Skip to content

Conversation

@vmoens
Copy link
Collaborator

@vmoens vmoens commented Jan 1, 2026

[ghstack-poisoned]
@pytorch-bot
Copy link

pytorch-bot bot commented Jan 1, 2026

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/rl/3287

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:

❌ 2 New Failures, 7 Pending, 9 Unrelated Failures

As of commit 6c18731 with merge base 7866d11 (image):

NEW FAILURES - The following jobs have failed:

FLAKY - The following jobs failed but were likely due to flakiness present on trunk:

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 1, 2026
vmoens added a commit that referenced this pull request Jan 1, 2026
ghstack-source-id: 34cd3ab
Pull-Request: #3287
vmoens added a commit that referenced this pull request Jan 1, 2026
ghstack-source-id: 34cd3ab
Pull-Request: #3287
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
@vmoens vmoens closed this Jan 1, 2026
@vmoens vmoens deleted the gh/vmoens/173/head branch January 1, 2026 07:27
@github-actions
Copy link

github-actions bot commented Jan 1, 2026

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

Total Benchmarks: 164. Improved: $\large\color{#35bf28}16$. 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] 81.6598μs 80.8505μs 12.3685 KOps/s 12.4930 KOps/s $\color{#d91a1a}-1.00\%$
test_tensor_to_bytestream_speed[torch.save] 0.1386ms 0.1383ms 7.2329 KOps/s 7.2716 KOps/s $\color{#d91a1a}-0.53\%$
test_tensor_to_bytestream_speed[untyped_storage] 0.1139s 0.1137s 8.7940 Ops/s 8.7932 Ops/s $+0.01\%$
test_tensor_to_bytestream_speed[numpy] 2.7122μs 2.7050μs 369.6884 KOps/s 364.5608 KOps/s $\color{#35bf28}+1.41\%$
test_tensor_to_bytestream_speed[safetensors] 37.0197μs 36.8187μs 27.1601 KOps/s 25.0845 KOps/s $\textbf{\color{#35bf28}+8.27\%}$
test_simple 0.5415s 0.5364s 1.8644 Ops/s 1.7648 Ops/s $\textbf{\color{#35bf28}+5.64\%}$
test_transformed 1.1020s 1.0999s 0.9091 Ops/s 0.8849 Ops/s $\color{#35bf28}+2.74\%$
test_serial 1.6179s 1.6164s 0.6187 Ops/s 0.6051 Ops/s $\color{#35bf28}+2.25\%$
test_parallel 1.0945s 1.0651s 0.9389 Ops/s 0.8612 Ops/s $\textbf{\color{#35bf28}+9.02\%}$
test_step_mdp_speed[True-True-True-True-True] 0.4856ms 43.6720μs 22.8980 KOps/s 23.2246 KOps/s $\color{#d91a1a}-1.41\%$
test_step_mdp_speed[True-True-True-True-False] 51.6710μs 24.3302μs 41.1012 KOps/s 40.9041 KOps/s $\color{#35bf28}+0.48\%$
test_step_mdp_speed[True-True-True-False-True] 48.6410μs 24.1452μs 41.4161 KOps/s 39.1249 KOps/s $\textbf{\color{#35bf28}+5.86\%}$
test_step_mdp_speed[True-True-True-False-False] 37.8100μs 13.3683μs 74.8039 KOps/s 71.9452 KOps/s $\color{#35bf28}+3.97\%$
test_step_mdp_speed[True-True-False-True-True] 89.5320μs 47.0039μs 21.2748 KOps/s 21.3161 KOps/s $\color{#d91a1a}-0.19\%$
