Skip to content

[Example] Dreamer: distributed prof profiling integration#3461

Closed
vmoens wants to merge 2 commits intogh/vmoens/222/basefrom
gh/vmoens/222/head
Closed

[Example] Dreamer: distributed prof profiling integration#3461
vmoens wants to merge 2 commits intogh/vmoens/222/basefrom
gh/vmoens/222/head

Conversation

@vmoens
Copy link
Collaborator

@vmoens vmoens commented Feb 6, 2026 •
edited
Loading

Stack from ghstack (oldest at bottom):

Integrate the prof distributed profiler into the Dreamer training loop
and collector workers for coordinated cross-process profiling.

  • dreamer_utils.py: add create_prof_handle(), extend DreamerProfiler
    with prof_handle param, step(), shm_name property, finish() cleanup.
    Add prof_shm_name param to make_collector with PROF_SHM_NAME env var.
  • dreamer.py: create prof_handle early, pass shm_name to collector,
    wrap training phases with _prof_context (sample, world_model, actor,
    value, weight_update), call profiler.finish() at cleanup.
  • config.yaml: add profiling.distributed block, raise total_optim_steps
    to 70 for prof window.
  • _runner.py: worker reads PROF_SHM_NAME/PROF_ENABLED env vars and
    calls prof.prepare() to join profiling. Wraps rollout in prof context.

Co-authored-by: Cursor [email protected]

@vmoens
Update
c734f07 
[ghstack-poisoned]
vmoens added a commit that referenced this pull request Feb 6, 2026
@vmoens @cursoragent
[Example] Dreamer: distributed prof profiling integration
08ea80e 
Integrate the prof distributed profiler into the Dreamer training loop
and collector workers for coordinated cross-process profiling.

- dreamer_utils.py: add create_prof_handle(), extend DreamerProfiler
  with prof_handle param, step(), shm_name property, finish() cleanup.
  Add prof_shm_name param to make_collector with PROF_SHM_NAME env var.
- dreamer.py: create prof_handle early, pass shm_name to collector,
  wrap training phases with _prof_context (sample, world_model, actor,
  value, weight_update), call profiler.finish() at cleanup.
- config.yaml: add profiling.distributed block, raise total_optim_steps
  to 70 for prof window.
- _runner.py: worker reads PROF_SHM_NAME/PROF_ENABLED env vars and
  calls prof.prepare() to join profiling. Wraps rollout in prof context.

Co-authored-by: Cursor <[email protected]>
ghstack-source-id: 9fcc2e8
Pull-Request: #3461
@vmoens vmoens mentioned this pull request Feb 6, 2026
Merged
@pytorch-bot
Copy link

pytorch-bot bot commented Feb 6, 2026 •
edited
Loading

🔗 Helpful Links

🧪 See artifacts and rendered test results at hud.pytorch.org/pr/pytorch/rl/3461

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

❌ 1 New Failure

As of commit 81a8672 with merge base ab49b59 (image):

NEW FAILURE - The following job has failed:

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

This was referenced Feb 6, 2026
Merged
Closed
Closed
@github-actions
Copy link
Contributor

github-actions bot commented Feb 6, 2026

⚠️ PR Title Label Error

Unknown or invalid prefix [Example].

Current title: [Example] Dreamer: distributed prof profiling integration

Supported Prefixes (case-sensitive)

Your PR title must start with exactly one of these prefixes:

Prefix Label Applied Example
[BugFix] BugFix [BugFix] Fix memory leak in collector
[Feature] Feature [Feature] Add new optimizer
[Doc] or [Docs] Documentation [Doc] Update installation guide
[Refactor] Refactoring [Refactor] Clean up module imports
[CI] CI [CI] Fix workflow permissions
[Test] or [Tests] Tests [Tests] Add unit tests for buffer
[Environment] or [Environments] Environments [Environments] Add Gymnasium support
[Data] Data [Data] Fix replay buffer sampling
[Performance] or [Perf] Performance [Performance] Optimize tensor ops
[BC-Breaking] bc breaking [BC-Breaking] Remove deprecated API
[Deprecation] Deprecation [Deprecation] Mark old function

Note: Common variations like singular/plural are supported (e.g., [Doc] or [Docs]).

@github-actions
Copy link
Contributor

github-actions bot commented Feb 6, 2026

⚠️ PR Title Label Error

Unknown or invalid prefix [Example].

Current title: [Example] Dreamer: distributed prof profiling integration

Supported Prefixes (case-sensitive)

Your PR title must start with exactly one of these prefixes:

Prefix Label Applied Example
[BugFix] BugFix [BugFix] Fix memory leak in collector
[Feature] Feature [Feature] Add new optimizer
[Doc] or [Docs] Documentation [Doc] Update installation guide
[Refactor] Refactoring [Refactor] Clean up module imports
[CI] CI [CI] Fix workflow permissions
[Test] or [Tests] Tests [Tests] Add unit tests for buffer
[Environment] or [Environments] Environments [Environments] Add Gymnasium support
[Data] Data [Data] Fix replay buffer sampling
[Performance] or [Perf] Performance [Performance] Optimize tensor ops
[BC-Breaking] bc breaking [BC-Breaking] Remove deprecated API
[Deprecation] Deprecation [Deprecation] Mark old function

Note: Common variations like singular/plural are supported (e.g., [Doc] or [Docs]).

@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 Feb 6, 2026
@vmoens
Update
81a8672 
[ghstack-poisoned]
vmoens added a commit that referenced this pull request Feb 6, 2026
@vmoens @cursoragent
[Example] Dreamer: distributed prof profiling integration
7a4b905 
Integrate the prof distributed profiler into the Dreamer training loop
and collector workers for coordinated cross-process profiling.

- dreamer_utils.py: add create_prof_handle(), extend DreamerProfiler
  with prof_handle param, step(), shm_name property, finish() cleanup.
  Add prof_shm_name param to make_collector with PROF_SHM_NAME env var.
- dreamer.py: create prof_handle early, pass shm_name to collector,
  wrap training phases with _prof_context (sample, world_model, actor,
  value, weight_update), call profiler.finish() at cleanup.
- config.yaml: add profiling.distributed block, raise total_optim_steps
  to 70 for prof window.
- _runner.py: worker reads PROF_SHM_NAME/PROF_ENABLED env vars and
  calls prof.prepare() to join profiling. Wraps rollout in prof context.

Co-authored-by: Cursor <[email protected]>
ghstack-source-id: fb5d3fa
Pull-Request: #3461
@github-actions
Copy link
Contributor

github-actions bot commented Feb 6, 2026

⚠️ PR Title Label Error

Unknown or invalid prefix [Example].

Current title: [Example] Dreamer: distributed prof profiling integration

Supported Prefixes (case-sensitive)

Your PR title must start with exactly one of these prefixes:

Prefix Label Applied Example
[BugFix] BugFix [BugFix] Fix memory leak in collector
[Feature] Feature [Feature] Add new optimizer
[Doc] or [Docs] Documentation [Doc] Update installation guide
[Refactor] Refactoring [Refactor] Clean up module imports
[CI] CI [CI] Fix workflow permissions
[Test] or [Tests] Tests [Tests] Add unit tests for buffer
[Environment] or [Environments] Environments [Environments] Add Gymnasium support
[Data] Data [Data] Fix replay buffer sampling
[Performance] or [Perf] Performance [Performance] Optimize tensor ops
[BC-Breaking] bc breaking [BC-Breaking] Remove deprecated API
[Deprecation] Deprecation [Deprecation] Mark old function

Note: Common variations like singular/plural are supported (e.g., [Doc] or [Docs]).

@github-actions
Copy link
Contributor

github-actions bot commented Feb 6, 2026

$\color{#D29922}\textsf{\Large&amp;#x26A0;\kern{0.2cm}\normalsize Warning}$ Result of CPU Benchmark Tests

Total Benchmarks: 173. Improved: $\large\color{#35bf28}20$. Worsened: $\large\color{#d91a1a}18$.

