Skip to content

Conversation

@vmoens
Copy link
Collaborator

@vmoens vmoens commented Dec 31, 2025

Stack from ghstack (oldest at bottom):

[ghstack-poisoned]
@pytorch-bot
Copy link

pytorch-bot bot commented Dec 31, 2025

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/rl/3286

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

❗ 1 Active SEVs

There are 1 currently active SEVs. If your PR is affected, please view them below:

❌ 3 New Failures, 1 Cancelled Job

As of commit e09df60 with merge base 7866d11 (image):

NEW FAILURES - The following jobs have failed:

CANCELLED JOB - The following job was cancelled. Please retry:

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

vmoens added a commit that referenced this pull request Dec 31, 2025
ghstack-source-id: 2e1839f
Pull-Request: #3286
@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 Dec 31, 2025
@github-actions
Copy link

github-actions bot commented Dec 31, 2025

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

Total Benchmarks: 164. Improved: $\large\color{#35bf28}21$. Worsened: $\large\color{#d91a1a}11$.

Expand to view detailed results
Name Max Mean Ops Ops on Repo HEAD Change
test_tensor_to_bytestream_speed[pickle] 80.6607μs 78.7569μs 12.6973 KOps/s 12.7928 KOps/s $\color{#d91a1a}-0.75\%$
test_tensor_to_bytestream_speed[torch.save] 0.1378ms 0.1375ms 7.2715 KOps/s 7.4505 KOps/s $\color{#d91a1a}-2.40\%$
test_tensor_to_bytestream_speed[untyped_storage] 0.1127s 0.1124s 8.8976 Ops/s 8.7910 Ops/s $\color{#35bf28}+1.21\%$
test_tensor_to_bytestream_speed[numpy] 2.6250μs 2.6081μs 383.4170 KOps/s 370.7444 KOps/s $\color{#35bf28}+3.42\%$
test_tensor_to_bytestream_speed[safetensors] 37.9119μs 37.7308μs 26.5036 KOps/s 26.7834 KOps/s $\color{#d91a1a}-1.04\%$
test_simple 0.5318s 0.5265s 1.8994 Ops/s 1.8160 Ops/s $\color{#35bf28}+4.60\%$
test_transformed 1.0708s 1.0702s 0.9344 Ops/s 0.9080 Ops/s $\color{#35bf28}+2.91\%$
test_serial 1.5878s 1.5862s 0.6304 Ops/s 0.6207 Ops/s $\color{#35bf28}+1.57\%$
test_parallel 1.1896s 1.0983s 0.9105 Ops/s 0.8559 Ops/s $\textbf{\color{#35bf28}+6.38\%}$
test_step_mdp_speed[True-True-True-True-True] 0.1548ms 43.3986μs 23.0422 KOps/s 22.5654 KOps/s $\color{#35bf28}+2.11\%$
test_step_mdp_speed[True-True-True-True-False] 49.4210μs 24.4497μs 40.9004 KOps/s 41.0379 KOps/s $\color{#d91a1a}-0.34\%$
test_step_mdp_speed[True-True-True-False-True] 65.0710μs 24.0757μs 41.5356 KOps/s 40.4931 KOps/s $\color{#35bf28}+2.57\%$
test_step_mdp_speed[True-True-True-False-False] 47.9300μs 13.5527μs 73.7859 KOps/s 73.3435 KOps/s $\color{#35bf28}+0.60\%$
test_step_mdp_speed[True-True-False-True-True] 77.4610μs 46.1526μs 21.6672 KOps/s 21.3814 KOps/s $\color{#35bf28}+1.34\%$
test_step_mdp_speed[True-True-False-True-False] 55.7910μs 27.0346μs 36.9896 KOps/s 36.7850 KOps/s $\color{#35bf28}+0.56\%$