test_step_mdp_speed[True-True-False-True-False] 50.9910μs 26.9089μs 37.1624 KOps/s 37.2540 KOps/s $\color{#d91a1a}-0.25\%$
test_step_mdp_speed[True-True-False-False-True] 65.4510μs 26.9970μs 37.0412 KOps/s 36.5823 KOps/s $\color{#35bf28}+1.25\%$
test_step_mdp_speed[True-True-False-False-False] 46.7310μs 16.2274μs 61.6243 KOps/s 60.1265 KOps/s $\color{#35bf28}+2.49\%$
test_step_mdp_speed[True-False-True-True-True] 87.5320μs 48.9481μs 20.4298 KOps/s 19.8614 KOps/s $\color{#35bf28}+2.86\%$
test_step_mdp_speed[True-False-True-True-False] 63.9110μs 29.4055μs 34.0072 KOps/s 33.5501 KOps/s $\color{#35bf28}+1.36\%$
test_step_mdp_speed[True-False-True-False-True] 52.4210μs 26.6402μs 37.5373 KOps/s 36.4977 KOps/s $\color{#35bf28}+2.85\%$
test_step_mdp_speed[True-False-True-False-False] 37.1710μs 15.9728μs 62.6062 KOps/s 61.1223 KOps/s $\color{#35bf28}+2.43\%$
test_step_mdp_speed[True-False-False-True-True] 87.9820μs 51.0550μs 19.5867 KOps/s 19.2991 KOps/s $\color{#35bf28}+1.49\%$
test_step_mdp_speed[True-False-False-True-False] 58.0310μs 31.8512μs 31.3960 KOps/s 30.8463 KOps/s $\color{#35bf28}+1.78\%$
test_step_mdp_speed[True-False-False-False-True] 55.7810μs 28.6537μs 34.8995 KOps/s 34.0808 KOps/s $\color{#35bf28}+2.40\%$
test_step_mdp_speed[True-False-False-False-False] 49.7610μs 18.7719μs 53.2711 KOps/s 51.9948 KOps/s $\color{#35bf28}+2.45\%$
test_step_mdp_speed[False-True-True-True-True] 77.2220μs 49.1186μs 20.3589 KOps/s 20.1372 KOps/s $\color{#35bf28}+1.10\%$
test_step_mdp_speed[False-True-True-True-False] 54.9520μs 29.9893μs 33.3452 KOps/s 34.2291 KOps/s $\color{#d91a1a}-2.58\%$
test_step_mdp_speed[False-True-True-False-True] 2.3121ms 31.1274μs 32.1260 KOps/s 31.8885 KOps/s $\color{#35bf28}+0.74\%$
test_step_mdp_speed[False-True-True-False-False] 42.6010μs 17.5406μs 57.0107 KOps/s 56.6052 KOps/s $\color{#35bf28}+0.72\%$
test_step_mdp_speed[False-True-False-True-True] 82.4020μs 50.8405μs 19.6694 KOps/s 19.3495 KOps/s $\color{#35bf28}+1.65\%$
test_step_mdp_speed[False-True-False-True-False] 60.3710μs 32.1890μs 31.0665 KOps/s 31.6501 KOps/s $\color{#d91a1a}-1.84\%$
test_step_mdp_speed[False-True-False-False-True] 64.7120μs 33.2271μs 30.0959 KOps/s 29.8240 KOps/s $\color{#35bf28}+0.91\%$
test_step_mdp_speed[False-True-False-False-False] 56.4410μs 20.3024μs 49.2552 KOps/s 49.7069 KOps/s $\color{#d91a1a}-0.91\%$
test_step_mdp_speed[False-False-True-True-True] 87.7120μs 53.6208μs 18.6495 KOps/s 18.5577 KOps/s $\color{#35bf28}+0.49\%$
test_step_mdp_speed[False-False-True-True-False] 66.0310μs 34.5617μs 28.9337 KOps/s 28.7028 KOps/s $\color{#35bf28}+0.80\%$
test_step_mdp_speed[False-False-True-False-True] 65.1620μs 32.8888μs 30.4055 KOps/s 30.6385 KOps/s $\color{#d91a1a}-0.76\%$