Expand to view detailed results
Name Max Mean Ops Ops on Repo HEAD Change
test_tensor_to_bytestream_speed[pickle] 81.9714μs 80.4178μs 12.4351 KOps/s 12.5940 KOps/s $\color{#d91a1a}-1.26\%$
test_tensor_to_bytestream_speed[torch.save] 0.1396ms 0.1391ms 7.1882 KOps/s 7.1587 KOps/s $\color{#35bf28}+0.41\%$
test_tensor_to_bytestream_speed[untyped_storage] 0.1199s 0.1196s 8.3610 Ops/s 8.9507 Ops/s $\textbf{\color{#d91a1a}-6.59\%}$
test_tensor_to_bytestream_speed[numpy] 2.5592μs 2.5532μs 391.6659 KOps/s 378.9859 KOps/s $\color{#35bf28}+3.35\%$
test_tensor_to_bytestream_speed[safetensors] 37.5680μs 37.3735μs 26.7569 KOps/s 25.9256 KOps/s $\color{#35bf28}+3.21\%$
test_simple 0.5491s 0.5470s 1.8281 Ops/s 1.7340 Ops/s $\textbf{\color{#35bf28}+5.43\%}$
test_transformed 1.1308s 1.1279s 0.8866 Ops/s 0.8638 Ops/s $\color{#35bf28}+2.64\%$
test_serial 1.6755s 1.6728s 0.5978 Ops/s 0.5899 Ops/s $\color{#35bf28}+1.34\%$
test_parallel 1.1615s 1.0616s 0.9419 Ops/s 0.9476 Ops/s $\color{#d91a1a}-0.60\%$
test_step_mdp_speed[True-True-True-True-True] 0.2114ms 43.7915μs 22.8355 KOps/s 21.9813 KOps/s $\color{#35bf28}+3.89\%$
test_step_mdp_speed[True-True-True-True-False] 53.9420μs 24.6790μs 40.5203 KOps/s 39.1385 KOps/s $\color{#35bf28}+3.53\%$
test_step_mdp_speed[True-True-True-False-True] 56.3220μs 24.6898μs 40.5026 KOps/s 39.4965 KOps/s $\color{#35bf28}+2.55\%$
test_step_mdp_speed[True-True-True-False-False] 45.6420μs 13.5929μs 73.5681 KOps/s 71.9130 KOps/s $\color{#35bf28}+2.30\%$
test_step_mdp_speed[True-True-False-True-True] 83.3530μs 46.9908μs 21.2808 KOps/s 20.9800 KOps/s $\color{#35bf28}+1.43\%$
test_step_mdp_speed[True-True-False-True-False] 53.2120μs 27.1959μs 36.7702 KOps/s 35.6000 KOps/s $\color{#35bf28}+3.29\%$
test_step_mdp_speed[True-True-False-False-True] 60.4530μs 26.8183μs 37.2880 KOps/s 36.0685 KOps/s $\color{#35bf28}+3.38\%$
test_step_mdp_speed[True-True-False-False-False] 43.8910μs 16.3734μs 61.0747 KOps/s 59.9966 KOps/s $\color{#35bf28}+1.80\%$
test_step_mdp_speed[True-False-True-True-True] 82.5230μs 50.3530μs 19.8598 KOps/s 19.8513 KOps/s $\color{#35bf28}+0.04\%$
test_step_mdp_speed[True-False-True-True-False] 68.0320μs 30.3181μs 32.9836 KOps/s 32.3942 KOps/s $\color{#35bf28}+1.82\%$
test_step_mdp_speed[True-False-True-False-True] 68.5630μs 27.2062μs 36.7563 KOps/s 35.9173 KOps/s $\color{#35bf28}+2.34\%$
test_step_mdp_speed[True-False-True-False-False] 44.2220μs 16.3452μs 61.1802 KOps/s 59.7286 KOps/s $\color{#35bf28}+2.43\%$
test_step_mdp_speed[True-False-False-True-True] 81.0630μs 51.2073μs 19.5285 KOps/s 18.9437 KOps/s $\color{#35bf28}+3.09\%$
test_step_mdp_speed[True-False-False-True-False] 63.3820μs 33.1215μs 30.1919 KOps/s 29.8508 KOps/s $\color{#35bf28}+1.14\%$
test_step_mdp_speed[True-False-False-False-True] 56.5020μs 30.1524μs 33.1649 KOps/s 33.0617 KOps/s $\color{#35bf28}+0.31\%$
test_step_mdp_speed[True-False-False-False-False] 62.2030μs 19.2947μs 51.8278 KOps/s 51.9971 KOps/s $\color{#d91a1a}-0.33\%$
test_step_mdp_speed[False-True-True-True-True] 88.5730μs 50.1685μs 19.9328 KOps/s 19.6631 KOps/s $\color{#35bf28}+1.37\%$
test_step_mdp_speed[False-True-True-True-False] 56.3820μs 30.1107μs 33.2108 KOps/s 32.3797 KOps/s $\color{#35bf28}+2.57\%$
test_step_mdp_speed[False-True-True-False-True] 2.3249ms 32.3444μs 30.9172 KOps/s 31.7798 KOps/s $\color{#d91a1a}-2.71\%$
test_step_mdp_speed[False-True-True-False-False] 49.5420μs 18.4213μs 54.2851 KOps/s 53.5546 KOps/s $\color{#35bf28}+1.36\%$
test_step_mdp_speed[False-True-False-True-True] 79.5320μs 52.4181μs 19.0774 KOps/s 18.1223 KOps/s $\textbf{\color{#35bf28}+5.27\%}$
test_step_mdp_speed[False-True-False-True-False] 70.5220μs 33.1365μs 30.1782 KOps/s 29.5167 KOps/s $\color{#35bf28}+2.24\%$
test_step_mdp_speed[False-True-False-False-True] 72.8330μs 33.6570μs 29.7115 KOps/s 28.7029 KOps/s $\color{#35bf28}+3.51\%$
test_step_mdp_speed[False-True-False-False-False] 54.6720μs 21.1092μs 47.3728 KOps/s 45.5479 KOps/s $\color{#35bf28}+4.01\%$
test_step_mdp_speed[False-False-True-True-True] 91.3530μs 55.7630μs 17.9330 KOps/s 17.8681 KOps/s $\color{#35bf28}+0.36\%$
test_step_mdp_speed[False-False-True-True-False] 67.3620μs 36.1521μs 27.6609 KOps/s 26.0794 KOps/s $\textbf{\color{#35bf28}+6.06\%}$
test_step_mdp_speed[False-False-True-False-True] 78.5330μs 34.3335μs 29.1261 KOps/s 29.2575 KOps/s $\color{#d91a1a}-0.45\%$
test_step_mdp_speed[False-False-True-False-False] 50.5710μs 20.5631μs 48.6307 KOps/s 47.9002 KOps/s $\color{#35bf28}+1.53\%$
test_step_mdp_speed[False-False-False-True-True] 95.0130μs 57.1223μs 17.5063 KOps/s 17.3075 KOps/s $\color{#35bf28}+1.15\%$
test_step_mdp_speed[False-False-False-True-False] 72.6030μs 38.1447μs 26.2160 KOps/s 24.5534 KOps/s $\textbf{\color{#35bf28}+6.77\%}$
test_step_mdp_speed[False-False-False-False-True] 70.7020μs 35.6375μs 28.0603 KOps/s 27.2804 KOps/s $\color{#35bf28}+2.86\%$
test_step_mdp_speed[False-False-False-False-False] 49.9320μs 23.4060μs 42.7241 KOps/s 41.8769 KOps/s $\color{#35bf28}+2.02\%$
test_non_tensor_env_rollout_speed[1000-single-True] 0.7540s 0.7498s 1.3337 Ops/s 1.2791 Ops/s $\color{#35bf28}+4.27\%$
test_non_tensor_env_rollout_speed[1000-single-False] 0.7288s 0.6361s 1.5722 Ops/s 1.5642 Ops/s $\color{#35bf28}+0.51\%$
test_non_tensor_env_rollout_speed[1000-serial-no-buffers-True] 1.7572s 1.6784s 0.5958 Ops/s 0.5918 Ops/s $\color{#35bf28}+0.67\%$
test_non_tensor_env_rollout_speed[1000-serial-no-buffers-False] 1.5364s 1.4610s 0.6845 Ops/s 0.6768 Ops/s $\color{#35bf28}+1.13\%$
test_non_tensor_env_rollout_speed[1000-serial-buffers-True] 2.0170s 1.9380s 0.5160 Ops/s 0.5138 Ops/s $\color{#35bf28}+0.42\%$
test_non_tensor_env_rollout_speed[1000-serial-buffers-False] 1.8055s 1.7162s 0.5827 Ops/s 0.5821 Ops/s $\color{#35bf28}+0.10\%$
test_non_tensor_env_rollout_speed[1000-parallel-no-buffers-True] 4.8064s 4.6847s 0.2135 Ops/s 0.2164 Ops/s $\color{#d91a1a}-1.34\%$
test_non_tensor_env_rollout_speed[1000-parallel-no-buffers-False] 4.5416s 4.4418s 0.2251 Ops/s 0.2233 Ops/s $\color{#35bf28}+0.82\%$
test_non_tensor_env_rollout_speed[1000-parallel-buffers-True] 1.9825s 1.8756s 0.5332 Ops/s 0.5290 Ops/s $\color{#35bf28}+0.78\%$
test_non_tensor_env_rollout_speed[1000-parallel-buffers-False] 1.7365s 1.6196s 0.6174 Ops/s 0.6173 Ops/s $\color{#35bf28}+0.02\%$
test_values[generalized_advantage_estimate-True-True] 11.3490ms 10.5744ms 94.5681 Ops/s 94.9608 Ops/s $\color{#d91a1a}-0.41\%$
test_values[vec_generalized_advantage_estimate-True-True] 13.3687ms 11.1628ms 89.5836 Ops/s 56.0668 Ops/s $\textbf{\color{#35bf28}+59.78\%}$
test_values[td0_return_estimate-False-False] 0.2454ms 0.1392ms 7.1850 KOps/s 7.6702 KOps/s $\textbf{\color{#d91a1a}-6.33\%}$
test_values[td1_return_estimate-False-False] 30.0891ms 28.6312ms 34.9269 Ops/s 34.8131 Ops/s $\color{#35bf28}+0.33\%$
test_values[vec_td1_return_estimate-False-False] 12.3284ms 11.2213ms 89.1164 Ops/s 55.4030 Ops/s $\textbf{\color{#35bf28}+60.85\%}$
test_values[td_lambda_return_estimate-True-False] 43.9387ms 42.3223ms 23.6282 Ops/s 23.4010 Ops/s $\color{#35bf28}+0.97\%$
test_values[vec_td_lambda_return_estimate-True-False] 11.4967ms 11.1779ms 89.4624 Ops/s 55.2546 Ops/s $\textbf{\color{#35bf28}+61.91\%}$