test_step_mdp_speed[True-True-False-False-True] 63.8210μs 27.4173μs 36.4734 KOps/s 36.4218 KOps/s $\color{#35bf28}+0.14\%$
test_step_mdp_speed[True-True-False-False-False] 40.5810μs 15.9804μs 62.5766 KOps/s 61.9913 KOps/s $\color{#35bf28}+0.94\%$
test_step_mdp_speed[True-False-True-True-True] 81.6010μs 48.4912μs 20.6223 KOps/s 20.3464 KOps/s $\color{#35bf28}+1.36\%$
test_step_mdp_speed[True-False-True-True-False] 57.1310μs 29.4071μs 34.0053 KOps/s 33.4085 KOps/s $\color{#35bf28}+1.79\%$
test_step_mdp_speed[True-False-True-False-True] 62.3110μs 26.5493μs 37.6658 KOps/s 36.3946 KOps/s $\color{#35bf28}+3.49\%$
test_step_mdp_speed[True-False-True-False-False] 40.8400μs 15.7054μs 63.6725 KOps/s 61.9786 KOps/s $\color{#35bf28}+2.73\%$
test_step_mdp_speed[True-False-False-True-True] 83.2910μs 50.8482μs 19.6664 KOps/s 19.3080 KOps/s $\color{#35bf28}+1.86\%$
test_step_mdp_speed[True-False-False-True-False] 80.0010μs 31.2378μs 32.0125 KOps/s 30.6960 KOps/s $\color{#35bf28}+4.29\%$
test_step_mdp_speed[True-False-False-False-True] 60.0610μs 28.5419μs 35.0362 KOps/s 33.6431 KOps/s $\color{#35bf28}+4.14\%$
test_step_mdp_speed[True-False-False-False-False] 50.2310μs 18.3830μs 54.3980 KOps/s 53.9105 KOps/s $\color{#35bf28}+0.90\%$
test_step_mdp_speed[False-True-True-True-True] 80.8810μs 48.6851μs 20.5402 KOps/s 20.4103 KOps/s $\color{#35bf28}+0.64\%$
test_step_mdp_speed[False-True-True-True-False] 65.1610μs 29.3097μs 34.1184 KOps/s 33.7605 KOps/s $\color{#35bf28}+1.06\%$
test_step_mdp_speed[False-True-True-False-True] 2.4784ms 30.7659μs 32.5035 KOps/s 31.7658 KOps/s $\color{#35bf28}+2.32\%$
test_step_mdp_speed[False-True-True-False-False] 44.8600μs 17.6363μs 56.7012 KOps/s 55.5597 KOps/s $\color{#35bf28}+2.05\%$
test_step_mdp_speed[False-True-False-True-True] 90.9210μs 51.2796μs 19.5009 KOps/s 19.7094 KOps/s $\color{#d91a1a}-1.06\%$
test_step_mdp_speed[False-True-False-True-False] 58.7810μs 31.7939μs 31.4526 KOps/s 30.4005 KOps/s $\color{#35bf28}+3.46\%$
test_step_mdp_speed[False-True-False-False-True] 73.1710μs 31.9728μs 31.2766 KOps/s 29.6447 KOps/s $\textbf{\color{#35bf28}+5.51\%}$
test_step_mdp_speed[False-True-False-False-False] 46.3300μs 20.1033μs 49.7431 KOps/s 48.1522 KOps/s $\color{#35bf28}+3.30\%$
test_step_mdp_speed[False-False-True-True-True] 96.9410μs 53.4271μs 18.7171 KOps/s 18.1262 KOps/s $\color{#35bf28}+3.26\%$
test_step_mdp_speed[False-False-True-True-False] 80.0310μs 34.9481μs 28.6138 KOps/s 28.5486 KOps/s $\color{#35bf28}+0.23\%$
test_step_mdp_speed[False-False-True-False-True] 67.1110μs 33.2261μs 30.0968 KOps/s 29.6235 KOps/s $\color{#35bf28}+1.60\%$
test_step_mdp_speed[False-False-True-False-False] 54.5710μs 20.3623μs 49.1103 KOps/s 48.1565 KOps/s $\color{#35bf28}+1.98\%$