test_step_mdp_speed[False-False-True-False-False] 60.9620μs 19.7674μs 50.5883 KOps/s 50.7126 KOps/s $\color{#d91a1a}-0.24\%$
test_step_mdp_speed[False-False-False-True-True] 88.7510μs 55.1117μs 18.1450 KOps/s 18.1918 KOps/s $\color{#d91a1a}-0.26\%$
test_step_mdp_speed[False-False-False-True-False] 76.2420μs 36.9766μs 27.0442 KOps/s 27.2317 KOps/s $\color{#d91a1a}-0.69\%$
test_step_mdp_speed[False-False-False-False-True] 65.4010μs 35.2739μs 28.3495 KOps/s 28.8007 KOps/s $\color{#d91a1a}-1.57\%$
test_step_mdp_speed[False-False-False-False-False] 51.9910μs 22.7731μs 43.9115 KOps/s 44.3360 KOps/s $\color{#d91a1a}-0.96\%$
test_non_tensor_env_rollout_speed[1000-single-True] 0.8653s 0.7648s 1.3076 Ops/s 1.3108 Ops/s $\color{#d91a1a}-0.24\%$
test_non_tensor_env_rollout_speed[1000-single-False] 0.7254s 0.6267s 1.5958 Ops/s 1.5800 Ops/s $\color{#35bf28}+1.00\%$
test_non_tensor_env_rollout_speed[1000-serial-no-buffers-True] 1.7414s 1.6610s 0.6020 Ops/s 0.5946 Ops/s $\color{#35bf28}+1.25\%$
test_non_tensor_env_rollout_speed[1000-serial-no-buffers-False] 1.5198s 1.4464s 0.6914 Ops/s 0.6873 Ops/s $\color{#35bf28}+0.59\%$
test_non_tensor_env_rollout_speed[1000-serial-buffers-True] 1.9990s 1.9145s 0.5223 Ops/s 0.5251 Ops/s $\color{#d91a1a}-0.53\%$
test_non_tensor_env_rollout_speed[1000-serial-buffers-False] 1.7832s 1.6955s 0.5898 Ops/s 0.5897 Ops/s $\color{#35bf28}+0.02\%$
test_non_tensor_env_rollout_speed[1000-parallel-no-buffers-True] 4.6847s 4.6290s 0.2160 Ops/s 0.2167 Ops/s $\color{#d91a1a}-0.30\%$
test_non_tensor_env_rollout_speed[1000-parallel-no-buffers-False] 4.4382s 4.3579s 0.2295 Ops/s 0.2278 Ops/s $\color{#35bf28}+0.71\%$
test_non_tensor_env_rollout_speed[1000-parallel-buffers-True] 2.1873s 1.9757s 0.5061 Ops/s 0.5184 Ops/s $\color{#d91a1a}-2.37\%$
test_non_tensor_env_rollout_speed[1000-parallel-buffers-False] 1.7253s 1.6433s 0.6085 Ops/s 0.5958 Ops/s $\color{#35bf28}+2.13\%$
test_values[generalized_advantage_estimate-True-True] 10.5042ms 9.9007ms 101.0029 Ops/s 100.9893 Ops/s $\color{#35bf28}+0.01\%$
test_values[vec_generalized_advantage_estimate-True-True] 20.3579ms 17.6097ms 56.7869 Ops/s 57.4081 Ops/s $\color{#d91a1a}-1.08\%$
test_values[td0_return_estimate-False-False] 0.1943ms 0.1262ms 7.9252 KOps/s 8.1797 KOps/s $\color{#d91a1a}-3.11\%$
test_values[td1_return_estimate-False-False] 26.6105ms 26.3415ms 37.9630 Ops/s 37.7141 Ops/s $\color{#35bf28}+0.66\%$
test_values[vec_td1_return_estimate-False-False] 18.0053ms 17.5518ms 56.9742 Ops/s 57.0675 Ops/s $\color{#d91a1a}-0.16\%$
test_values[td_lambda_return_estimate-True-False] 40.4684ms 39.1658ms 25.5325 Ops/s 25.3188 Ops/s $\color{#35bf28}+0.84\%$