test_gae_speed[generalized_advantage_estimate-False-1-512] 9.5669ms 9.4397ms 105.9355 Ops/s 106.1150 Ops/s $\color{#d91a1a}-0.17\%$
test_gae_speed[vec_generalized_advantage_estimate-True-1-512] 1.7128ms 1.4620ms 683.9894 Ops/s 648.7300 Ops/s $\textbf{\color{#35bf28}+5.44\%}$
test_gae_speed[vec_generalized_advantage_estimate-False-1-512] 0.5293ms 0.4329ms 2.3101 KOps/s 2.2752 KOps/s $\color{#35bf28}+1.53\%$
test_gae_speed[vec_generalized_advantage_estimate-True-32-512] 23.0991ms 22.6817ms 44.0885 Ops/s 28.7245 Ops/s $\textbf{\color{#35bf28}+53.49\%}$
test_gae_speed[vec_generalized_advantage_estimate-False-32-512] 2.1763ms 1.7392ms 574.9699 Ops/s 575.7714 Ops/s $\color{#d91a1a}-0.14\%$
test_dqn_speed[False-None] 1.5369ms 1.4186ms 704.9047 Ops/s 709.0873 Ops/s $\color{#d91a1a}-0.59\%$
test_dqn_speed[False-backward] 2.0576ms 1.9146ms 522.3048 Ops/s 521.5697 Ops/s $\color{#35bf28}+0.14\%$
test_dqn_speed[True-None] 0.9526ms 0.5647ms 1.7709 KOps/s 1.7898 KOps/s $\color{#d91a1a}-1.06\%$
test_dqn_speed[True-backward] 1.0518ms 1.0095ms 990.6108 Ops/s 956.9437 Ops/s $\color{#35bf28}+3.52\%$
test_dqn_speed[reduce-overhead-None] 0.9382ms 0.5464ms 1.8301 KOps/s 1.8115 KOps/s $\color{#35bf28}+1.02\%$
test_ddpg_speed[False-None] 3.2321ms 2.8543ms 350.3449 Ops/s 351.3500 Ops/s $\color{#d91a1a}-0.29\%$
test_ddpg_speed[False-backward] 4.2083ms 4.0510ms 246.8537 Ops/s 247.5544 Ops/s $\color{#d91a1a}-0.28\%$
test_ddpg_speed[True-None] 1.8064ms 1.4204ms 704.0450 Ops/s 696.7728 Ops/s $\color{#35bf28}+1.04\%$
test_ddpg_speed[True-backward] 2.5242ms 2.4000ms 416.6660 Ops/s 359.6709 Ops/s $\textbf{\color{#35bf28}+15.85\%}$
test_ddpg_speed[reduce-overhead-None] 1.5067ms 1.4151ms 706.6742 Ops/s 705.9793 Ops/s $\color{#35bf28}+0.10\%$
test_sac_speed[False-None] 8.9498ms 8.0501ms 124.2217 Ops/s 123.4766 Ops/s $\color{#35bf28}+0.60\%$
test_sac_speed[False-backward] 11.7339ms 11.2702ms 88.7292 Ops/s 88.2035 Ops/s $\color{#35bf28}+0.60\%$
test_sac_speed[True-None] 2.5568ms 2.1410ms 467.0633 Ops/s 452.9656 Ops/s $\color{#35bf28}+3.11\%$
test_sac_speed[True-backward] 4.1907ms 4.0056ms 249.6516 Ops/s 237.3599 Ops/s $\textbf{\color{#35bf28}+5.18\%}$
test_sac_speed[reduce-overhead-None] 2.2362ms 2.1406ms 467.1546 Ops/s 454.6438 Ops/s $\color{#35bf28}+2.75\%$
test_redq_speed[False-None] 11.0735ms 10.4411ms 95.7753 Ops/s 89.2847 Ops/s $\textbf{\color{#35bf28}+7.27\%}$
test_redq_speed[False-backward] 18.5904ms 17.5305ms 57.0434 Ops/s 57.4974 Ops/s $\color{#d91a1a}-0.79\%$
test_redq_speed[True-None] 4.6685ms 4.3776ms 228.4378 Ops/s 220.0057 Ops/s $\color{#35bf28}+3.83\%$
test_redq_speed[True-backward] 10.0418ms 9.4883ms 105.3933 Ops/s 102.7289 Ops/s $\color{#35bf28}+2.59\%$
test_redq_speed[reduce-overhead-None] 4.6276ms 4.3939ms 227.5907 Ops/s 233.3204 Ops/s $\color{#d91a1a}-2.46\%$
test_redq_deprec_speed[False-None] 11.6888ms 10.9540ms 91.2913 Ops/s 90.3695 Ops/s $\color{#35bf28}+1.02\%$
test_redq_deprec_speed[False-backward] 16.3804ms 15.8416ms 63.1251 Ops/s 63.1792 Ops/s $\color{#d91a1a}-0.09\%$
test_redq_deprec_speed[True-None] 3.9467ms 3.5893ms 278.6087 Ops/s 270.5826 Ops/s $\color{#35bf28}+2.97\%$
test_redq_deprec_speed[True-backward] 7.6970ms 7.3856ms 135.3992 Ops/s 128.6004 Ops/s $\textbf{\color{#35bf28}+5.29\%}$
test_redq_deprec_speed[reduce-overhead-None] 3.8376ms 3.5450ms 282.0890 Ops/s 209.2363 Ops/s $\textbf{\color{#35bf28}+34.82\%}$
test_td3_speed[False-None] 8.1192ms 8.0519ms 124.1942 Ops/s 122.4665 Ops/s $\color{#35bf28}+1.41\%$
test_td3_speed[False-backward] 11.3557ms 10.9572ms 91.2640 Ops/s 91.2569 Ops/s $+0.01\%$
test_td3_speed[True-None] 1.9121ms 1.8286ms 546.8723 Ops/s 530.3264 Ops/s $\color{#35bf28}+3.12\%$
test_td3_speed[True-backward] 3.7411ms 3.5937ms 278.2648 Ops/s 268.3287 Ops/s $\color{#35bf28}+3.70\%$
test_td3_speed[reduce-overhead-None] 1.8140ms 1.7867ms 559.6858 Ops/s 550.4837 Ops/s $\color{#35bf28}+1.67\%$
test_cql_speed[False-None] 28.5991ms 25.9187ms 38.5822 Ops/s 38.3861 Ops/s $\color{#35bf28}+0.51\%$
test_cql_speed[False-backward] 35.5012ms 34.9471ms 28.6147 Ops/s 27.9983 Ops/s $\color{#35bf28}+2.20\%$
test_cql_speed[True-None] 12.5895ms 12.2032ms 81.9457 Ops/s 79.6172 Ops/s $\color{#35bf28}+2.92\%$
test_cql_speed[True-backward] 22.8980ms 19.0495ms 52.4948 Ops/s 54.1337 Ops/s $\color{#d91a1a}-3.03\%$
test_cql_speed[reduce-overhead-None] 12.4942ms 12.1998ms 81.9683 Ops/s 80.3684 Ops/s $\color{#35bf28}+1.99\%$
test_a2c_speed[False-None] 5.6766ms 5.2866ms 189.1580 Ops/s 190.1216 Ops/s $\color{#d91a1a}-0.51\%$
test_a2c_speed[False-backward] 11.7866ms 11.4974ms 86.9765 Ops/s 86.7902 Ops/s $\color{#35bf28}+0.21\%$
test_a2c_speed[True-None] 3.8402ms 3.6646ms 272.8840 Ops/s 249.4012 Ops/s $\textbf{\color{#35bf28}+9.42\%}$
test_a2c_speed[True-backward] 8.7819ms 8.5023ms 117.6148 Ops/s 117.3691 Ops/s $\color{#35bf28}+0.21\%$
test_a2c_speed[reduce-overhead-None] 4.1320ms 3.7185ms 268.9263 Ops/s 268.8408 Ops/s $\color{#35bf28}+0.03\%$
test_ppo_speed[False-None] 6.3440ms 5.9664ms 167.6064 Ops/s 172.6181 Ops/s $\color{#d91a1a}-2.90\%$
test_ppo_speed[False-backward] 12.8446ms 12.5237ms 79.8484 Ops/s 82.5908 Ops/s $\color{#d91a1a}-3.32\%$
test_ppo_speed[True-None] 3.8051ms 3.6258ms 275.8015 Ops/s 274.5920 Ops/s $\color{#35bf28}+0.44\%$
test_ppo_speed[True-backward] 8.5872ms 8.3270ms 120.0916 Ops/s 119.8195 Ops/s $\color{#35bf28}+0.23\%$
test_ppo_speed[reduce-overhead-None] 4.1099ms 3.6293ms 275.5357 Ops/s 276.4191 Ops/s $\color{#d91a1a}-0.32\%$
test_reinforce_speed[False-None] 4.7382ms 4.4461ms 224.9162 Ops/s 224.8792 Ops/s $\color{#35bf28}+0.02\%$
test_reinforce_speed[False-backward] 7.4397ms 7.2591ms 137.7579 Ops/s 137.0032 Ops/s $\color{#35bf28}+0.55\%$
test_reinforce_speed[True-None] 3.3218ms 2.8318ms 353.1330 Ops/s 346.6011 Ops/s $\color{#35bf28}+1.88\%$
test_reinforce_speed[True-backward] 7.9572ms 7.7236ms 129.4726 Ops/s 130.3618 Ops/s $\color{#d91a1a}-0.68\%$
test_reinforce_speed[reduce-overhead-None] 3.3917ms 2.8786ms 347.3884 Ops/s 329.0131 Ops/s $\textbf{\color{#35bf28}+5.58\%}$
test_iql_speed[False-None] 20.5886ms 19.8018ms 50.5005 Ops/s 49.4309 Ops/s $\color{#35bf28}+2.16\%$
test_iql_speed[False-backward] 30.8469ms 30.1955ms 33.1175 Ops/s 33.1167 Ops/s $+0.00\%$
test_iql_speed[True-None] 10.1337ms 8.5815ms 116.5297 Ops/s 124.8100 Ops/s $\textbf{\color{#d91a1a}-6.63\%}$
test_iql_speed[True-backward] 16.8548ms 16.4580ms 60.7609 Ops/s 65.0262 Ops/s $\textbf{\color{#d91a1a}-6.56\%}$
test_iql_speed[reduce-overhead-None] 8.7686ms 8.4892ms 117.7964 Ops/s 124.4472 Ops/s $\textbf{\color{#d91a1a}-5.34\%}$
test_rb_sample[TensorDictReplayBuffer-ListStorage-RandomSampler-4000] 6.2981ms 6.1095ms 163.6782 Ops/s 162.7361 Ops/s $\color{#35bf28}+0.58\%$
test_rb_sample[TensorDictReplayBuffer-LazyMemmapStorage-RandomSampler-10000] 2.7427ms 0.3283ms 3.0464 KOps/s 3.5111 KOps/s $\textbf{\color{#d91a1a}-13.23\%}$
test_rb_sample[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-10000] 0.6078ms 0.3098ms 3.2280 KOps/s 3.3479 KOps/s $\color{#d91a1a}-3.58\%$
test_rb_sample[TensorDictReplayBuffer-ListStorage-SamplerWithoutReplacement-4000] 6.1695ms 5.9103ms 169.1970 Ops/s 169.4494 Ops/s $\color{#d91a1a}-0.15\%$
test_rb_sample[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-10000] 1.0979ms 0.3409ms 2.9334 KOps/s 3.5959 KOps/s $\textbf{\color{#d91a1a}-18.43\%}$