test_step_mdp_speed[False-False-False-True-True] 96.8510μs 55.8015μs 17.9207 KOps/s 17.8010 KOps/s $\color{#35bf28}+0.67\%$
test_step_mdp_speed[False-False-False-True-False] 77.0910μs 36.4591μs 27.4280 KOps/s 26.6928 KOps/s $\color{#35bf28}+2.75\%$
test_step_mdp_speed[False-False-False-False-True] 61.1310μs 34.2977μs 29.1565 KOps/s 28.1631 KOps/s $\color{#35bf28}+3.53\%$
test_step_mdp_speed[False-False-False-False-False] 58.5510μs 22.7343μs 43.9864 KOps/s 43.0817 KOps/s $\color{#35bf28}+2.10\%$
test_non_tensor_env_rollout_speed[1000-single-True] 0.8585s 0.7630s 1.3106 Ops/s 1.3116 Ops/s $\color{#d91a1a}-0.08\%$
test_non_tensor_env_rollout_speed[1000-single-False] 0.7312s 0.6291s 1.5897 Ops/s 1.6004 Ops/s $\color{#d91a1a}-0.67\%$
test_non_tensor_env_rollout_speed[1000-serial-no-buffers-True] 1.7371s 1.6561s 0.6038 Ops/s 0.6008 Ops/s $\color{#35bf28}+0.50\%$
test_non_tensor_env_rollout_speed[1000-serial-no-buffers-False] 1.5223s 1.4407s 0.6941 Ops/s 0.6899 Ops/s $\color{#35bf28}+0.61\%$
test_non_tensor_env_rollout_speed[1000-serial-buffers-True] 1.9788s 1.8941s 0.5280 Ops/s 0.5293 Ops/s $\color{#d91a1a}-0.25\%$
test_non_tensor_env_rollout_speed[1000-serial-buffers-False] 1.7683s 1.6782s 0.5959 Ops/s 0.5936 Ops/s $\color{#35bf28}+0.38\%$
test_non_tensor_env_rollout_speed[1000-parallel-no-buffers-True] 4.6365s 4.5070s 0.2219 Ops/s 0.2148 Ops/s $\color{#35bf28}+3.29\%$
test_non_tensor_env_rollout_speed[1000-parallel-no-buffers-False] 4.4766s 4.3720s 0.2287 Ops/s 0.2304 Ops/s $\color{#d91a1a}-0.73\%$
test_non_tensor_env_rollout_speed[1000-parallel-buffers-True] 2.0656s 1.9397s 0.5155 Ops/s 0.5147 Ops/s $\color{#35bf28}+0.17\%$
test_non_tensor_env_rollout_speed[1000-parallel-buffers-False] 1.7592s 1.6483s 0.6067 Ops/s 0.6062 Ops/s $\color{#35bf28}+0.08\%$
test_values[generalized_advantage_estimate-True-True] 9.8448ms 9.5656ms 104.5411 Ops/s 106.4311 Ops/s $\color{#d91a1a}-1.78\%$
test_values[vec_generalized_advantage_estimate-True-True] 19.4540ms 17.7406ms 56.3678 Ops/s 56.6116 Ops/s $\color{#d91a1a}-0.43\%$
test_values[td0_return_estimate-False-False] 0.2028ms 0.1239ms 8.0715 KOps/s 8.0155 KOps/s $\color{#35bf28}+0.70\%$
test_values[td1_return_estimate-False-False] 26.1609ms 25.4362ms 39.3140 Ops/s 39.6753 Ops/s $\color{#d91a1a}-0.91\%$
test_values[vec_td1_return_estimate-False-False] 18.3037ms 17.8885ms 55.9020 Ops/s 56.4306 Ops/s $\color{#d91a1a}-0.94\%$
test_values[td_lambda_return_estimate-True-False] 39.7447ms 38.2601ms 26.1369 Ops/s 26.7808 Ops/s $\color{#d91a1a}-2.40\%$
test_values[vec_td_lambda_return_estimate-True-False] 18.1111ms 17.8665ms 55.9708 Ops/s 56.2856 Ops/s $\color{#d91a1a}-0.56\%$
test_gae_speed[generalized_advantage_estimate-False-1-512] 8.9523ms 8.6110ms 116.1299 Ops/s 119.9997 Ops/s $\color{#d91a1a}-3.22\%$
