#35bf28}+0.19\%$
test_step_mdp_speed[False-False-False-False-False] 62.3810μs 23.3335μs 42.8569 KOps/s 43.5442 KOps/s $\color{#d91a1a}-1.58\%$
test_non_tensor_env_rollout_speed[1000-single-True] 0.8553s 0.7784s 1.2847 Ops/s 1.3330 Ops/s $\color{#d91a1a}-3.62\%$
test_non_tensor_env_rollout_speed[1000-single-False] 0.7198s 0.6264s 1.5965 Ops/s 1.6124 Ops/s $\color{#d91a1a}-0.99\%$
test_non_tensor_env_rollout_speed[1000-serial-no-buffers-True] 1.7157s 1.6438s 0.6083 Ops/s 0.6086 Ops/s $\color{#d91a1a}-0.04\%$
test_non_tensor_env_rollout_speed[1000-serial-no-buffers-False] 1.5158s 1.4268s 0.7009 Ops/s 0.6942 Ops/s $\color{#35bf28}+0.97\%$
test_non_tensor_env_rollout_speed[1000-serial-buffers-True] 1.9724s 1.9003s 0.5262 Ops/s 0.5308 Ops/s $\color{#d91a1a}-0.86\%$
test_non_tensor_env_rollout_speed[1000-serial-buffers-False] 1.7487s 1.6743s 0.5973 Ops/s 0.5989 Ops/s $\color{#d91a1a}-0.27\%$
test_non_tensor_env_rollout_speed[1000-parallel-no-buffers-True] 4.7108s 4.6057s 0.2171 Ops/s 0.2164 Ops/s $\color{#35bf28}+0.34\%$
test_non_tensor_env_rollout_speed[1000-parallel-no-buffers-False] 4.4723s 4.4232s 0.2261 Ops/s 0.2273 Ops/s $\color{#d91a1a}-0.52\%$
test_non_tensor_env_rollout_speed[1000-parallel-buffers-True] 2.0115s 1.9598s 0.5103 Ops/s 0.5181 Ops/s $\color{#d91a1a}-1.51\%$
test_non_tensor_env_rollout_speed[1000-parallel-buffers-False] 1.7095s 1.6335s 0.6122 Ops/s 0.6047 Ops/s $\color{#35bf28}+1.23\%$
test_values[generalized_advantage_estimate-True-True] 20.6851ms 20.0740ms 49.8158 Ops/s 47.8646 Ops/s $\color{#35bf28}+4.08\%$
test_values[vec_generalized_advantage_estimate-True-True] 0.1505s 3.9264ms 254.6893 Ops/s 263.8428 Ops/s $\color{#d91a1a}-3.47\%$
test_values[td0_return_estimate-False-False] 0.1083ms 83.2405μs 12.0134 KOps/s 12.0811 KOps/s $\color{#d91a1a}-0.56\%$
test_values[td1_return_estimate-False-False] 50.3510ms 48.2511ms 20.7249 Ops/s 19.7549 Ops/s $\color{#35bf28}+4.91\%$
test_values[vec_td1_return_estimate-False-False] 1.3575ms 1.0847ms 921.9161 Ops/s 924.0249 Ops/s $\color{#d91a1a}-0.23\%$
test_values[td_lambda_return_estimate-True-False] 82.1661ms 78.8329ms 12.6851 Ops/s 12.3123 Ops/s $\color{#35bf28}+3.03\%$
test_values[vec_td_lambda_return_estimate-True-False] 1.2569ms 1.0795ms 926.3572 Ops/s 922.6791 Ops/s $\color{#35bf28}+0.40\%$
test_gae_speed[generalized_advantage_estimate-False-1-512] 20.5748ms 20.3891ms 49.0457 Ops/s 46.0184 Ops/s $\textbf{\color{#35bf28}+6.58\%}$
test_gae_speed[vec_generalized_advantage_estimate-True-1-512] 1.0190ms 0.7460ms 1.3405 KOps/s 1.3466 KOps/s $\color{#d91a1a}-0.45\%$
test_gae_speed[vec_generalized_advantage_estimate-False-1-512] 0.7571ms 0.6815ms 1.4673 KOps/s 1.4212 KOps/s $\color{#35bf28}+3.25\%$