test_rb_sample[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-10000] 0.6830ms 0.2906ms 3.4413 KOps/s 3.7554 KOps/s $\textbf{\color{#d91a1a}-8.36\%}$
test_rb_sample[TensorDictReplayBuffer-LazyMemmapStorage-sampler6-10000] 1.6234ms 1.3482ms 741.7349 Ops/s 781.6581 Ops/s $\textbf{\color{#d91a1a}-5.11\%}$
test_rb_sample[TensorDictReplayBuffer-LazyTensorStorage-sampler7-10000] 1.6129ms 1.2799ms 781.3063 Ops/s 831.5050 Ops/s $\textbf{\color{#d91a1a}-6.04\%}$
test_rb_sample[TensorDictPrioritizedReplayBuffer-ListStorage-None-4000] 9.5821ms 6.2098ms 161.0359 Ops/s 166.3845 Ops/s $\color{#d91a1a}-3.21\%$
test_rb_sample[TensorDictPrioritizedReplayBuffer-LazyMemmapStorage-None-10000] 1.8910ms 0.4528ms 2.2085 KOps/s 2.1058 KOps/s $\color{#35bf28}+4.88\%$
test_rb_sample[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-10000] 0.6669ms 0.4510ms 2.2172 KOps/s 2.3662 KOps/s $\textbf{\color{#d91a1a}-6.30\%}$
test_rb_iterate[TensorDictReplayBuffer-ListStorage-RandomSampler-4000] 6.0343ms 5.9319ms 168.5810 Ops/s 169.1624 Ops/s $\color{#d91a1a}-0.34\%$
test_rb_iterate[TensorDictReplayBuffer-LazyMemmapStorage-RandomSampler-10000] 0.9126ms 0.2811ms 3.5569 KOps/s 2.8756 KOps/s $\textbf{\color{#35bf28}+23.69\%}$
test_rb_iterate[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-10000] 0.4932ms 0.2635ms 3.7955 KOps/s 3.7588 KOps/s $\color{#35bf28}+0.98\%$
test_rb_iterate[TensorDictReplayBuffer-ListStorage-SamplerWithoutReplacement-4000] 6.2174ms 5.9519ms 168.0136 Ops/s 171.6379 Ops/s $\color{#d91a1a}-2.11\%$
test_rb_iterate[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-10000] 0.7580ms 0.3400ms 2.9412 KOps/s 3.6017 KOps/s $\textbf{\color{#d91a1a}-18.34\%}$
test_rb_iterate[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-10000] 0.4775ms 0.2904ms 3.4435 KOps/s 3.7553 KOps/s $\textbf{\color{#d91a1a}-8.30\%}$
test_rb_iterate[TensorDictPrioritizedReplayBuffer-ListStorage-None-4000] 6.2133ms 6.0734ms 164.6524 Ops/s 167.4929 Ops/s $\color{#d91a1a}-1.70\%$
test_rb_iterate[TensorDictPrioritizedReplayBuffer-LazyMemmapStorage-None-10000] 2.2860ms 0.4653ms 2.1492 KOps/s 2.1560 KOps/s $\color{#d91a1a}-0.31\%$
test_rb_iterate[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-10000] 0.6956ms 0.4515ms 2.2147 KOps/s 2.3217 KOps/s $\color{#d91a1a}-4.61\%$
test_rb_populate[TensorDictReplayBuffer-ListStorage-RandomSampler-400] 6.4152ms 4.9674ms 201.3108 Ops/s 57.1616 Ops/s $\textbf{\color{#35bf28}+252.18\%}$
test_rb_populate[TensorDictReplayBuffer-LazyMemmapStorage-RandomSampler-400] 8.9539ms 2.1538ms 464.2880 Ops/s 546.4382 Ops/s $\textbf{\color{#d91a1a}-15.03\%}$
test_rb_populate[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-400] 10.2039ms 1.2687ms 788.1990 Ops/s 913.4330 Ops/s $\textbf{\color{#d91a1a}-13.71\%}$
test_rb_populate[TensorDictReplayBuffer-ListStorage-SamplerWithoutReplacement-400] 0.5439s 15.8645ms 63.0337 Ops/s 196.6234 Ops/s $\textbf{\color{#d91a1a}-67.94\%}$
test_rb_populate[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-400] 3.8493ms 1.8154ms 550.8380 Ops/s 564.7911 Ops/s $\color{#d91a1a}-2.47\%$
test_rb_populate[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-400] 8.8579ms 1.1923ms 838.7207 Ops/s 1.1377 KOps/s $\textbf{\color{#d91a1a}-26.28\%}$
test_rb_populate[TensorDictPrioritizedReplayBuffer-ListStorage-None-400] 6.7149ms 5.2759ms 189.5429 Ops/s 59.1997 Ops/s $\textbf{\color{#35bf28}+220.18\%}$
test_rb_populate[TensorDictPrioritizedReplayBuffer-LazyMemmapStorage-None-400] 11.0944ms 2.0709ms 482.8840 Ops/s 521.9777 Ops/s $\textbf{\color{#d91a1a}-7.49\%}$
test_rb_populate[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-400] 3.5603ms 1.0817ms 924.4834 Ops/s 962.4025 Ops/s $\color{#d91a1a}-3.94\%$
test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-10000-10000-100-True] 37.7485ms 35.5808ms 28.1050 Ops/s 27.7201 Ops/s $\color{#35bf28}+1.39\%$
test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-10000-10000-100-False] 20.0912ms 18.2182ms 54.8901 Ops/s 55.2792 Ops/s $\color{#d91a1a}-0.70\%$
test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-100000-10000-100-True] 41.5286ms 36.9829ms 27.0395 Ops/s 26.5401 Ops/s $\color{#35bf28}+1.88\%$
test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-100000-10000-100-False] 19.7997ms 18.3343ms 54.5427 Ops/s 55.1110 Ops/s $\color{#d91a1a}-1.03\%$
test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-1000000-10000-100-True] 40.6718ms 38.7394ms 25.8135 Ops/s 25.1131 Ops/s $\color{#35bf28}+2.79\%$
test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-1000000-10000-100-False] 21.6333ms 20.0846ms 49.7893 Ops/s 49.8894 Ops/s $\color{#d91a1a}-0.20\%$
test_storage_write_lazystack[50-img_shape0-small] 0.8535ms 0.2195ms 4.5561 KOps/s 4.3975 KOps/s $\color{#35bf28}+3.61\%$
test_storage_write_lazystack[100-img_shape1-atari] 1.6452ms 1.3896ms 719.6240 Ops/s 722.5334 Ops/s $\color{#d91a1a}-0.40\%$
test_storage_write_lazystack[100-img_shape2-large_img] 2.6357ms 2.3896ms 418.4725 Ops/s 420.4122 Ops/s $\color{#d91a1a}-0.46\%$
test_storage_write_lazystack[200-img_shape3-large_batch] 3.1042ms 2.9287ms 341.4426 Ops/s 346.7655 Ops/s $\color{#d91a1a}-1.54\%$
test_storage_write_contiguous[50-img_shape0-small] 0.2154ms 0.1329ms 7.5248 KOps/s 7.4564 KOps/s $\color{#35bf28}+0.92\%$
test_storage_write_contiguous[100-img_shape1-atari] 0.3627ms 0.1847ms 5.4130 KOps/s 5.0930 KOps/s $\textbf{\color{#35bf28}+6.28\%}$
test_storage_write_contiguous[100-img_shape2-large_img] 1.8900ms 1.7354ms 576.2464 Ops/s 585.6858 Ops/s $\color{#d91a1a}-1.61\%$
test_storage_write_contiguous[200-img_shape3-large_batch] 1.4558ms 1.3051ms 766.2198 Ops/s 786.1812 Ops/s $\color{#d91a1a}-2.54\%$
test_collector_stack_then_write[50-img_shape0-small] 1.2360ms 1.1327ms 882.8553 Ops/s 892.7334 Ops/s $\color{#d91a1a}-1.11\%$
test_collector_stack_then_write[100-img_shape1-atari] 7.6040ms 3.6612ms 273.1361 Ops/s 280.8150 Ops/s $\color{#d91a1a}-2.73\%$
test_collector_stack_then_write[100-img_shape2-large_img] 10.4480ms 5.7237ms 174.7123 Ops/s 171.3989 Ops/s $\color{#35bf28}+1.93\%$
test_collector_stack_then_write[200-img_shape3-large_batch] 15.0928ms 7.1324ms 140.2058 Ops/s 136.8342 Ops/s $\color{#35bf28}+2.46\%$
test_collector_lazystack_then_write[50-img_shape0-small] 0.5009ms 0.2762ms 3.6202 KOps/s 3.6312 KOps/s $\color{#d91a1a}-0.31\%$
test_collector_lazystack_then_write[100-img_shape1-atari] 1.6548ms 1.5129ms 660.9723 Ops/s 669.9642 Ops/s $\color{#d91a1a}-1.34\%$
test_collector_lazystack_then_write[100-img_shape2-large_img] 2.6262ms 2.5023ms 399.6355 Ops/s 402.3489 Ops/s $\color{#d91a1a}-0.67\%$
test_collector_lazystack_then_write[200-img_shape3-large_batch] 3.2585ms 3.1306ms 319.4291 Ops/s 321.1255 Ops/s $\color{#d91a1a}-0.53\%$
test_collector_without_rb[100-img_shape0-atari] 35.3018ms 34.8078ms 28.7292 Ops/s 28.7168 Ops/s $\color{#35bf28}+0.04\%$
test_collector_without_rb[200-img_shape1-large_batch] 68.5428ms 68.3163ms 14.6378 Ops/s 14.5689 Ops/s $\color{#35bf28}+0.47\%$
test_collector_with_rb[100-img_shape0-atari] 43.6421ms 43.0471ms 23.2304 Ops/s 23.3731 Ops/s $\color{#d91a1a}-0.61\%$
test_collector_with_rb[200-img_shape1-large_batch] 0.1077s 86.6539ms 11.5402 Ops/s 11.9794 Ops/s $\color{#d91a1a}-3.67\%$