test_gae_speed[vec_generalized_advantage_estimate-True-1-512] 1.8433ms 1.5438ms 647.7560 Ops/s 648.7006 Ops/s $\color{#d91a1a}-0.15\%$
test_gae_speed[vec_generalized_advantage_estimate-False-1-512] 0.6311ms 0.3902ms 2.5629 KOps/s 2.5544 KOps/s $\color{#35bf28}+0.34\%$
test_gae_speed[vec_generalized_advantage_estimate-True-32-512] 34.9595ms 34.1830ms 29.2543 Ops/s 28.5141 Ops/s $\color{#35bf28}+2.60\%$
test_gae_speed[vec_generalized_advantage_estimate-False-32-512] 2.0193ms 1.6873ms 592.6753 Ops/s 590.6771 Ops/s $\color{#35bf28}+0.34\%$
test_dqn_speed[False-None] 1.6990ms 1.3552ms 737.9016 Ops/s 737.5123 Ops/s $\color{#35bf28}+0.05\%$
test_dqn_speed[False-backward] 1.9721ms 1.8590ms 537.9199 Ops/s 526.0057 Ops/s $\color{#35bf28}+2.27\%$
test_dqn_speed[True-None] 0.6640ms 0.5237ms 1.9093 KOps/s 1.7967 KOps/s $\textbf{\color{#35bf28}+6.27\%}$
test_dqn_speed[True-backward] 1.0446ms 0.9589ms 1.0429 KOps/s 1.0265 KOps/s $\color{#35bf28}+1.59\%$
test_dqn_speed[reduce-overhead-None] 0.6818ms 0.5109ms 1.9574 KOps/s 1.8591 KOps/s $\textbf{\color{#35bf28}+5.29\%}$
test_dqn_speed[reduce-overhead-backward] 0.9889ms 0.9391ms 1.0648 KOps/s 874.5016 Ops/s $\textbf{\color{#35bf28}+21.76\%}$
test_ddpg_speed[False-None] 3.1922ms 2.7779ms 359.9882 Ops/s 353.5464 Ops/s $\color{#35bf28}+1.82\%$
test_ddpg_speed[False-backward] 4.0186ms 3.9257ms 254.7337 Ops/s 254.5096 Ops/s $\color{#35bf28}+0.09\%$
test_ddpg_speed[True-None] 1.7238ms 1.3523ms 739.4948 Ops/s 715.3529 Ops/s $\color{#35bf28}+3.37\%$
test_ddpg_speed[True-backward] 2.3539ms 2.2946ms 435.7989 Ops/s 367.6617 Ops/s $\textbf{\color{#35bf28}+18.53\%}$
test_ddpg_speed[reduce-overhead-None] 1.6785ms 1.3372ms 747.8480 Ops/s 705.5035 Ops/s $\textbf{\color{#35bf28}+6.00\%}$
test_ddpg_speed[reduce-overhead-backward] 2.3490ms 2.2849ms 437.6593 Ops/s 407.0704 Ops/s $\textbf{\color{#35bf28}+7.51\%}$
test_sac_speed[False-None] 8.0247ms 7.5518ms 132.4180 Ops/s 130.0024 Ops/s $\color{#35bf28}+1.86\%$
test_sac_speed[False-backward] 11.1466ms 10.7810ms 92.7554 Ops/s 90.5824 Ops/s $\color{#35bf28}+2.40\%$
test_sac_speed[True-None] 2.3378ms 2.0992ms 476.3832 Ops/s 460.3219 Ops/s $\color{#35bf28}+3.49\%$
test_sac_speed[True-backward] 4.1570ms 3.8933ms 256.8503 Ops/s 242.3612 Ops/s $\textbf{\color{#35bf28}+5.98\%}$
test_sac_speed[reduce-overhead-None] 2.4744ms 2.0699ms 483.1169 Ops/s 459.6062 Ops/s $\textbf{\color{#35bf28}+5.12\%}$
test_sac_speed[reduce-overhead-backward] 4.1420ms 3.9436ms 253.5736 Ops/s 218.2286 Ops/s $\textbf{\color{#35bf28}+16.20\%}$
test_redq_speed[False-None] 14.8192ms 10.3823ms 96.3177 Ops/s 97.2396 Ops/s $\color{#d91a1a}-0.95\%$
test_redq_speed[False-backward] 18.5814ms 17.7179ms 56.4401 Ops/s 58.2367 Ops/s $\color{#d91a1a}-3.09\%$
test_redq_speed[True-None] 4.7279ms 4.4585ms 224.2929 Ops/s 225.7651 Ops/s $\color{#d91a1a}-0.65\%$