test_values[vec_td_lambda_return_estimate-True-False] 18.1535ms 17.5927ms 56.8416 Ops/s 56.8366 Ops/s $+0.01\%$
test_gae_speed[generalized_advantage_estimate-False-1-512] 9.0369ms 8.8557ms 112.9220 Ops/s 113.3020 Ops/s $\color{#d91a1a}-0.34\%$
test_gae_speed[vec_generalized_advantage_estimate-True-1-512] 1.7123ms 1.4624ms 683.8154 Ops/s 698.7539 Ops/s $\color{#d91a1a}-2.14\%$
test_gae_speed[vec_generalized_advantage_estimate-False-1-512] 0.5537ms 0.4111ms 2.4326 KOps/s 2.4336 KOps/s $\color{#d91a1a}-0.04\%$
test_gae_speed[vec_generalized_advantage_estimate-True-32-512] 35.5861ms 33.9089ms 29.4908 Ops/s 29.1097 Ops/s $\color{#35bf28}+1.31\%$
test_gae_speed[vec_generalized_advantage_estimate-False-32-512] 2.1033ms 1.7017ms 587.6451 Ops/s 587.1360 Ops/s $\color{#35bf28}+0.09\%$
test_dqn_speed[False-None] 1.7324ms 1.3838ms 722.6255 Ops/s 728.7783 Ops/s $\color{#d91a1a}-0.84\%$
test_dqn_speed[False-backward] 2.0306ms 1.8992ms 526.5407 Ops/s 532.6016 Ops/s $\color{#d91a1a}-1.14\%$
test_dqn_speed[True-None] 0.9130ms 0.5141ms 1.9453 KOps/s 1.9461 KOps/s $\color{#d91a1a}-0.04\%$
test_dqn_speed[True-backward] 1.0164ms 0.9599ms 1.0417 KOps/s 864.5885 Ops/s $\textbf{\color{#35bf28}+20.49\%}$
test_dqn_speed[reduce-overhead-None] 0.7654ms 0.5093ms 1.9634 KOps/s 1.8962 KOps/s $\color{#35bf28}+3.54\%$
test_dqn_speed[reduce-overhead-backward] 1.0023ms 0.9489ms 1.0539 KOps/s 876.3949 Ops/s $\textbf{\color{#35bf28}+20.25\%}$
test_ddpg_speed[False-None] 3.1752ms 2.8106ms 355.7995 Ops/s 343.3714 Ops/s $\color{#35bf28}+3.62\%$
test_ddpg_speed[False-backward] 4.1196ms 4.0138ms 249.1431 Ops/s 250.8739 Ops/s $\color{#d91a1a}-0.69\%$
test_ddpg_speed[True-None] 1.7106ms 1.3546ms 738.2397 Ops/s 721.4250 Ops/s $\color{#35bf28}+2.33\%$
test_ddpg_speed[True-backward] 2.3731ms 2.3115ms 432.6140 Ops/s 432.9649 Ops/s $\color{#d91a1a}-0.08\%$
test_ddpg_speed[reduce-overhead-None] 1.4758ms 1.3478ms 741.9450 Ops/s 733.5357 Ops/s $\color{#35bf28}+1.15\%$
test_ddpg_speed[reduce-overhead-backward] 2.3512ms 2.3025ms 434.3142 Ops/s 437.3298 Ops/s $\color{#d91a1a}-0.69\%$
test_sac_speed[False-None] 8.2505ms 7.7788ms 128.5543 Ops/s 128.1179 Ops/s $\color{#35bf28}+0.34\%$
test_sac_speed[False-backward] 11.8530ms 11.0104ms 90.8236 Ops/s 88.6029 Ops/s $\color{#35bf28}+2.51\%$
test_sac_speed[True-None] 2.5037ms 2.0729ms 482.4133 Ops/s 462.0676 Ops/s $\color{#35bf28}+4.40\%$
test_sac_speed[True-backward] 4.0355ms 3.9115ms 255.6581 Ops/s 231.3529 Ops/s $\textbf{\color{#35bf28}+10.51\%}$
test_sac_speed[reduce-overhead-None] 2.4062ms 2.0562ms 486.3454 Ops/s 477.8505 Ops/s $\color{#35bf28}+1.78\%$
test_sac_speed[reduce-overhead-backward] 4.0084ms 3.8823ms 257.5794 Ops/s 245.6610 Ops/s $\color{#35bf28}+4.85\%$