@github-actions
Copy link
Contributor

github-actions bot commented Feb 6, 2026

$\color{#D29922}\textsf{\Large&amp;#x26A0;\kern{0.2cm}\normalsize Warning}$ Result of GPU Benchmark Tests

Total Benchmarks: 172. Improved: $\large\color{#35bf28}10$. Worsened: $\large\color{#d91a1a}12$.

Expand to view detailed results
Name Max Mean Ops Ops on Repo HEAD Change
test_tensor_to_bytestream_speed[pickle] 79.7302μs 78.6843μs 12.7090 KOps/s 12.7862 KOps/s $\color{#d91a1a}-0.60\%$
test_tensor_to_bytestream_speed[torch.save] 0.1417ms 0.1390ms 7.1960 KOps/s 7.3818 KOps/s $\color{#d91a1a}-2.52\%$
test_tensor_to_bytestream_speed[untyped_storage] 0.1076s 0.1052s 9.5028 Ops/s 9.5499 Ops/s $\color{#d91a1a}-0.49\%$
test_tensor_to_bytestream_speed[numpy] 2.4537μs 2.4461μs 408.8122 KOps/s 408.9838 KOps/s $\color{#d91a1a}-0.04\%$
test_tensor_to_bytestream_speed[safetensors] 38.7597μs 37.1732μs 26.9011 KOps/s 27.3002 KOps/s $\color{#d91a1a}-1.46\%$
test_simple 0.7800s 0.7789s 1.2838 Ops/s 1.2364 Ops/s $\color{#35bf28}+3.83\%$
test_transformed 1.5125s 1.4199s 0.7043 Ops/s 0.6930 Ops/s $\color{#35bf28}+1.62\%$
test_serial 2.3707s 2.2772s 0.4391 Ops/s 0.4266 Ops/s $\color{#35bf28}+2.93\%$
test_parallel 1.9546s 1.8576s 0.5383 Ops/s 0.5437 Ops/s $\color{#d91a1a}-0.98\%$
test_step_mdp_speed[True-True-True-True-True] 0.2658ms 43.6062μs 22.9325 KOps/s 22.3227 KOps/s $\color{#35bf28}+2.73\%$
test_step_mdp_speed[True-True-True-True-False] 49.1710μs 25.0229μs 39.9634 KOps/s 39.9852 KOps/s $\color{#d91a1a}-0.05\%$
test_step_mdp_speed[True-True-True-False-True] 67.9810μs 24.2887μs 41.1714 KOps/s 41.3346 KOps/s $\color{#d91a1a}-0.39\%$
test_step_mdp_speed[True-True-True-False-False] 41.2400μs 13.4824μs 74.1708 KOps/s 75.6136 KOps/s $\color{#d91a1a}-1.91\%$
test_step_mdp_speed[True-True-False-True-True] 77.8510μs 46.7818μs 21.3758 KOps/s 21.9292 KOps/s $\color{#d91a1a}-2.52\%$
test_step_mdp_speed[True-True-False-True-False] 68.6610μs 27.1009μs 36.8991 KOps/s 37.0535 KOps/s $\color{#d91a1a}-0.42\%$
test_step_mdp_speed[True-True-False-False-True] 53.0610μs 27.3194μs 36.6041 KOps/s 37.3614 KOps/s $\color{#d91a1a}-2.03\%$
test_step_mdp_speed[True-True-False-False-False] 44.1310μs 16.2117μs 61.6838 KOps/s 60.9351 KOps/s $\color{#35bf28}+1.23\%$
test_step_mdp_speed[True-False-True-True-True] 99.4910μs 49.8634μs 20.0548 KOps/s 20.2444 KOps/s $\color{#d91a1a}-0.94\%$
test_step_mdp_speed[True-False-True-True-False] 63.5510μs 30.0311μs 33.2988 KOps/s 32.8250 KOps/s $\color{#35bf28}+1.44\%$
test_step_mdp_speed[True-False-True-False-True] 56.5010μs 27.0921μs 36.9111 KOps/s 36.5858 KOps/s $\color{#35bf28}+0.89\%$
test_step_mdp_speed[True-False-True-False-False] 50.4710μs 16.3000μs 61.3497 KOps/s 61.5019 KOps/s $\color{#d91a1a}-0.25\%$
test_step_mdp_speed[True-False-False-True-True] 91.0720μs 51.4239μs 19.4462 KOps/s 19.4152 KOps/s $\color{#35bf28}+0.16\%$
test_step_mdp_speed[True-False-False-True-False] 98.2720μs 32.1518μs 31.1024 KOps/s 30.6289 KOps/s $\color{#35bf28}+1.55\%$
test_step_mdp_speed[True-False-False-False-True] 61.6100μs 29.7844μs 33.5747 KOps/s 33.9981 KOps/s $\color{#d91a1a}-1.25\%$
test_step_mdp_speed[True-False-False-False-False] 51.4410μs 18.9064μs 52.8920 KOps/s 52.8714 KOps/s $\color{#35bf28}+0.04\%$
test_step_mdp_speed[False-True-True-True-True] 86.0910μs 49.4534μs 20.2211 KOps/s 20.5868 KOps/s $\color{#d91a1a}-1.78\%$
test_step_mdp_speed[False-True-True-True-False] 61.5910μs 30.1782μs 33.1365 KOps/s 33.2673 KOps/s $\color{#d91a1a}-0.39\%$
test_step_mdp_speed[False-True-True-False-True] 2.3422ms 30.7947μs 32.4731 KOps/s 32.8806 KOps/s $\color{#d91a1a}-1.24\%$
test_step_mdp_speed[False-True-True-False-False] 60.3000μs 18.0111μs 55.5213 KOps/s 56.3973 KOps/s $\color{#d91a1a}-1.55\%$
test_step_mdp_speed[False-True-False-True-True] 81.4810μs 52.3299μs 19.1095 KOps/s 19.3300 KOps/s $\color{#d91a1a}-1.14\%$
test_step_mdp_speed[False-True-False-True-False] 63.7910μs 32.6421μs 30.6353 KOps/s 30.9460 KOps/s $\color{#d91a1a}-1.00\%$
test_step_mdp_speed[False-True-False-False-True] 71.6210μs 33.4908μs 29.8590 KOps/s 30.8141 KOps/s $\color{#d91a1a}-3.10\%$
test_step_mdp_speed[False-True-False-False-False] 47.9410μs 20.7483μs 48.1968 KOps/s 49.3681 KOps/s $\color{#d91a1a}-2.37\%$
test_step_mdp_speed[False-False-True-True-True] 86.7110μs 54.8677μs 18.2257 KOps/s 18.2194 KOps/s $\color{#35bf28}+0.03\%$
test_step_mdp_speed[False-False-True-True-False] 80.3110μs 35.2581μs 28.3623 KOps/s 28.2548 KOps/s $\color{#35bf28}+0.38\%$
test_step_mdp_speed[False-False-True-False-True] 67.7510μs 33.5490μs 29.8072 KOps/s 29.8033 KOps/s $\color{#35bf28}+0.01\%$
test_step_mdp_speed[False-False-True-False-False] 68.9110μs 20.6498μs 48.4266 KOps/s 49.5428 KOps/s $\color{#d91a1a}-2.25\%$
test_step_mdp_speed[False-False-False-True-True] 91.1610μs 56.6479μs 17.6529 KOps/s 17.8347 KOps/s $\color{#d91a1a}-1.02\%$
test_step_mdp_speed[False-False-False-True-False] 65.6610μs 37.2144μs 26.8713 KOps/s 26.4944 KOps/s $\color{#35bf28}+1.42\%$
test_step_mdp_speed[False-False-False-False-True] 67.7010μs 35.2198μs 28.3931 KOps/s 28.3440 KOps/s $\color{#35bf28}+0.17\%$
test_step_mdp_speed[False-False-False-False-False] 51.2300μs 22.8150μs 43.8309 KOps/s 44.6531 KOps/s $\color{#d91a1a}-1.84\%$
test_non_tensor_env_rollout_speed[1000-single-True] 0.8534s 0.7639s 1.3090 Ops/s 1.3267 Ops/s $\color{#d91a1a}-1.33\%$
test_non_tensor_env_rollout_speed[1000-single-False] 0.7187s 0.6234s 1.6040 Ops/s 1.6103 Ops/s $\color{#d91a1a}-0.39\%$
test_non_tensor_env_rollout_speed[1000-serial-no-buffers-True] 1.7367s 1.6446s 0.6081 Ops/s 0.6077 Ops/s $\color{#35bf28}+0.07\%$