test_gae_speed[vec_generalized_advantage_estimate-True-32-512] 1.5759ms 1.4841ms 673.7934 Ops/s 664.9210 Ops/s $\color{#35bf28}+1.33\%$
test_gae_speed[vec_generalized_advantage_estimate-False-32-512] 0.7620ms 0.6874ms 1.4548 KOps/s 1.3980 KOps/s $\color{#35bf28}+4.06\%$
test_dqn_speed[False-None] 1.6174ms 1.5163ms 659.5146 Ops/s 649.6032 Ops/s $\color{#35bf28}+1.53\%$
test_dqn_speed[False-backward] 2.2065ms 2.1525ms 464.5689 Ops/s 458.0660 Ops/s $\color{#35bf28}+1.42\%$
test_dqn_speed[True-None] 1.1154ms 0.5594ms 1.7876 KOps/s 1.7413 KOps/s $\color{#35bf28}+2.66\%$
test_dqn_speed[True-backward] 1.1771ms 1.0996ms 909.3866 Ops/s 826.8140 Ops/s $\textbf{\color{#35bf28}+9.99\%}$
test_dqn_speed[reduce-overhead-None] 0.7426ms 0.6022ms 1.6606 KOps/s 1.5751 KOps/s $\textbf{\color{#35bf28}+5.43\%}$
test_ddpg_speed[False-None] 3.3112ms 2.8711ms 348.2934 Ops/s 334.2682 Ops/s $\color{#35bf28}+4.20\%$
test_ddpg_speed[False-backward] 4.5713ms 4.1639ms 240.1591 Ops/s 230.3048 Ops/s $\color{#35bf28}+4.28\%$
test_ddpg_speed[True-None] 1.4109ms 1.3122ms 762.0904 Ops/s 754.5719 Ops/s $\color{#35bf28}+1.00\%$
test_ddpg_speed[True-backward] 2.4503ms 2.3936ms 417.7888 Ops/s 392.4719 Ops/s $\textbf{\color{#35bf28}+6.45\%}$
test_ddpg_speed[reduce-overhead-None] 1.4634ms 1.3506ms 740.4328 Ops/s 738.2295 Ops/s $\color{#35bf28}+0.30\%$
test_sac_speed[False-None] 8.8326ms 8.3252ms 120.1170 Ops/s 120.8103 Ops/s $\color{#d91a1a}-0.57\%$
test_sac_speed[False-backward] 12.0712ms 11.3233ms 88.3138 Ops/s 86.6774 Ops/s $\color{#35bf28}+1.89\%$
test_sac_speed[True-None] 1.9882ms 1.8957ms 527.5234 Ops/s 547.6659 Ops/s $\color{#d91a1a}-3.68\%$
test_sac_speed[True-backward] 4.1321ms 3.6552ms 273.5838 Ops/s 272.9379 Ops/s $\color{#35bf28}+0.24\%$
test_sac_speed[reduce-overhead-None] 18.1962ms 10.3848ms 96.2950 Ops/s 96.5697 Ops/s $\color{#d91a1a}-0.28\%$
test_redq_deprec_speed[False-None] 9.8530ms 9.2739ms 107.8296 Ops/s 107.8752 Ops/s $\color{#d91a1a}-0.04\%$
test_redq_deprec_speed[False-backward] 13.0785ms 12.6138ms 79.2782 Ops/s 79.3905 Ops/s $\color{#d91a1a}-0.14\%$
test_redq_deprec_speed[True-None] 2.6309ms 2.5494ms 392.2469 Ops/s 393.5451 Ops/s $\color{#d91a1a}-0.33\%$
test_redq_deprec_speed[True-backward] 4.8763ms 4.3747ms 228.5854 Ops/s 228.2024 Ops/s $\color{#35bf28}+0.17\%$
test_redq_deprec_speed[reduce-overhead-None] 15.3833ms 9.4848ms 105.4320 Ops/s 89.3308 Ops/s $\textbf{\color{#35bf28}+18.02\%}$
test_td3_speed[False-None] 8.2287ms 8.1717ms 122.3741 Ops/s 122.2628 Ops/s $\color{#35bf28}+0.09\%$
test_td3_speed[False-backward] 11.2897ms 10.7723ms 92.8303 Ops/s 92.6419 Ops/s $\color{#35bf28}+0.20\%$
test_td3_speed[True-None] 1.7048ms 1.6824ms 594.3836 Ops/s 574.9839 Ops/s $\color{#35bf28}+3.37\%$
test_td3_speed[True-backward] 3.8111ms 3.3335ms 299.9860 Ops/s 315.1683 Ops/s $\color{#d91a1a}-4.82\%$