test_redq_speed[False-None] 15.6640ms 10.9147ms 91.6194 Ops/s 94.7259 Ops/s $\color{#d91a1a}-3.28\%$
test_redq_speed[False-backward] 17.7697ms 17.1937ms 58.1608 Ops/s 56.5939 Ops/s $\color{#35bf28}+2.77\%$
test_redq_speed[True-None] 4.6983ms 4.3004ms 232.5380 Ops/s 233.1447 Ops/s $\color{#d91a1a}-0.26\%$
test_redq_speed[True-backward] 9.7167ms 9.4605ms 105.7031 Ops/s 108.3883 Ops/s $\color{#d91a1a}-2.48\%$
test_redq_speed[reduce-overhead-None] 4.4349ms 4.2263ms 236.6122 Ops/s 225.4446 Ops/s $\color{#35bf28}+4.95\%$
test_redq_speed[reduce-overhead-backward] 9.8611ms 9.5938ms 104.2342 Ops/s 101.9173 Ops/s $\color{#35bf28}+2.27\%$
test_redq_deprec_speed[False-None] 11.2014ms 10.7836ms 92.7331 Ops/s 90.5259 Ops/s $\color{#35bf28}+2.44\%$
test_redq_deprec_speed[False-backward] 16.1005ms 15.4983ms 64.5234 Ops/s 64.9864 Ops/s $\color{#d91a1a}-0.71\%$
test_redq_deprec_speed[True-None] 3.8640ms 3.4996ms 285.7459 Ops/s 273.9947 Ops/s $\color{#35bf28}+4.29\%$
test_redq_deprec_speed[True-backward] 7.7539ms 7.3042ms 136.9076 Ops/s 133.6517 Ops/s $\color{#35bf28}+2.44\%$
test_redq_deprec_speed[reduce-overhead-None] 3.9304ms 3.5117ms 284.7608 Ops/s 285.3779 Ops/s $\color{#d91a1a}-0.22\%$
test_redq_deprec_speed[reduce-overhead-backward] 7.4998ms 7.3303ms 136.4193 Ops/s 136.9209 Ops/s $\color{#d91a1a}-0.37\%$
test_td3_speed[False-None] 7.9676ms 7.8127ms 127.9970 Ops/s 124.1956 Ops/s $\color{#35bf28}+3.06\%$
test_td3_speed[False-backward] 11.0861ms 10.6262ms 94.1071 Ops/s 91.4621 Ops/s $\color{#35bf28}+2.89\%$
test_td3_speed[True-None] 1.8217ms 1.7603ms 568.0694 Ops/s 560.7088 Ops/s $\color{#35bf28}+1.31\%$
test_td3_speed[True-backward] 3.6555ms 3.5272ms 283.5074 Ops/s 229.1210 Ops/s $\textbf{\color{#35bf28}+23.74\%}$
test_td3_speed[reduce-overhead-None] 1.7733ms 1.7321ms 577.3248 Ops/s 574.3903 Ops/s $\color{#35bf28}+0.51\%$
test_td3_speed[reduce-overhead-backward] 3.8722ms 3.5491ms 281.7630 Ops/s 242.1248 Ops/s $\textbf{\color{#35bf28}+16.37\%}$
test_cql_speed[False-None] 29.2247ms 25.4715ms 39.2596 Ops/s 39.0366 Ops/s $\color{#35bf28}+0.57\%$
test_cql_speed[False-backward] 37.8846ms 34.5430ms 28.9495 Ops/s 28.7927 Ops/s $\color{#35bf28}+0.54\%$
test_cql_speed[True-None] 12.6444ms 12.0866ms 82.7361 Ops/s 82.0903 Ops/s $\color{#35bf28}+0.79\%$
test_cql_speed[True-backward] 18.6733ms 18.1163ms 55.1989 Ops/s 56.2284 Ops/s $\color{#d91a1a}-1.83\%$
test_cql_speed[reduce-overhead-None] 12.6186ms 12.2577ms 81.5814 Ops/s 82.9112 Ops/s $\color{#d91a1a}-1.60\%$
test_cql_speed[reduce-overhead-backward] 20.1978ms 18.2087ms 54.9189 Ops/s 56.6254 Ops/s $\color{#d91a1a}-3.01\%$
test_a2c_speed[False-None] 5.7910ms 5.3720ms 186.1505 Ops/s 190.2691 Ops/s $\color{#d91a1a}-2.16\%$