test_redq_speed[True-backward] 10.3319ms 9.8451ms 101.5734 Ops/s 101.1473 Ops/s $\color{#35bf28}+0.42\%$
test_redq_speed[reduce-overhead-None] 4.8452ms 4.3775ms 228.4399 Ops/s 232.4606 Ops/s $\color{#d91a1a}-1.73\%$
test_redq_speed[reduce-overhead-backward] 10.5671ms 10.1240ms 98.7756 Ops/s 105.0323 Ops/s $\textbf{\color{#d91a1a}-5.96\%}$
test_redq_deprec_speed[False-None] 11.6592ms 11.1887ms 89.3762 Ops/s 93.0592 Ops/s $\color{#d91a1a}-3.96\%$
test_redq_deprec_speed[False-backward] 16.5283ms 15.9972ms 62.5109 Ops/s 64.8961 Ops/s $\color{#d91a1a}-3.68\%$
test_redq_deprec_speed[True-None] 4.0388ms 3.6685ms 272.5878 Ops/s 280.1267 Ops/s $\color{#d91a1a}-2.69\%$
test_redq_deprec_speed[True-backward] 8.0370ms 7.7823ms 128.4971 Ops/s 135.3857 Ops/s $\textbf{\color{#d91a1a}-5.09\%}$
test_redq_deprec_speed[reduce-overhead-None] 4.0216ms 3.6315ms 275.3719 Ops/s 278.8382 Ops/s $\color{#d91a1a}-1.24\%$
test_redq_deprec_speed[reduce-overhead-backward] 7.9098ms 7.6391ms 130.9054 Ops/s 113.1350 Ops/s $\textbf{\color{#35bf28}+15.71\%}$
test_td3_speed[False-None] 8.6316ms 7.7724ms 128.6611 Ops/s 129.5695 Ops/s $\color{#d91a1a}-0.70\%$
test_td3_speed[False-backward] 10.8614ms 10.4749ms 95.4659 Ops/s 94.3009 Ops/s $\color{#35bf28}+1.24\%$
test_td3_speed[True-None] 1.8682ms 1.7962ms 556.7409 Ops/s 543.6375 Ops/s $\color{#35bf28}+2.41\%$
test_td3_speed[True-backward] 3.8687ms 3.6183ms 276.3716 Ops/s 271.0223 Ops/s $\color{#35bf28}+1.97\%$
test_td3_speed[reduce-overhead-None] 1.8460ms 1.7763ms 562.9789 Ops/s 545.1965 Ops/s $\color{#35bf28}+3.26\%$
test_td3_speed[reduce-overhead-backward] 3.7656ms 3.6272ms 275.6965 Ops/s 250.1046 Ops/s $\textbf{\color{#35bf28}+10.23\%}$
test_cql_speed[False-None] 30.0480ms 26.2661ms 38.0719 Ops/s 38.9104 Ops/s $\color{#d91a1a}-2.16\%$
test_cql_speed[False-backward] 40.5934ms 35.2990ms 28.3295 Ops/s 28.7991 Ops/s $\color{#d91a1a}-1.63\%$
test_cql_speed[True-None] 13.2090ms 12.4405ms 80.3826 Ops/s 79.5494 Ops/s $\color{#35bf28}+1.05\%$
test_cql_speed[True-backward] 19.1172ms 18.6450ms 53.6338 Ops/s 55.7209 Ops/s $\color{#d91a1a}-3.75\%$
test_cql_speed[reduce-overhead-None] 13.0930ms 12.5076ms 79.9513 Ops/s 79.1501 Ops/s $\color{#35bf28}+1.01\%$
test_cql_speed[reduce-overhead-backward] 18.9247ms 18.5484ms 53.9129 Ops/s 56.7718 Ops/s $\textbf{\color{#d91a1a}-5.04\%}$
test_a2c_speed[False-None] 5.9187ms 5.3989ms 185.2225 Ops/s 183.0265 Ops/s $\color{#35bf28}+1.20\%$
test_a2c_speed[False-backward] 12.5709ms 11.8776ms 84.1921 Ops/s 83.9919 Ops/s $\color{#35bf28}+0.24\%$
test_a2c_speed[True-None] 3.9802ms 3.6562ms 273.5062 Ops/s 261.9484 Ops/s $\color{#35bf28}+4.41\%$
test_a2c_speed[True-backward] 9.2657ms 8.6884ms 115.0964 Ops/s 110.0886 Ops/s $\color{#35bf28}+4.55\%$