test_td3_speed[reduce-overhead-None] 53.7501ms 23.1070ms 43.2769 Ops/s 42.0084 Ops/s $\color{#35bf28}+3.02\%$
test_cql_speed[False-None] 17.4424ms 17.1348ms 58.3609 Ops/s 58.2710 Ops/s $\color{#35bf28}+0.15\%$
test_cql_speed[False-backward] 23.1583ms 22.7144ms 44.0250 Ops/s 44.6697 Ops/s $\color{#d91a1a}-1.44\%$
test_cql_speed[True-None] 3.4626ms 3.3690ms 296.8265 Ops/s 293.7303 Ops/s $\color{#35bf28}+1.05\%$
test_cql_speed[True-backward] 6.0998ms 5.6672ms 176.4550 Ops/s 173.7237 Ops/s $\color{#35bf28}+1.57\%$
test_cql_speed[reduce-overhead-None] 18.1450ms 11.4632ms 87.2355 Ops/s 87.2634 Ops/s $\color{#d91a1a}-0.03\%$
test_a2c_speed[False-None] 3.9088ms 3.2052ms 311.9920 Ops/s 311.3559 Ops/s $\color{#35bf28}+0.20\%$
test_a2c_speed[False-backward] 6.7823ms 6.3617ms 157.1898 Ops/s 155.4671 Ops/s $\color{#35bf28}+1.11\%$
test_a2c_speed[True-None] 1.4360ms 1.3316ms 750.9965 Ops/s 740.5779 Ops/s $\color{#35bf28}+1.41\%$
test_a2c_speed[True-backward] 3.4383ms 3.1622ms 316.2313 Ops/s 319.1776 Ops/s $\color{#d91a1a}-0.92\%$
test_a2c_speed[reduce-overhead-None] 1.0392ms 0.9379ms 1.0662 KOps/s 1.0497 KOps/s $\color{#35bf28}+1.57\%$
test_ppo_speed[False-None] 3.9064ms 3.8013ms 263.0706 Ops/s 262.7424 Ops/s $\color{#35bf28}+0.12\%$
test_ppo_speed[False-backward] 7.5227ms 7.1448ms 139.9625 Ops/s 139.1628 Ops/s $\color{#35bf28}+0.57\%$
test_ppo_speed[True-None] 1.4907ms 1.3901ms 719.3765 Ops/s 699.6277 Ops/s $\color{#35bf28}+2.82\%$
test_ppo_speed[True-backward] 3.3885ms 3.3276ms 300.5191 Ops/s 313.3248 Ops/s $\color{#d91a1a}-4.09\%$
test_ppo_speed[reduce-overhead-None] 1.0827ms 1.0100ms 990.1099 Ops/s 969.7585 Ops/s $\color{#35bf28}+2.10\%$
test_reinforce_speed[False-None] 2.3413ms 2.2631ms 441.8815 Ops/s 436.9746 Ops/s $\color{#35bf28}+1.12\%$
test_reinforce_speed[False-backward] 3.4336ms 3.3875ms 295.2018 Ops/s 294.3939 Ops/s $\color{#35bf28}+0.27\%$
test_reinforce_speed[True-None] 1.3349ms 1.2555ms 796.4827 Ops/s 766.8917 Ops/s $\color{#35bf28}+3.86\%$
test_reinforce_speed[True-backward] 3.2975ms 3.1606ms 316.3939 Ops/s 321.7567 Ops/s $\color{#d91a1a}-1.67\%$
test_reinforce_speed[reduce-overhead-None] 0.4620s 9.9834ms 100.1659 Ops/s 101.4008 Ops/s $\color{#d91a1a}-1.22\%$
test_iql_speed[False-None] 10.1364ms 9.4095ms 106.2752 Ops/s 105.7595 Ops/s $\color{#35bf28}+0.49\%$
test_iql_speed[False-backward] 13.9848ms 13.4388ms 74.4113 Ops/s 74.2646 Ops/s $\color{#35bf28}+0.20\%$
test_iql_speed[True-None] 2.4489ms 2.2165ms 451.1585 Ops/s 451.0793 Ops/s $\color{#35bf28}+0.02\%$
test_iql_speed[True-backward] 5.3980ms 5.0470ms 198.1376 Ops/s 204.3585 Ops/s $\color{#d91a1a}-3.04\%$
test_iql_speed[reduce-overhead-None] 0.7029s 12.2220ms 81.8194 Ops/s 76.8788 Ops/s $\textbf{\color{#35bf28}+6.43\%}$