test_non_tensor_env_rollout_speed[1000-serial-no-buffers-False] 1.4944s 1.4183s 0.7051 Ops/s 0.6989 Ops/s $\color{#35bf28}+0.88\%$
test_non_tensor_env_rollout_speed[1000-serial-buffers-True] 1.9684s 1.8873s 0.5299 Ops/s 0.5270 Ops/s $\color{#35bf28}+0.54\%$
test_non_tensor_env_rollout_speed[1000-serial-buffers-False] 1.7447s 1.6655s 0.6004 Ops/s 0.5956 Ops/s $\color{#35bf28}+0.81\%$
test_non_tensor_env_rollout_speed[1000-parallel-no-buffers-True] 4.6796s 4.5663s 0.2190 Ops/s 0.2152 Ops/s $\color{#35bf28}+1.77\%$
test_non_tensor_env_rollout_speed[1000-parallel-no-buffers-False] 4.4785s 4.4016s 0.2272 Ops/s 0.2244 Ops/s $\color{#35bf28}+1.25\%$
test_non_tensor_env_rollout_speed[1000-parallel-buffers-True] 1.9794s 1.8922s 0.5285 Ops/s 0.5262 Ops/s $\color{#35bf28}+0.44\%$
test_non_tensor_env_rollout_speed[1000-parallel-buffers-False] 1.6777s 1.5889s 0.6294 Ops/s 0.6212 Ops/s $\color{#35bf28}+1.31\%$
test_values[generalized_advantage_estimate-True-True] 21.4074ms 20.7367ms 48.2238 Ops/s 48.5099 Ops/s $\color{#d91a1a}-0.59\%$
test_values[vec_generalized_advantage_estimate-True-True] 0.1385s 3.6941ms 270.7011 Ops/s 287.6128 Ops/s $\textbf{\color{#d91a1a}-5.88\%}$
test_values[td0_return_estimate-False-False] 0.1140ms 81.9209μs 12.2069 KOps/s 11.8899 KOps/s $\color{#35bf28}+2.67\%$
test_values[td1_return_estimate-False-False] 52.0084ms 49.9712ms 20.0115 Ops/s 20.5872 Ops/s $\color{#d91a1a}-2.80\%$
test_values[vec_td1_return_estimate-False-False] 1.3740ms 1.0864ms 920.5035 Ops/s 916.4313 Ops/s $\color{#35bf28}+0.44\%$
test_values[td_lambda_return_estimate-True-False] 79.6306ms 79.0792ms 12.6456 Ops/s 12.5459 Ops/s $\color{#35bf28}+0.79\%$
test_values[vec_td_lambda_return_estimate-True-False] 1.2884ms 1.0847ms 921.8726 Ops/s 917.6236 Ops/s $\color{#35bf28}+0.46\%$
test_gae_speed[generalized_advantage_estimate-False-1-512] 20.9546ms 20.7761ms 48.1323 Ops/s 48.0376 Ops/s $\color{#35bf28}+0.20\%$
test_gae_speed[vec_generalized_advantage_estimate-True-1-512] 1.0647ms 0.7511ms 1.3313 KOps/s 1.3112 KOps/s $\color{#35bf28}+1.54\%$
test_gae_speed[vec_generalized_advantage_estimate-False-1-512] 0.7418ms 0.6788ms 1.4732 KOps/s 1.4637 KOps/s $\color{#35bf28}+0.65\%$
test_gae_speed[vec_generalized_advantage_estimate-True-32-512] 1.5288ms 1.4898ms 671.2185 Ops/s 668.1889 Ops/s $\color{#35bf28}+0.45\%$
test_gae_speed[vec_generalized_advantage_estimate-False-32-512] 0.8129ms 0.6930ms 1.4429 KOps/s 1.4285 KOps/s $\color{#35bf28}+1.01\%$
test_dqn_speed[False-None] 1.5949ms 1.5154ms 659.8942 Ops/s 653.2036 Ops/s $\color{#35bf28}+1.02\%$
test_dqn_speed[False-backward] 2.3956ms 2.1737ms 460.0397 Ops/s 453.4731 Ops/s $\color{#35bf28}+1.45\%$
test_dqn_speed[True-None] 1.1564ms 0.5614ms 1.7812 KOps/s 1.7688 KOps/s $\color{#35bf28}+0.70\%$
test_dqn_speed[True-backward] 1.1301ms 1.0928ms 915.0575 Ops/s 909.1358 Ops/s $\color{#35bf28}+0.65\%$
test_dqn_speed[reduce-overhead-None] 0.7440ms 0.5795ms 1.7257 KOps/s 1.6625 KOps/s $\color{#35bf28}+3.80\%$
test_ddpg_speed[False-None] 3.2263ms 2.8679ms 348.6892 Ops/s 348.8748 Ops/s $\color{#d91a1a}-0.05\%$
test_ddpg_speed[False-backward] 4.6257ms 4.2010ms 238.0405 Ops/s 239.6936 Ops/s $\color{#d91a1a}-0.69\%$
test_ddpg_speed[True-None] 1.3662ms 1.3029ms 767.5388 Ops/s 759.6154 Ops/s $\color{#35bf28}+1.04\%$
test_ddpg_speed[True-backward] 2.6152ms 2.3788ms 420.3881 Ops/s 395.8590 Ops/s $\textbf{\color{#35bf28}+6.20\%}$
test_ddpg_speed[reduce-overhead-None] 1.5376ms 1.3567ms 737.0580 Ops/s 733.2170 Ops/s $\color{#35bf28}+0.52\%$
test_sac_speed[False-None] 9.1393ms 8.4387ms 118.5011 Ops/s 119.7708 Ops/s $\color{#d91a1a}-1.06\%$
test_sac_speed[False-backward] 12.0666ms 11.5029ms 86.9347 Ops/s 87.3794 Ops/s $\color{#d91a1a}-0.51\%$
test_sac_speed[True-None] 1.8534ms 1.8016ms 555.0469 Ops/s 552.0576 Ops/s $\color{#35bf28}+0.54\%$
test_sac_speed[True-backward] 3.4713ms 3.4154ms 292.7938 Ops/s 275.9848 Ops/s $\textbf{\color{#35bf28}+6.09\%}$
test_sac_speed[reduce-overhead-None] 19.5643ms 11.0288ms 90.6721 Ops/s 82.1177 Ops/s $\textbf{\color{#35bf28}+10.42\%}$
test_redq_deprec_speed[False-None] 9.8630ms 9.2749ms 107.8184 Ops/s 106.5011 Ops/s $\color{#35bf28}+1.24\%$
test_redq_deprec_speed[False-backward] 13.0648ms 12.4553ms 80.2873 Ops/s 77.1979 Ops/s $\color{#35bf28}+4.00\%$
test_redq_deprec_speed[True-None] 2.6596ms 2.5229ms 396.3769 Ops/s 396.4184 Ops/s $\color{#d91a1a}-0.01\%$
test_redq_deprec_speed[True-backward] 4.5795ms 4.2927ms 232.9516 Ops/s 228.9328 Ops/s $\color{#35bf28}+1.76\%$
test_redq_deprec_speed[reduce-overhead-None] 16.3617ms 9.8832ms 101.1817 Ops/s 100.7439 Ops/s $\color{#35bf28}+0.43\%$
test_td3_speed[False-None] 8.2081ms 8.1300ms 123.0013 Ops/s 120.1896 Ops/s $\color{#35bf28}+2.34\%$
test_td3_speed[False-backward] 12.4791ms 10.9977ms 90.9283 Ops/s 91.6869 Ops/s $\color{#d91a1a}-0.83\%$
test_td3_speed[True-None] 1.7305ms 1.6391ms 610.0739 Ops/s 592.4491 Ops/s $\color{#35bf28}+2.97\%$
test_td3_speed[True-backward] 3.3213ms 3.2410ms 308.5515 Ops/s 300.4419 Ops/s $\color{#35bf28}+2.70\%$
test_td3_speed[reduce-overhead-None] 72.3890ms 24.7920ms 40.3355 Ops/s 40.6983 Ops/s $\color{#d91a1a}-0.89\%$
test_cql_speed[False-None] 17.3035ms 17.0478ms 58.6585 Ops/s 57.9531 Ops/s $\color{#35bf28}+1.22\%$
test_cql_speed[False-backward] 22.8615ms 22.4263ms 44.5905 Ops/s 43.2893 Ops/s $\color{#35bf28}+3.01\%$
test_cql_speed[True-None] 3.4561ms 3.2446ms 308.2056 Ops/s 306.3068 Ops/s $\color{#35bf28}+0.62\%$
test_cql_speed[True-backward] 5.9536ms 5.3766ms 185.9904 Ops/s 179.5191 Ops/s $\color{#35bf28}+3.60\%$
test_cql_speed[reduce-overhead-None] 19.5312ms 11.9722ms 83.5271 Ops/s 83.6281 Ops/s $\color{#d91a1a}-0.12\%$