test_a2c_speed[False-backward] 11.9268ms 11.6232ms 86.0351 Ops/s 87.3891 Ops/s $\color{#d91a1a}-1.55\%$
test_a2c_speed[True-None] 3.8227ms 3.6351ms 275.0977 Ops/s 273.4953 Ops/s $\color{#35bf28}+0.59\%$
test_a2c_speed[True-backward] 8.7574ms 8.4366ms 118.5317 Ops/s 108.2664 Ops/s $\textbf{\color{#35bf28}+9.48\%}$
test_a2c_speed[reduce-overhead-None] 4.0494ms 3.6648ms 272.8632 Ops/s 278.0159 Ops/s $\color{#d91a1a}-1.85\%$
test_a2c_speed[reduce-overhead-backward] 8.7768ms 8.5808ms 116.5388 Ops/s 116.2402 Ops/s $\color{#35bf28}+0.26\%$
test_ppo_speed[False-None] 6.1262ms 5.7704ms 173.2977 Ops/s 172.2851 Ops/s $\color{#35bf28}+0.59\%$
test_ppo_speed[False-backward] 12.6200ms 12.2250ms 81.7996 Ops/s 81.3448 Ops/s $\color{#35bf28}+0.56\%$
test_ppo_speed[True-None] 3.7053ms 3.5547ms 281.3189 Ops/s 277.7863 Ops/s $\color{#35bf28}+1.27\%$
test_ppo_speed[True-backward] 8.7089ms 8.2474ms 121.2507 Ops/s 119.8606 Ops/s $\color{#35bf28}+1.16\%$
test_ppo_speed[reduce-overhead-None] 3.7430ms 3.5476ms 281.8789 Ops/s 276.8824 Ops/s $\color{#35bf28}+1.80\%$
test_ppo_speed[reduce-overhead-backward] 8.9872ms 8.5475ms 116.9938 Ops/s 113.3598 Ops/s $\color{#35bf28}+3.21\%$
test_reinforce_speed[False-None] 4.6861ms 4.4903ms 222.7026 Ops/s 215.5389 Ops/s $\color{#35bf28}+3.32\%$
test_reinforce_speed[False-backward] 7.5553ms 7.3069ms 136.8572 Ops/s 132.6961 Ops/s $\color{#35bf28}+3.14\%$
test_reinforce_speed[True-None] 3.2252ms 2.7984ms 357.3455 Ops/s 345.0769 Ops/s $\color{#35bf28}+3.56\%$
test_reinforce_speed[True-backward] 7.8068ms 7.5808ms 131.9118 Ops/s 124.7869 Ops/s $\textbf{\color{#35bf28}+5.71\%}$
test_reinforce_speed[reduce-overhead-None] 3.2422ms 2.8087ms 356.0398 Ops/s 352.3921 Ops/s $\color{#35bf28}+1.04\%$
test_reinforce_speed[reduce-overhead-backward] 8.1029ms 7.6738ms 130.3141 Ops/s 127.8754 Ops/s $\color{#35bf28}+1.91\%$
test_iql_speed[False-None] 20.3097ms 19.5691ms 51.1009 Ops/s 49.1448 Ops/s $\color{#35bf28}+3.98\%$
test_iql_speed[False-backward] 31.3003ms 29.8137ms 33.5416 Ops/s 33.3571 Ops/s $\color{#35bf28}+0.55\%$
test_iql_speed[True-None] 8.7818ms 8.3108ms 120.3251 Ops/s 117.8951 Ops/s $\color{#35bf28}+2.06\%$
test_iql_speed[True-backward] 16.9234ms 16.3680ms 61.0948 Ops/s 61.1918 Ops/s $\color{#d91a1a}-0.16\%$
test_iql_speed[reduce-overhead-None] 8.7059ms 8.3472ms 119.8005 Ops/s 117.8935 Ops/s $\color{#35bf28}+1.62\%$
test_iql_speed[reduce-overhead-backward] 17.2854ms 16.6940ms 59.9018 Ops/s 60.0262 Ops/s $\color{#d91a1a}-0.21\%$
test_rb_sample[TensorDictReplayBuffer-ListStorage-RandomSampler-4000] 7.7605ms 5.9344ms 168.5096 Ops/s 168.2613 Ops/s $\color{#35bf28}+0.15\%$
test_rb_sample[TensorDictReplayBuffer-LazyMemmapStorage-RandomSampler-10000] 0.5617ms 0.3321ms 3.0109 KOps/s 3.2379 KOps/s $\textbf{\color{#d91a1a}-7.01\%}$