test_rb_sample[TensorDictReplayBuffer-ListStorage-RandomSampler-4000] 6.2562ms 5.8417ms 171.1817 Ops/s 170.6355 Ops/s $\color{#35bf28}+0.32\%$
test_rb_sample[TensorDictReplayBuffer-LazyMemmapStorage-RandomSampler-10000] 0.8925ms 0.3380ms 2.9587 KOps/s 2.8654 KOps/s $\color{#35bf28}+3.26\%$
test_rb_sample[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-10000] 0.5881ms 0.2928ms 3.4150 KOps/s 3.0045 KOps/s $\textbf{\color{#35bf28}+13.66\%}$
test_rb_sample[TensorDictReplayBuffer-ListStorage-SamplerWithoutReplacement-4000] 5.9413ms 5.6001ms 178.5692 Ops/s 177.3981 Ops/s $\color{#35bf28}+0.66\%$
test_rb_sample[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-10000] 2.3379ms 0.3208ms 3.1171 KOps/s 3.5608 KOps/s $\textbf{\color{#d91a1a}-12.46\%}$
test_rb_sample[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-10000] 0.5274ms 0.3329ms 3.0040 KOps/s 3.8237 KOps/s $\textbf{\color{#d91a1a}-21.44\%}$
test_rb_sample[TensorDictReplayBuffer-LazyMemmapStorage-sampler6-10000] 1.5908ms 1.3938ms 717.4716 Ops/s 785.3088 Ops/s $\textbf{\color{#d91a1a}-8.64\%}$
test_rb_sample[TensorDictReplayBuffer-LazyTensorStorage-sampler7-10000] 1.6150ms 1.3129ms 761.6855 Ops/s 845.3702 Ops/s $\textbf{\color{#d91a1a}-9.90\%}$
test_rb_sample[TensorDictPrioritizedReplayBuffer-ListStorage-None-4000] 5.8616ms 5.7513ms 173.8734 Ops/s 171.9650 Ops/s $\color{#35bf28}+1.11\%$
test_rb_sample[TensorDictPrioritizedReplayBuffer-LazyMemmapStorage-None-10000] 0.8527ms 0.5007ms 1.9971 KOps/s 1.9241 KOps/s $\color{#35bf28}+3.79\%$
test_rb_sample[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-10000] 0.7320ms 0.4823ms 2.0733 KOps/s 1.9685 KOps/s $\textbf{\color{#35bf28}+5.32\%}$
test_rb_iterate[TensorDictReplayBuffer-ListStorage-RandomSampler-4000] 5.8220ms 5.6558ms 176.8091 Ops/s 176.0357 Ops/s $\color{#35bf28}+0.44\%$
test_rb_iterate[TensorDictReplayBuffer-LazyMemmapStorage-RandomSampler-10000] 1.1304ms 0.3122ms 3.2034 KOps/s 3.4497 KOps/s $\textbf{\color{#d91a1a}-7.14\%}$
test_rb_iterate[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-10000] 0.5683ms 0.3216ms 3.1096 KOps/s 3.7006 KOps/s $\textbf{\color{#d91a1a}-15.97\%}$
test_rb_iterate[TensorDictReplayBuffer-ListStorage-SamplerWithoutReplacement-4000] 5.8586ms 5.5713ms 179.4916 Ops/s 176.6939 Ops/s $\color{#35bf28}+1.58\%$
test_rb_iterate[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-10000] 1.8700ms 0.3596ms 2.7809 KOps/s 3.5743 KOps/s $\textbf{\color{#d91a1a}-22.20\%}$
test_rb_iterate[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-10000] 0.6019ms 0.3472ms 2.8805 KOps/s 2.6796 KOps/s $\textbf{\color{#35bf28}+7.50\%}$
test_rb_iterate[TensorDictPrioritizedReplayBuffer-ListStorage-None-4000] 5.8260ms 5.7352ms 174.3623 Ops/s 171.5570 Ops/s $\color{#35bf28}+1.64\%$