test_a2c_speed[False-None] 4.0005ms 3.2113ms 311.3957 Ops/s 307.5976 Ops/s $\color{#35bf28}+1.23\%$
test_a2c_speed[False-backward] 6.5720ms 6.2236ms 160.6780 Ops/s 154.6403 Ops/s $\color{#35bf28}+3.90\%$
test_a2c_speed[True-None] 1.3772ms 1.3002ms 769.1123 Ops/s 741.8681 Ops/s $\color{#35bf28}+3.67\%$
test_a2c_speed[True-backward] 3.1153ms 2.9702ms 336.6741 Ops/s 322.1185 Ops/s $\color{#35bf28}+4.52\%$
test_a2c_speed[reduce-overhead-None] 1.0546ms 0.9752ms 1.0254 KOps/s 1.0343 KOps/s $\color{#d91a1a}-0.86\%$
test_ppo_speed[False-None] 4.0429ms 3.8793ms 257.7768 Ops/s 256.4023 Ops/s $\color{#35bf28}+0.54\%$
test_ppo_speed[False-backward] 7.4520ms 7.0350ms 142.1471 Ops/s 137.2273 Ops/s $\color{#35bf28}+3.59\%$
test_ppo_speed[True-None] 1.4911ms 1.4128ms 707.8307 Ops/s 710.2691 Ops/s $\color{#d91a1a}-0.34\%$
test_ppo_speed[True-backward] 3.2865ms 3.2372ms 308.9124 Ops/s 302.8261 Ops/s $\color{#35bf28}+2.01\%$
test_ppo_speed[reduce-overhead-None] 1.1037ms 1.0291ms 971.6951 Ops/s 949.0906 Ops/s $\color{#35bf28}+2.38\%$
test_reinforce_speed[False-None] 2.4146ms 2.2636ms 441.7722 Ops/s 435.8453 Ops/s $\color{#35bf28}+1.36\%$
test_reinforce_speed[False-backward] 3.8636ms 3.4418ms 290.5418 Ops/s 289.4021 Ops/s $\color{#35bf28}+0.39\%$
test_reinforce_speed[True-None] 1.4444ms 1.2750ms 784.3392 Ops/s 771.1695 Ops/s $\color{#35bf28}+1.71\%$
test_reinforce_speed[True-backward] 3.0641ms 3.0216ms 330.9528 Ops/s 327.0702 Ops/s $\color{#35bf28}+1.19\%$
test_reinforce_speed[reduce-overhead-None] 17.5721ms 9.5031ms 105.2292 Ops/s 104.4530 Ops/s $\color{#35bf28}+0.74\%$
test_iql_speed[False-None] 9.9641ms 9.3900ms 106.4965 Ops/s 105.5151 Ops/s $\color{#35bf28}+0.93\%$
test_iql_speed[False-backward] 13.8885ms 13.4048ms 74.6002 Ops/s 73.7726 Ops/s $\color{#35bf28}+1.12\%$
test_iql_speed[True-None] 2.3561ms 2.1589ms 463.2091 Ops/s 456.2250 Ops/s $\color{#35bf28}+1.53\%$
test_iql_speed[True-backward] 5.1153ms 4.8117ms 207.8271 Ops/s 202.5516 Ops/s $\color{#35bf28}+2.60\%$
test_iql_speed[reduce-overhead-None] 18.2412ms 10.5763ms 94.5507 Ops/s 94.4060 Ops/s $\color{#35bf28}+0.15\%$
test_rb_sample[TensorDictReplayBuffer-ListStorage-RandomSampler-4000] 6.2469ms 5.7540ms 173.7925 Ops/s 173.5837 Ops/s $\color{#35bf28}+0.12\%$
test_rb_sample[TensorDictReplayBuffer-LazyMemmapStorage-RandomSampler-10000] 0.9689ms 0.2815ms 3.5526 KOps/s 3.0118 KOps/s $\textbf{\color{#35bf28}+17.95\%}$
test_rb_sample[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-10000] 0.5151ms 0.2652ms 3.7705 KOps/s 3.2519 KOps/s $\textbf{\color{#35bf28}+15.95\%}$
test_rb_sample[TensorDictReplayBuffer-ListStorage-SamplerWithoutReplacement-4000] 5.8722ms 5.5397ms 180.5141 Ops/s 181.7990 Ops/s $\color{#d91a1a}-0.71\%$
test_rb_sample[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-10000] 0.5273ms 0.2931ms 3.4118 KOps/s 3.2901 KOps/s $\color{#35bf28}+3.70\%$
test_rb_sample[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-10000] 0.6565ms 0.3060ms 3.2682 KOps/s 3.2746 KOps/s $\color{#d91a1a}-0.20\%$
test_rb_sample[TensorDictReplayBuffer-LazyMemmapStorage-sampler6-10000] 1.6957ms 1.2694ms 787.7550 Ops/s 785.5102 Ops/s $\color{#35bf28}+0.29\%$
test_rb_sample[TensorDictReplayBuffer-LazyTensorStorage-sampler7-10000] 1.4422ms 1.1905ms 839.9995 Ops/s 843.7813 Ops/s $\color{#d91a1a}-0.45\%$
test_rb_sample[TensorDictPrioritizedReplayBuffer-ListStorage-None-4000] 6.1995ms 5.7503ms 173.9045 Ops/s 176.4271 Ops/s $\color{#d91a1a}-1.43\%$
test_rb_sample[TensorDictPrioritizedReplayBuffer-LazyMemmapStorage-None-10000] 2.3271ms 0.5208ms 1.9201 KOps/s 2.2984 KOps/s $\textbf{\color{#d91a1a}-16.46\%}$
test_rb_sample[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-10000] 0.7188ms 0.5031ms 1.9876 KOps/s 2.4311 KOps/s $\textbf{\color{#d91a1a}-18.24\%}$
test_rb_iterate[TensorDictReplayBuffer-ListStorage-RandomSampler-4000] 6.1610ms 5.6825ms 175.9798 Ops/s 173.7787 Ops/s $\color{#35bf28}+1.27\%$
test_rb_iterate[TensorDictReplayBuffer-LazyMemmapStorage-RandomSampler-10000] 0.7530ms 0.3541ms 2.8239 KOps/s 3.5642 KOps/s $\textbf{\color{#d91a1a}-20.77\%}$
test_rb_iterate[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-10000] 0.6045ms 0.3401ms 2.9406 KOps/s 3.8058 KOps/s $\textbf{\color{#d91a1a}-22.73\%}$
test_rb_iterate[TensorDictReplayBuffer-ListStorage-SamplerWithoutReplacement-4000] 6.2045ms 5.6275ms 177.6973 Ops/s 177.6494 Ops/s $\color{#35bf28}+0.03\%$
test_rb_iterate[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-10000] 1.1376ms 0.3523ms 2.8388 KOps/s 3.3293 KOps/s $\textbf{\color{#d91a1a}-14.73\%}$
test_rb_iterate[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-10000] 0.5416ms 0.3398ms 2.9428 KOps/s 3.6793 KOps/s $\textbf{\color{#d91a1a}-20.02\%}$
test_rb_iterate[TensorDictPrioritizedReplayBuffer-ListStorage-None-4000] 6.0688ms 5.7928ms 172.6279 Ops/s 170.7655 Ops/s $\color{#35bf28}+1.09\%$
test_rb_iterate[TensorDictPrioritizedReplayBuffer-LazyMemmapStorage-None-10000] 1.9738ms 0.5299ms 1.8872 KOps/s 2.1507 KOps/s $\textbf{\color{#d91a1a}-12.25\%}$
test_rb_iterate[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-10000] 0.6174ms 0.4160ms 2.4037 KOps/s 2.3177 KOps/s $\color{#35bf28}+3.71\%$
test_rb_populate[TensorDictReplayBuffer-ListStorage-RandomSampler-400] 0.6080s 17.2109ms 58.1026 Ops/s 199.5611 Ops/s $\textbf{\color{#d91a1a}-70.88\%}$
test_rb_populate[TensorDictReplayBuffer-LazyMemmapStorage-RandomSampler-400] 9.4041ms 2.0445ms 489.1091 Ops/s 433.0783 Ops/s $\textbf{\color{#35bf28}+12.94\%}$