test_a2c_speed[reduce-overhead-None] 4.0000ms 3.6984ms 270.3853 Ops/s 269.8199 Ops/s $\color{#35bf28}+0.21\%$
test_a2c_speed[reduce-overhead-backward] 9.1956ms 8.8275ms 113.2829 Ops/s 112.3178 Ops/s $\color{#35bf28}+0.86\%$
test_ppo_speed[False-None] 6.3563ms 5.8501ms 170.9385 Ops/s 171.4266 Ops/s $\color{#d91a1a}-0.28\%$
test_ppo_speed[False-backward] 12.7534ms 12.3290ms 81.1098 Ops/s 80.1744 Ops/s $\color{#35bf28}+1.17\%$
test_ppo_speed[True-None] 3.8408ms 3.5988ms 277.8698 Ops/s 271.7741 Ops/s $\color{#35bf28}+2.24\%$
test_ppo_speed[True-backward] 10.3015ms 8.7226ms 114.6448 Ops/s 109.7712 Ops/s $\color{#35bf28}+4.44\%$
test_ppo_speed[reduce-overhead-None] 3.9598ms 3.5312ms 283.1878 Ops/s 277.3105 Ops/s $\color{#35bf28}+2.12\%$
test_ppo_speed[reduce-overhead-backward] 9.1496ms 8.7891ms 113.7771 Ops/s 107.5039 Ops/s $\textbf{\color{#35bf28}+5.84\%}$
test_reinforce_speed[False-None] 5.0289ms 4.5768ms 218.4929 Ops/s 218.4546 Ops/s $\color{#35bf28}+0.02\%$
test_reinforce_speed[False-backward] 7.9906ms 7.4584ms 134.0771 Ops/s 136.0492 Ops/s $\color{#d91a1a}-1.45\%$
test_reinforce_speed[True-None] 3.3048ms 2.8966ms 345.2325 Ops/s 343.6968 Ops/s $\color{#35bf28}+0.45\%$
test_reinforce_speed[True-backward] 8.1513ms 7.7220ms 129.5006 Ops/s 128.2175 Ops/s $\color{#35bf28}+1.00\%$
test_reinforce_speed[reduce-overhead-None] 3.0797ms 2.8557ms 350.1712 Ops/s 339.0176 Ops/s $\color{#35bf28}+3.29\%$
test_reinforce_speed[reduce-overhead-backward] 8.2723ms 7.8958ms 126.6502 Ops/s 118.5778 Ops/s $\textbf{\color{#35bf28}+6.81\%}$
test_iql_speed[False-None] 25.8545ms 19.7719ms 50.5767 Ops/s 48.8293 Ops/s $\color{#35bf28}+3.58\%$
test_iql_speed[False-backward] 34.7667ms 30.3020ms 33.0011 Ops/s 32.5635 Ops/s $\color{#35bf28}+1.34\%$
test_iql_speed[True-None] 9.0253ms 8.5253ms 117.2976 Ops/s 114.7842 Ops/s $\color{#35bf28}+2.19\%$
test_iql_speed[True-backward] 17.7480ms 16.9341ms 59.0525 Ops/s 58.5868 Ops/s $\color{#35bf28}+0.79\%$
test_iql_speed[reduce-overhead-None] 8.9670ms 8.5803ms 116.5460 Ops/s 112.2421 Ops/s $\color{#35bf28}+3.83\%$
test_iql_speed[reduce-overhead-backward] 17.8262ms 17.2778ms 57.8778 Ops/s 58.3167 Ops/s $\color{#d91a1a}-0.75\%$
test_rb_sample[TensorDictReplayBuffer-ListStorage-RandomSampler-4000] 7.9289ms 5.9307ms 168.6133 Ops/s 168.3663 Ops/s $\color{#35bf28}+0.15\%$
test_rb_sample[TensorDictReplayBuffer-LazyMemmapStorage-RandomSampler-10000] 0.5386ms 0.3226ms 3.1003 KOps/s 2.9743 KOps/s $\color{#35bf28}+4.24\%$
test_rb_sample[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-10000] 0.5988ms 0.3037ms 3.2926 KOps/s 3.0668 KOps/s $\textbf{\color{#35bf28}+7.36\%}$
test_rb_sample[TensorDictReplayBuffer-ListStorage-SamplerWithoutReplacement-4000] 6.1358ms 5.7582ms 173.6648 Ops/s 176.2683 Ops/s $\color{#d91a1a}-1.48\%$