test_rb_sample[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-10000] 0.7796ms 0.2858ms 3.4992 KOps/s 3.4147 KOps/s $\color{#35bf28}+2.47\%$
test_rb_sample[TensorDictReplayBuffer-ListStorage-SamplerWithoutReplacement-4000] 5.9479ms 5.6369ms 177.4036 Ops/s 174.4014 Ops/s $\color{#35bf28}+1.72\%$
test_rb_sample[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-10000] 0.7556ms 0.3276ms 3.0526 KOps/s 3.4993 KOps/s $\textbf{\color{#d91a1a}-12.77\%}$
test_rb_sample[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-10000] 0.6909ms 0.2594ms 3.8554 KOps/s 3.6504 KOps/s $\textbf{\color{#35bf28}+5.62\%}$
test_rb_sample[TensorDictReplayBuffer-LazyMemmapStorage-sampler6-10000] 1.6533ms 1.3765ms 726.4785 Ops/s 805.7921 Ops/s $\textbf{\color{#d91a1a}-9.84\%}$
test_rb_sample[TensorDictReplayBuffer-LazyTensorStorage-sampler7-10000] 1.6845ms 1.3042ms 766.7409 Ops/s 855.4914 Ops/s $\textbf{\color{#d91a1a}-10.37\%}$
test_rb_sample[TensorDictPrioritizedReplayBuffer-ListStorage-None-4000] 9.6055ms 5.9597ms 167.7943 Ops/s 169.8205 Ops/s $\color{#d91a1a}-1.19\%$
test_rb_sample[TensorDictPrioritizedReplayBuffer-LazyMemmapStorage-None-10000] 1.2042ms 0.4620ms 2.1645 KOps/s 2.3103 KOps/s $\textbf{\color{#d91a1a}-6.31\%}$
test_rb_sample[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-10000] 0.8148ms 0.4568ms 2.1894 KOps/s 2.3300 KOps/s $\textbf{\color{#d91a1a}-6.04\%}$
test_rb_iterate[TensorDictReplayBuffer-ListStorage-RandomSampler-4000] 5.8009ms 5.6390ms 177.3379 Ops/s 174.8414 Ops/s $\color{#35bf28}+1.43\%$
test_rb_iterate[TensorDictReplayBuffer-LazyMemmapStorage-RandomSampler-10000] 0.6814ms 0.3622ms 2.7610 KOps/s 3.1857 KOps/s $\textbf{\color{#d91a1a}-13.33\%}$
test_rb_iterate[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-10000] 0.4723ms 0.2644ms 3.7828 KOps/s 3.8401 KOps/s $\color{#d91a1a}-1.49\%$
test_rb_iterate[TensorDictReplayBuffer-ListStorage-SamplerWithoutReplacement-4000] 5.8153ms 5.5520ms 180.1157 Ops/s 176.1567 Ops/s $\color{#35bf28}+2.25\%$
test_rb_iterate[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-10000] 0.6809ms 0.2920ms 3.4242 KOps/s 3.6862 KOps/s $\textbf{\color{#d91a1a}-7.11\%}$
test_rb_iterate[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-10000] 0.5152ms 0.2742ms 3.6466 KOps/s 2.6801 KOps/s $\textbf{\color{#35bf28}+36.06\%}$
test_rb_iterate[TensorDictPrioritizedReplayBuffer-ListStorage-None-4000] 8.6409ms 5.6926ms 175.6678 Ops/s 170.2881 Ops/s $\color{#35bf28}+3.16\%$
test_rb_iterate[TensorDictPrioritizedReplayBuffer-LazyMemmapStorage-None-10000] 1.4095ms 0.4611ms 2.1686 KOps/s 1.8974 KOps/s $\textbf{\color{#35bf28}+14.29\%}$
test_rb_iterate[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-10000] 0.7381ms 0.4470ms 2.2372 KOps/s 2.0768 KOps/s $\textbf{\color{#35bf28}+7.73\%}$
test_rb_populate[TensorDictReplayBuffer-ListStorage-RandomSampler-400] 0.7919s 20.6179ms 48.5016 Ops/s 194.9508 Ops/s $\textbf{\color{#d91a1a}-75.12\%}$
test_rb_populate[TensorDictReplayBuffer-LazyMemmapStorage-RandomSampler-400] 9.3434ms 2.0802ms 480.7212 Ops/s 492.5144 Ops/s $\color{#d91a1a}-2.39\%$
test_rb_populate[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-400] 8.0359ms 1.2224ms 818.0679 Ops/s 860.2643 Ops/s $\color{#d91a1a}-4.91\%$
test_rb_populate[TensorDictReplayBuffer-ListStorage-SamplerWithoutReplacement-400] 6.9791ms 4.9711ms 201.1640 Ops/s 196.5054 Ops/s $\color{#35bf28}+2.37\%$
test_rb_populate[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-400] 8.9776ms 2.0526ms 487.1980 Ops/s 485.9723 Ops/s $\color{#35bf28}+0.25\%$
test_rb_populate[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-400] 7.1194ms 1.1992ms 833.9181 Ops/s 849.8006 Ops/s $\color{#d91a1a}-1.87\%$
test_rb_populate[TensorDictPrioritizedReplayBuffer-ListStorage-None-400] 6.7647ms 5.1474ms 194.2728 Ops/s 52.9081 Ops/s $\textbf{\color{#35bf28}+267.19\%}$
test_rb_populate[TensorDictPrioritizedReplayBuffer-LazyMemmapStorage-None-400] 0.5766s 13.7391ms 72.7848 Ops/s 515.5104 Ops/s $\textbf{\color{#d91a1a}-85.88\%}$
test_rb_populate[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-400] 2.3290ms 1.1351ms 880.9691 Ops/s 1.0061 KOps/s $\textbf{\color{#d91a1a}-12.44\%}$
test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-10000-10000-100-True] 35.9763ms 33.2069ms 30.1142 Ops/s 29.0778 Ops/s $\color{#35bf28}+3.56\%$
test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-10000-10000-100-False] 19.3202ms 17.6550ms 56.6412 Ops/s 57.0756 Ops/s $\color{#d91a1a}-0.76\%$
test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-100000-10000-100-True] 39.5608ms 34.4778ms 29.0042 Ops/s 28.4706 Ops/s $\color{#35bf28}+1.87\%$
test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-100000-10000-100-False] 19.5032ms 17.8462ms 56.0344 Ops/s 56.2281 Ops/s $\color{#d91a1a}-0.34\%$
test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-1000000-10000-100-True] 40.8709ms 36.2357ms 27.5971 Ops/s 26.8966 Ops/s $\color{#35bf28}+2.60\%$
test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-1000000-10000-100-False] 20.9603ms 19.2265ms 52.0116 Ops/s 51.9448 Ops/s $\color{#35bf28}+0.13\%$

vmoens added a commit that referenced this pull request Jan 1, 2026
ghstack-source-id: b6179e0
Pull-Request: #3286

amend

ghstack-source-id: b6179e0
Pull-Request: #3287
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.

2 participants