test_rb_populate[TensorDictReplayBuffer-LazyTensorStorage-RandomSampler-400] 10.4729ms 1.3188ms 758.2758 Ops/s 1.0600 KOps/s $\textbf{\color{#d91a1a}-28.46\%}$
test_rb_populate[TensorDictReplayBuffer-ListStorage-SamplerWithoutReplacement-400] 7.5527ms 5.1836ms 192.9150 Ops/s 50.4632 Ops/s $\textbf{\color{#35bf28}+282.29\%}$
test_rb_populate[TensorDictReplayBuffer-LazyMemmapStorage-SamplerWithoutReplacement-400] 3.9898ms 1.8153ms 550.8852 Ops/s 528.8491 Ops/s $\color{#35bf28}+4.17\%$
test_rb_populate[TensorDictReplayBuffer-LazyTensorStorage-SamplerWithoutReplacement-400] 1.0937ms 0.9505ms 1.0521 KOps/s 868.0691 Ops/s $\textbf{\color{#35bf28}+21.20\%}$
test_rb_populate[TensorDictPrioritizedReplayBuffer-ListStorage-None-400] 8.3841ms 5.4134ms 184.7253 Ops/s 188.0729 Ops/s $\color{#d91a1a}-1.78\%$
test_rb_populate[TensorDictPrioritizedReplayBuffer-LazyMemmapStorage-None-400] 4.2034ms 1.9038ms 525.2550 Ops/s 507.0348 Ops/s $\color{#35bf28}+3.59\%$
test_rb_populate[TensorDictPrioritizedReplayBuffer-LazyTensorStorage-None-400] 3.8913ms 1.1681ms 856.0703 Ops/s 930.5498 Ops/s $\textbf{\color{#d91a1a}-8.00\%}$
test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-10000-10000-100-True] 37.6822ms 35.7817ms 27.9472 Ops/s 27.5141 Ops/s $\color{#35bf28}+1.57\%$
test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-10000-10000-100-False] 0.5867s 29.5420ms 33.8502 Ops/s 54.0830 Ops/s $\textbf{\color{#d91a1a}-37.41\%}$
test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-100000-10000-100-True] 39.5651ms 36.6801ms 27.2627 Ops/s 26.9916 Ops/s $\color{#35bf28}+1.00\%$
test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-100000-10000-100-False] 19.7069ms 18.1734ms 55.0255 Ops/s 53.7949 Ops/s $\color{#35bf28}+2.29\%$
test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-1000000-10000-100-True] 40.3450ms 38.1327ms 26.2242 Ops/s 25.6627 Ops/s $\color{#35bf28}+2.19\%$
test_rb_extend_sample[ReplayBuffer-LazyTensorStorage-RandomSampler-1000000-10000-100-False] 21.3706ms 19.8437ms 50.3938 Ops/s 50.3783 Ops/s $\color{#35bf28}+0.03\%$
test_storage_write_lazystack[50-img_shape0-small] 0.8727ms 0.2150ms 4.6501 KOps/s 4.3202 KOps/s $\textbf{\color{#35bf28}+7.64\%}$
test_storage_write_lazystack[100-img_shape1-atari] 1.9024ms 1.4402ms 694.3281 Ops/s 695.1602 Ops/s $\color{#d91a1a}-0.12\%$
test_storage_write_lazystack[100-img_shape2-large_img] 2.8296ms 2.2987ms 435.0303 Ops/s 428.5054 Ops/s $\color{#35bf28}+1.52\%$
test_storage_write_lazystack[200-img_shape3-large_batch] 3.4655ms 2.9646ms 337.3140 Ops/s 332.7779 Ops/s $\color{#35bf28}+1.36\%$
test_storage_write_contiguous[50-img_shape0-small] 0.3172ms 0.1653ms 6.0488 KOps/s 6.0895 KOps/s $\color{#d91a1a}-0.67\%$
test_storage_write_contiguous[100-img_shape1-atari] 0.3493ms 0.2318ms 4.3140 KOps/s 4.3761 KOps/s $\color{#d91a1a}-1.42\%$
test_storage_write_contiguous[100-img_shape2-large_img] 2.0105ms 1.8523ms 539.8592 Ops/s 545.7263 Ops/s $\color{#d91a1a}-1.08\%$
test_storage_write_contiguous[200-img_shape3-large_batch] 1.8751ms 1.4152ms 706.6159 Ops/s 709.4145 Ops/s $\color{#d91a1a}-0.39\%$
test_collector_stack_then_write[50-img_shape0-small] 1.5713ms 1.1156ms 896.4015 Ops/s 889.9915 Ops/s $\color{#35bf28}+0.72\%$
test_collector_stack_then_write[100-img_shape1-atari] 4.0076ms 3.5782ms 279.4734 Ops/s 266.5387 Ops/s $\color{#35bf28}+4.85\%$
test_collector_stack_then_write[100-img_shape2-large_img] 11.2621ms 5.9898ms 166.9499 Ops/s 174.0638 Ops/s $\color{#d91a1a}-4.09\%$
test_collector_stack_then_write[200-img_shape3-large_batch] 7.9761ms 7.0882ms 141.0804 Ops/s 139.5463 Ops/s $\color{#35bf28}+1.10\%$
test_collector_lazystack_then_write[50-img_shape0-small] 0.7353ms 0.2739ms 3.6512 KOps/s 3.6915 KOps/s $\color{#d91a1a}-1.09\%$
test_collector_lazystack_then_write[100-img_shape1-atari] 2.0138ms 1.5637ms 639.5094 Ops/s 642.2523 Ops/s $\color{#d91a1a}-0.43\%$
test_collector_lazystack_then_write[100-img_shape2-large_img] 2.8762ms 2.4231ms 412.6955 Ops/s 415.9917 Ops/s $\color{#d91a1a}-0.79\%$
test_collector_lazystack_then_write[200-img_shape3-large_batch] 3.5916ms 3.1790ms 314.5635 Ops/s 313.4528 Ops/s $\color{#35bf28}+0.35\%$
test_collector_without_rb[100-img_shape0-atari] 35.7897ms 34.4612ms 29.0181 Ops/s 29.0867 Ops/s $\color{#d91a1a}-0.24\%$
test_collector_without_rb[200-img_shape1-large_batch] 67.9170ms 67.1821ms 14.8849 Ops/s 14.8704 Ops/s $\color{#35bf28}+0.10\%$
test_collector_with_rb[100-img_shape0-atari] 43.5685ms 42.0977ms 23.7543 Ops/s 23.9272 Ops/s $\color{#d91a1a}-0.72\%$
test_collector_with_rb[200-img_shape1-large_batch] 83.9824ms 82.6520ms 12.0989 Ops/s 12.2174 Ops/s $\color{#d91a1a}-0.97\%$
test_collector_without_rb_cuda[100-img_shape0-atari] 57.0012ms 56.7959ms 17.6069 Ops/s 17.5768 Ops/s $\color{#35bf28}+0.17\%$
test_collector_without_rb_cuda[200-img_shape1-large_batch] 0.1186s 0.1142s 8.7557 Ops/s 4.7531 Ops/s $\textbf{\color{#35bf28}+84.21\%}$
test_collector_with_rb_cuda[100-img_shape0-atari] 62.4753ms 62.0260ms 16.1223 Ops/s 15.9481 Ops/s $\color{#35bf28}+1.09\%$
test_collector_with_rb_cuda[200-img_shape1-large_batch] 0.1295s 0.1249s 8.0094 Ops/s 7.9673 Ops/s $\color{#35bf28}+0.53\%$

@vmoens
Copy link
Collaborator Author

vmoens commented Feb 7, 2026

Closing: prof instrumentation not needed in the final stack.

@vmoens vmoens closed this Feb 7, 2026
@vmoens vmoens deleted the gh/vmoens/222/head branch February 7, 2026 15:44
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Reviewers

No reviews

Assignees

No one assigned

Labels

CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. Collectors sota-implementations/

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

1 participant

@vmoens