test_rb_sample[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-10000] 0.6900ms 0.2749ms 3.6374 KOps/s 2.9386 KOps/s $\textbf{\color{#35bf28}+23.78\%}$
test_rb_sample[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-10000] 0.4979ms 0.2573ms 3.8869 KOps/s 3.1391 KOps/s $\textbf{\color{#35bf28}+23.82\%}$
test_rb_sample[TensorDictReplayBuffer-LazyMemmapStorage-sampler6-10000] 1.4696ms 1.2035ms 830.8789 Ops/s 747.9604 Ops/s $\textbf{\color{#35bf28}+11.09\%}$
test_rb_sample[TensorDictReplayBuffer-LazyTensorStorage-sampler7-10000] 1.5521ms 1.1529ms 867.3462 Ops/s 811.5878 Ops/s $\textbf{\color{#35bf28}+6.87\%}$
test_rb_sample[TensorDictPrioritizedReplayBuffer-ListStorage-None-4000] 9.4653ms 5.8952ms 169.6298 Ops/s 169.7005 Ops/s $\color{#d91a1a}-0.04\%$
test_rb_sample[TensorDictPrioritizedReplayBuffer-LazyMemmapStorage-None-10000] 0.9638ms 0.4368ms 2.2894 KOps/s 2.3245 KOps/s $\color{#d91a1a}-1.51\%$
test_rb_sample[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-10000] 0.8571ms 0.4135ms 2.4181 KOps/s 2.4007 KOps/s $\color{#35bf28}+0.73\%$
test_rb_iterate[TensorDictReplayBuffer-ListStorage-RandomSampler-4000] 5.8779ms 5.7032ms 175.3413 Ops/s 173.5338 Ops/s $\color{#35bf28}+1.04\%$
test_rb_iterate[TensorDictReplayBuffer-LazyMemmapStorage-RandomSampler-10000] 0.6503ms 0.2812ms 3.5562 KOps/s 3.6023 KOps/s $\color{#d91a1a}-1.28\%$
test_rb_iterate[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-10000] 0.5813ms 0.3394ms 2.9461 KOps/s 3.8354 KOps/s $\textbf{\color{#d91a1a}-23.19\%}$
test_rb_iterate[TensorDictReplayBuffer-ListStorage-SamplerWithoutReplacement-4000] 5.8923ms 5.5181ms 181.2228 Ops/s 176.2900 Ops/s $\color{#35bf28}+2.80\%$
test_rb_iterate[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-10000] 2.1801ms 0.3646ms 2.7429 KOps/s 3.6081 KOps/s $\textbf{\color{#d91a1a}-23.98\%}$
test_rb_iterate[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-10000] 0.5926ms 0.3517ms 2.8430 KOps/s 3.8199 KOps/s $\textbf{\color{#d91a1a}-25.58\%}$
test_rb_iterate[TensorDictPrioritizedReplayBuffer-ListStorage-None-4000] 5.8121ms 5.7104ms 175.1190 Ops/s 171.3061 Ops/s $\color{#35bf28}+2.23\%$
test_rb_iterate[TensorDictPrioritizedReplayBuffer-LazyMemmapStorage-None-10000] 1.5107ms 0.5020ms 1.9919 KOps/s 2.1350 KOps/s $\textbf{\color{#d91a1a}-6.70\%}$
test_rb_iterate[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-10000] 0.6691ms 0.4880ms 2.0493 KOps/s 2.2149 KOps/s $\textbf{\color{#d91a1a}-7.48\%}$
test_rb_populate[TensorDictReplayBuffer-ListStorage-RandomSampler-400] 6.4905ms 5.0506ms 197.9973 Ops/s 196.7521 Ops/s $\color{#35bf28}+0.63\%$
test_rb_populate[TensorDictReplayBuffer-LazyMemmapStorage-RandomSampler-400] 10.2505ms 2.3255ms 430.0186 Ops/s 488.4179 Ops/s $\textbf{\color{#d91a1a}-11.96\%}$
test_rb_populate[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-400] 6.2921ms 1.1774ms 849.3045 Ops/s 844.3685 Ops/s $\color{#35bf28}+0.58\%$
test_rb_populate[TensorDictReplayBuffer-ListStorage-SamplerWithoutReplacement-400] 6.5156ms 4.9863ms 200.5507 Ops/s 197.5461 Ops/s $\color{#35bf28}+1.52\%$
test_rb_populate[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-400] 10.5600ms 2.3847ms 419.3316 Ops/s 476.5703 Ops/s $\textbf{\color{#d91a1a}-12.01\%}$
test_rb_populate[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-400] 3.3274ms 1.1159ms 896.1700 Ops/s 861.1014 Ops/s $\color{#35bf28}+4.07\%$
test_rb_populate[TensorDictPrioritizedReplayBuffer-ListStorage-None-400] 0.6418s 17.8175ms 56.1246 Ops/s 52.4748 Ops/s $\textbf{\color{#35bf28}+6.96\%}$
test_rb_populate[TensorDictPrioritizedReplayBuffer-LazyMemmapStorage-None-400] 8.3784ms 2.1270ms 470.1452 Ops/s 535.1200 Ops/s $\textbf{\color{#d91a1a}-12.14\%}$
test_rb_populate[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-400] 8.6675ms 1.3552ms 737.8930 Ops/s 720.0894 Ops/s $\color{#35bf28}+2.47\%$
test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-10000-10000-100-True] 36.5410ms 33.1378ms 30.1770 Ops/s 29.9517 Ops/s $\color{#35bf28}+0.75\%$
test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-10000-10000-100-False] 19.8079ms 17.5237ms 57.0656 Ops/s 57.8582 Ops/s $\color{#d91a1a}-1.37\%$
test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-100000-10000-100-True] 36.2550ms 34.1553ms 29.2781 Ops/s 28.2468 Ops/s $\color{#35bf28}+3.65\%$
test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-100000-10000-100-False] 19.0792ms 17.5392ms 57.0153 Ops/s 55.1455 Ops/s $\color{#35bf28}+3.39\%$
test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-1000000-10000-100-True] 40.7866ms 36.0590ms 27.7323 Ops/s 27.2750 Ops/s $\color{#35bf28}+1.68\%$
test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-1000000-10000-100-False] 20.7391ms 19.0855ms 52.3958 Ops/s 51.6647 Ops/s $\color{#35bf28}+1.42\%$
@vmoens vmoens added the CI Has to do with CI setup (e.g. wheels & builds, tests...) label Dec 31, 2025
[ghstack-poisoned]
vmoens added a commit that referenced this pull request Dec 31, 2025
ghstack-source-id: 7c92a6d
Pull-Request: #3286
[ghstack-poisoned]
vmoens added a commit that referenced this pull request Dec 31, 2025
ghstack-source-id: d5fae60
Pull-Request: #3286
[ghstack-poisoned]
@vmoens vmoens mentioned this pull request Jan 1, 2026
[ghstack-poisoned]
vmoens added a commit that referenced this pull request Jan 1, 2026
ghstack-source-id: 34cd3ab
Pull-Request: #3286

amend

ghstack-source-id: 34cd3ab
Pull-Request: #3287
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

CI Has to do with CI setup (e.g. wheels & builds, tests...) CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed.

2 participants