test_rb_iterate[TensorDictPrioritizedReplayBuffer-LazyMemmapStorage-None-10000] 2.0216ms 0.4329ms 2.3098 KOps/s 2.1068 KOps/s $\textbf{\color{#35bf28}+9.64\%}$
test_rb_iterate[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-10000] 0.6334ms 0.4140ms 2.4154 KOps/s 2.1607 KOps/s $\textbf{\color{#35bf28}+11.79\%}$
test_rb_populate[TensorDictReplayBuffer-ListStorage-RandomSampler-400] 6.5434ms 4.9236ms 203.1017 Ops/s 51.4527 Ops/s $\textbf{\color{#35bf28}+294.73\%}$
test_rb_populate[TensorDictReplayBuffer-LazyMemmapStorage-RandomSampler-400] 5.1080ms 2.1608ms 462.7971 Ops/s 548.6925 Ops/s $\textbf{\color{#d91a1a}-15.65\%}$
test_rb_populate[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-400] 2.1071ms 1.1276ms 886.8334 Ops/s 820.8934 Ops/s $\textbf{\color{#35bf28}+8.03\%}$
test_rb_populate[TensorDictReplayBuffer-ListStorage-SamplerWithoutReplacement-400] 0.5875s 16.6988ms 59.8844 Ops/s 198.6873 Ops/s $\textbf{\color{#d91a1a}-69.86\%}$
test_rb_populate[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-400] 4.0187ms 1.8352ms 544.9088 Ops/s 502.7860 Ops/s $\textbf{\color{#35bf28}+8.38\%}$
test_rb_populate[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-400] 1.1428ms 0.9049ms 1.1052 KOps/s 865.9699 Ops/s $\textbf{\color{#35bf28}+27.62\%}$
test_rb_populate[TensorDictPrioritizedReplayBuffer-ListStorage-None-400] 9.3697ms 5.2318ms 191.1401 Ops/s 51.3246 Ops/s $\textbf{\color{#35bf28}+272.41\%}$
test_rb_populate[TensorDictPrioritizedReplayBuffer-LazyMemmapStorage-None-400] 9.5073ms 1.9893ms 502.6901 Ops/s 509.3102 Ops/s $\color{#d91a1a}-1.30\%$
test_rb_populate[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-400] 3.5804ms 1.1231ms 890.4171 Ops/s 676.7165 Ops/s $\textbf{\color{#35bf28}+31.58\%}$
test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-10000-10000-100-True] 37.6057ms 33.9415ms 29.4624 Ops/s 29.0836 Ops/s $\color{#35bf28}+1.30\%$
test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-10000-10000-100-False] 18.9983ms 17.5272ms 57.0543 Ops/s 56.0395 Ops/s $\color{#35bf28}+1.81\%$
test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-100000-10000-100-True] 39.0873ms 35.1707ms 28.4328 Ops/s 28.4322 Ops/s $+0.00\%$
test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-100000-10000-100-False] 19.6507ms 17.9871ms 55.5955 Ops/s 55.5815 Ops/s $\color{#35bf28}+0.03\%$
test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-1000000-10000-100-True] 37.1131ms 36.7202ms 27.2330 Ops/s 26.8106 Ops/s $\color{#35bf28}+1.58\%$
test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-1000000-10000-100-False] 20.6899ms 19.5366ms 51.1860 Ops/s 51.4833 Ops/s $\color{#d91a1a}-0.58\%$

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.

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant