Skip to content

[Test] Reorganize test suite into per-area subdirectories - #1726

Merged
vmoens merged 3 commits into
mainfrom
claude/friendly-jang-a55564
Jun 15, 2026
Merged

[Test] Reorganize test suite into per-area subdirectories#1726
vmoens merged 3 commits into
mainfrom
claude/friendly-jang-a55564

Conversation

@vmoens

@vmoens vmoens commented Jun 15, 2026

Copy link
Copy Markdown
Collaborator

What

Reorganizes the flat test/ directory into per-area subdirectories, and
splits the 16k-line test/test_tensordict.py (607 KB) into nine themed
modules.

test_tensordict.py → test/tensordict/ (each class moved intact):

file classes
test_generic.py TestGeneric
test_methods.py TestTensorDicts
test_lazy.py TestLazyStackedTensorDict
test_nameddims.py TestNamedDims
test_nontensor.py TestNonTensorData, TestMetaData, TestUnbatchedTensor
test_repr.py TestTensorDictRepr, TestSetPrintoptions
test_mp.py TestMPInplace, TestTensorDictMP, TestMap, TestFCD
test_errors.py TestErrorMessage, TestErrors, TestLock
test_misc.py TestPointwiseOps, TestJacobians, TestTensorDictsRequiresGrad, TestMakeTensorDict, TestSubclassing, TestLikeConstructors, TestMemmap, TestFreeThreading, TestFromSchema

test_tabular.py also moved here.

Other files moved by area (pure git mv, history preserved):
tensorclass/ (tensorclass, typedtensordict, compatibility), nn/ (nn, fx,
functorch), compile/, memmap/ (memmap, h5, td_memmap_split), store/,
distributed/ (distributed, ucxx), utils/ (utils, setup, *_split).

_utils_internal.py, conftest.py, and smoke_test.py stay at the test/
root.

Why

test_tensordict.py had grown to 16,015 lines, making it slow to open,
navigate, and review, and the flat test/ directory mixed ~24 unrelated
suites. Per-area subdirs also let you run a slice (e.g. pytest test/nn).

How it stays correct

  • No tests lost or duplicated: every class and method was sliced out via
    AST (not hand-editing). A programmatic conservation check confirms the nine
    new files contain exactly the original 25 classes / 714 methods, with no
    class appearing twice and identical method sets.
  • Imports keep resolving: the bare from _utils_internal import ... used
    by several suites relies on test/ being on sys.path. Keeping
    _utils_internal.py + conftest.py at the test/ root preserves this — the
    root conftest.py makes pytest prepend test/ before any subdir module
    loads (verified with a standalone pytest sandbox). The fbcode branch uses the
    absolute pytorch.tensordict.test._utils_internal, unaffected by location.
  • No __init__.py anywhere under test/, so pytest's prepend import mode
    is unchanged; testpaths = test auto-discovers the new layout with no config
    change.
  • Helpers/fixtures travel with their sole consumer (device_fixture /
    _path_td_sync → generic, _compare_tensors_identity / _to_float →
    methods, _pool_fixt → mp).
  • New files: unused imports trimmed (ruff F401), formatted with ufmt
    (black + usort), clean under flake8 --config=setup.cfg.
  • .github/labeler.yml globs repointed to the new paths.

Note

The full pytest suite was not run locally (no torch environment available), so
the live collection/run is left to CI. The source-level conservation check and
the import-mechanism sandbox give high confidence the move is
behavior-preserving. Separately, Meta-internal fbcode TARGETS that enumerate
test paths will need a matching internal update.

🤖 Generated with Claude Code

@vmoens @claude
[Test] Reorganize test suite into per-area subdirectories
377b3de 
Split the 16k-line test/test_tensordict.py into nine themed modules under
test/tensordict/ (generic, methods, lazy, nameddims, nontensor, repr, mp,
errors, misc) and move the remaining flat test files into per-area
subdirectories: tensorclass/, nn/, compile/, memmap/, store/, distributed/,
utils/.

Shared infrastructure (_utils_internal.py, conftest.py) and smoke_test.py
stay at the test/ root, so the bare `from _utils_internal import ...` imports
keep resolving via the root conftest's sys.path insertion. No __init__.py is
added, which keeps pytest's prepend import mode intact. pytest auto-discovery
(testpaths = test) picks up the new layout with no config change.

The split preserves every class and test method exactly (25 classes, 714
methods, verified via AST). Helpers and fixtures travel with their sole
consumer (device_fixture/_path_td_sync -> generic, _compare_tensors_identity/
_to_float -> methods, _pool_fixt -> mp). Per-file unused imports were trimmed
and the new files formatted with ufmt.

Updated .github/labeler.yml globs to the new test paths.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
@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 Jun 15, 2026
@github-actions

github-actions Bot commented Jun 15, 2026 •
edited
Loading

Copy link
Copy Markdown
Contributor

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

Total Benchmarks: 344. Improved: $\large\color{#35bf28}10$. Worsened: $\large\color{#d91a1a}51$.

Expand to view detailed results
Name Max Mean Ops Ops on Repo HEAD Change
test_collector_preallocated_write[stack_out-cpu-32-2] 0.2167ms 0.1723ms 5.8030 KOps/s 5.9464 KOps/s $\color{#d91a1a}-2.41\%$
test_collector_preallocated_write[stack_out-cpu-32-8] 0.3467ms 0.2977ms 3.3593 KOps/s 3.3796 KOps/s $\color{#d91a1a}-0.60\%$
test_collector_preallocated_write[stack_out-cpu-32-32] 0.8340ms 0.7848ms 1.2743 KOps/s 1.2736 KOps/s $\color{#35bf28}+0.05\%$
test_collector_preallocated_write[stack_out-cpu-4096-2] 0.4841ms 0.3669ms 2.7259 KOps/s 2.8670 KOps/s $\color{#d91a1a}-4.92\%$
test_collector_preallocated_write[stack_out-cpu-4096-8] 1.5134ms 1.1712ms 853.8058 Ops/s 923.6196 Ops/s $\textbf{\color{#d91a1a}-7.56\%}$
test_collector_preallocated_write[stack_out-cpu-4096-32] 7.6386ms 7.3349ms 136.3344 Ops/s 133.7101 Ops/s $\color{#35bf28}+1.96\%$
test_collector_preallocated_write[stack_out-cuda-32-2] 0.3191ms 0.2647ms 3.7782 KOps/s 3.9720 KOps/s $\color{#d91a1a}-4.88\%$
test_collector_preallocated_write[stack_out-cuda-32-8] 0.4844ms 0.3821ms 2.6170 KOps/s 2.6857 KOps/s $\color{#d91a1a}-2.56\%$
test_collector_preallocated_write[stack_out-cuda-32-32] 0.9086ms 0.8309ms 1.2035 KOps/s 1.2077 KOps/s $\color{#d91a1a}-0.35\%$
test_collector_preallocated_write[stack_out-cuda-4096-2] 0.3205ms 0.2683ms 3.7268 KOps/s 3.9831 KOps/s $\textbf{\color{#d91a1a}-6.44\%}$
test_collector_preallocated_write[stack_out-cuda-4096-8] 0.4770ms 0.3842ms 2.6030 KOps/s 2.6749 KOps/s $\color{#d91a1a}-2.69\%$
test_collector_preallocated_write[stack_out-cuda-4096-32] 0.9780ms 0.8910ms 1.1223 KOps/s 1.1260 KOps/s $\color{#d91a1a}-0.33\%$
test_collector_preallocated_write[set_at_loop-cpu-32-2] 0.2768ms 0.2210ms 4.5250 KOps/s 4.6678 KOps/s $\color{#d91a1a}-3.06\%$
test_collector_preallocated_write[set_at_loop-cpu-32-8] 0.9634ms 0.8555ms 1.1689 KOps/s 1.2057 KOps/s $\color{#d91a1a}-3.05\%$
test_collector_preallocated_write[set_at_loop-cpu-32-32] 3.5476ms 3.3966ms 294.4158 Ops/s 302.6606 Ops/s $\color{#d91a1a}-2.72\%$
test_collector_preallocated_write[set_at_loop-cpu-4096-2] 0.7299ms 0.3809ms 2.6253 KOps/s 2.6829 KOps/s $\color{#d91a1a}-2.15\%$
test_collector_preallocated_write[set_at_loop-cpu-4096-8] 2.0183ms 1.8302ms 546.3822 Ops/s 582.5468 Ops/s $\textbf{\color{#d91a1a}-6.21\%}$
test_collector_preallocated_write[set_at_loop-cpu-4096-32] 14.1827ms 10.8060ms 92.5412 Ops/s 96.8442 Ops/s $\color{#d91a1a}-4.44\%$
test_collector_preallocated_write[set_at_loop-cuda-32-2] 0.4421ms 0.4087ms 2.4465 KOps/s 2.5368 KOps/s $\color{#d91a1a}-3.56\%$
test_collector_preallocated_write[set_at_loop-cuda-32-8] 1.6813ms 1.6315ms 612.9186 Ops/s 669.4744 Ops/s $\textbf{\color{#d91a1a}-8.45\%}$
test_collector_preallocated_write[set_at_loop-cuda-32-32] 6.3562ms 5.9967ms 166.7592 Ops/s 172.0395 Ops/s $\color{#d91a1a}-3.07\%$
test_collector_preallocated_write[set_at_loop-cuda-4096-2] 0.4865ms 0.4374ms 2.2862 KOps/s 2.5341 KOps/s $\textbf{\color{#d91a1a}-9.78\%}$
test_collector_preallocated_write[set_at_loop-cuda-4096-8] 1.8628ms 1.6394ms 609.9718 Ops/s 678.3881 Ops/s $\textbf{\color{#d91a1a}-10.09\%}$
test_collector_preallocated_write[set_at_loop-cuda-4096-32] 6.6243ms 6.2874ms 159.0490 Ops/s 172.4564 Ops/s $\textbf{\color{#d91a1a}-7.77\%}$
test_collector_preallocated_write[copy_at-cpu-32-2] 0.6235ms 0.1852ms 5.4007 KOps/s 5.7258 KOps/s $\textbf{\color{#d91a1a}-5.68\%}$
test_collector_preallocated_write[copy_at-cpu-32-8] 0.8088ms 0.7140ms 1.4006 KOps/s 1.4680 KOps/s $\color{#d91a1a}-4.59\%$
test_collector_preallocated_write[copy_at-cpu-32-32] 2.9923ms 2.8439ms 351.6355 Ops/s 369.5797 Ops/s $\color{#d91a1a}-4.86\%$
test_collector_preallocated_write[copy_at-cpu-4096-2] 0.6765ms 0.3237ms 3.0893 KOps/s 3.1134 KOps/s $\color{#d91a1a}-0.77\%$
test_collector_preallocated_write[copy_at-cpu-4096-8] 1.7244ms 1.5678ms 637.8316 Ops/s 678.7030 Ops/s $\textbf{\color{#d91a1a}-6.02\%}$
test_collector_preallocated_write[copy_at-cpu-4096-32] 10.0603ms 9.7004ms 103.0885 Ops/s 103.4233 Ops/s $\color{#d91a1a}-0.32\%$
test_collector_preallocated_write[copy_at-cuda-32-2] 0.4130ms 0.3348ms 2.9872 KOps/s 3.0180 KOps/s $\color{#d91a1a}-1.02\%$
test_collector_preallocated_write[copy_at-cuda-32-8] 1.3286ms 1.2490ms 800.6634 Ops/s 811.5106 Ops/s $\color{#d91a1a}-1.34\%$
test_collector_preallocated_write[copy_at-cuda-32-32] 5.0069ms 4.8996ms 204.0982 Ops/s 207.1656 Ops/s $\color{#d91a1a}-1.48\%$
test_collector_preallocated_write[copy_at-cuda-4096-2] 0.4645ms 0.3359ms 2.9769 KOps/s 3.0047 KOps/s $\color{#d91a1a}-0.93\%$
test_collector_preallocated_write[copy_at-cuda-4096-8] 1.3566ms 1.2562ms 796.0265 Ops/s 800.5745 Ops/s $\color{#d91a1a}-0.57\%$
test_collector_preallocated_write[copy_at-cuda-4096-32] 5.0262ms 4.9260ms 203.0024 Ops/s 206.8564 Ops/s $\color{#d91a1a}-1.86\%$
test_collector_preallocated_write[manual_leaf_copy-cpu-32-2] 0.1646ms 0.1260ms 7.9385 KOps/s 8.3541 KOps/s $\color{#d91a1a}-4.97\%$
test_collector_preallocated_write[manual_leaf_copy-cpu-32-8] 0.5494ms 0.4948ms 2.0210 KOps/s 2.1320 KOps/s $\textbf{\color{#d91a1a}-5.21\%}$
test_collector_preallocated_write[manual_leaf_copy-cpu-32-32] 2.0408ms 1.9622ms 509.6435 Ops/s 537.0241 Ops/s $\textbf{\color{#d91a1a}-5.10\%}$
test_collector_preallocated_write[manual_leaf_copy-cpu-4096-2] 0.3219ms 0.2647ms 3.7782 KOps/s 3.9728 KOps/s $\color{#d91a1a}-4.90\%$
test_collector_preallocated_write[manual_leaf_copy-cpu-4096-8] 1.4384ms 1.3130ms 761.6394 Ops/s 819.1389 Ops/s $\textbf{\color{#d91a1a}-7.02\%}$
test_collector_preallocated_write[manual_leaf_copy-cpu-4096-32] 8.9913ms 8.7427ms 114.3814 Ops/s 114.2618 Ops/s $\color{#35bf28}+0.10\%$
test_collector_preallocated_write[manual_leaf_copy-cuda-32-2] 0.3618ms 0.3073ms 3.2541 KOps/s 3.6059 KOps/s $\textbf{\color{#d91a1a}-9.76\%}$
test_collector_preallocated_write[manual_leaf_copy-cuda-32-8] 1.1789ms 1.0937ms 914.3280 Ops/s 971.7650 Ops/s $\textbf{\color{#d91a1a}-5.91\%}$
test_collector_preallocated_write[manual_leaf_copy-cuda-32-32] 4.5008ms 4.2841ms 233.4217 Ops/s 248.3297 Ops/s $\textbf{\color{#d91a1a}-6.00\%}$
test_collector_preallocated_write[manual_leaf_copy-cuda-4096-2] 0.3700ms 0.3103ms 3.2222 KOps/s 3.5971 KOps/s $\textbf{\color{#d91a1a}-10.42\%}$
test_collector_preallocated_write[manual_leaf_copy-cuda-4096-8] 1.1984ms 1.1094ms 901.4265 Ops/s 964.4074 Ops/s $\textbf{\color{#d91a1a}-6.53\%}$
test_collector_preallocated_write[manual_leaf_copy-cuda-4096-32] 4.5168ms 4.3213ms 231.4114 Ops/s 245.9014 Ops/s $\textbf{\color{#d91a1a}-5.89\%}$
test_wide_nested_preallocated_write[stack_out-cpu-128-128] 16.5082ms 16.3120ms 61.3045 Ops/s 61.3363 Ops/s $\color{#d91a1a}-0.05\%$
test_wide_nested_preallocated_write[stack_out-cpu-128-256] 32.9923ms 32.7000ms 30.5811 Ops/s 30.4856 Ops/s $\color{#35bf28}+0.31\%$
test_wide_nested_preallocated_write[stack_out-cpu-4096-128] 86.1483ms 85.7261ms 11.6651 Ops/s 11.7364 Ops/s $\color{#d91a1a}-0.61\%$
test_wide_nested_preallocated_write[stack_out-cpu-4096-256] 0.2066s 0.2054s 4.8676 Ops/s 5.0085 Ops/s $\color{#d91a1a}-2.81\%$
test_wide_nested_preallocated_write[stack_out-cuda-128-128] 14.0258ms 13.8609ms 72.1454 Ops/s 72.3238 Ops/s $\color{#d91a1a}-0.25\%$
test_wide_nested_preallocated_write[stack_out-cuda-128-256] 27.5933ms 26.9279ms 37.1362 Ops/s 37.0408 Ops/s $\color{#35bf28}+0.26\%$
test_wide_nested_preallocated_write[stack_out-cuda-4096-128] 14.0690ms 13.9426ms 71.7229 Ops/s 71.0450 Ops/s $\color{#35bf28}+0.95\%$
test_wide_nested_preallocated_write[stack_out-cuda-4096-256] 27.9829ms 27.3010ms 36.6287 Ops/s 36.4757 Ops/s $\color{#35bf28}+0.42\%$
test_wide_nested_preallocated_write[set_at_loop-cpu-128-128] 57.2902ms 56.5218ms 17.6923 Ops/s 19.2294 Ops/s $\textbf{\color{#d91a1a}-7.99\%}$
test_wide_nested_preallocated_write[set_at_loop-cpu-128-256] 0.1150s 0.1145s 8.7352 Ops/s 9.5980 Ops/s $\textbf{\color{#d91a1a}-8.99\%}$
test_wide_nested_preallocated_write[set_at_loop-cpu-4096-128] 0.1502s 0.1491s 6.7089 Ops/s 6.7686 Ops/s $\color{#d91a1a}-0.88\%$
test_wide_nested_preallocated_write[set_at_loop-cpu-4096-256] 0.3287s 0.3267s 3.0606 Ops/s 3.1266 Ops/s $\color{#d91a1a}-2.11\%$
test_wide_nested_preallocated_write[set_at_loop-cuda-128-128] 77.0629ms 76.6115ms 13.0529 Ops/s 13.5200 Ops/s $\color{#d91a1a}-3.46\%$
test_wide_nested_preallocated_write[set_at_loop-cuda-128-256] 0.1538s 0.1528s 6.5461 Ops/s 6.7457 Ops/s $\color{#d91a1a}-2.96\%$
test_wide_nested_preallocated_write[set_at_loop-cuda-4096-128] 77.6103ms 77.2306ms 12.9482 Ops/s 13.4071 Ops/s $\color{#d91a1a}-3.42\%$
test_wide_nested_preallocated_write[set_at_loop-cuda-4096-256] 0.1546s 0.1540s 6.4924 Ops/s 6.7067 Ops/s $\color{#d91a1a}-3.19\%$
test_wide_nested_preallocated_write[copy_at-cpu-128-128] 31.0394ms 30.8767ms 32.3869 Ops/s 33.1684 Ops/s $\color{#d91a1a}-2.36\%$
test_wide_nested_preallocated_write[copy_at-cpu-128-256] 62.9821ms 62.6630ms 15.9584 Ops/s 16.3614 Ops/s $\color{#d91a1a}-2.46\%$
test_wide_nested_preallocated_write[copy_at-cpu-4096-128] 0.1245s 0.1230s 8.1283 Ops/s 8.2844 Ops/s $\color{#d91a1a}-1.88\%$
test_wide_nested_preallocated_write[copy_at-cpu-4096-256] 0.2714s 0.2701s 3.7017 Ops/s 3.7426 Ops/s $\color{#d91a1a}-1.09\%$
test_wide_nested_preallocated_write[copy_at-cuda-128-128] 44.8687ms 44.7220ms 22.3603 Ops/s 22.9002 Ops/s $\color{#d91a1a}-2.36\%$
test_wide_nested_preallocated_write[copy_at-cuda-128-256] 89.8714ms 89.5575ms 11.1660 Ops/s 11.4825 Ops/s $\color{#d91a1a}-2.76\%$
test_wide_nested_preallocated_write[copy_at-cuda-4096-128] 45.3923ms 45.0639ms 22.1907 Ops/s 22.8111 Ops/s $\color{#d91a1a}-2.72\%$
test_wide_nested_preallocated_write[copy_at-cuda-4096-256] 90.6632ms 90.2170ms 11.0844 Ops/s 11.4097 Ops/s $\color{#d91a1a}-2.85\%$
test_wide_nested_preallocated_write[manual_leaf_copy-cpu-128-128] 21.7275ms 21.4284ms 46.6671 Ops/s 49.4983 Ops/s $\textbf{\color{#d91a1a}-5.72\%}$
test_wide_nested_preallocated_write[manual_leaf_copy-cpu-128-256] 44.1768ms 43.8940ms 22.7821 Ops/s 24.1307 Ops/s $\textbf{\color{#d91a1a}-5.59\%}$
test_wide_nested_preallocated_write[manual_leaf_copy-cpu-4096-128] 0.1169s 0.1131s 8.8448 Ops/s 9.3899 Ops/s $\textbf{\color{#d91a1a}-5.80\%}$
test_wide_nested_preallocated_write[manual_leaf_copy-cpu-4096-256] 0.2508s 0.2490s 4.0161 Ops/s 4.0894 Ops/s $\color{#d91a1a}-1.79\%$
test_wide_nested_preallocated_write[manual_leaf_copy-cuda-128-128] 39.1980ms 38.9784ms 25.6553 Ops/s 27.1266 Ops/s $\textbf{\color{#d91a1a}-5.42\%}$
test_wide_nested_preallocated_write[manual_leaf_copy-cuda-128-256] 78.5880ms 78.0244ms 12.8165 Ops/s 13.7224 Ops/s $\textbf{\color{#d91a1a}-6.60\%}$
test_wide_nested_preallocated_write[manual_leaf_copy-cuda-4096-128] 39.2728ms 38.9481ms 25.6752 Ops/s 27.2218 Ops/s $\textbf{\color{#d91a1a}-5.68\%}$
test_wide_nested_preallocated_write[manual_leaf_copy-cuda-4096-256] 78.3296ms 77.4563ms 12.9105 Ops/s 13.6276 Ops/s $\textbf{\color{#d91a1a}-5.26\%}$
test_plain_set_nested 43.4610μs 15.6210μs 64.0164 KOps/s 65.5028 KOps/s $\color{#d91a1a}-2.27\%$
test_plain_set_stack_nested 44.6210μs 15.4689μs 64.6460 KOps/s 65.7018 KOps/s $\color{#d91a1a}-1.61\%$
test_plain_set_nested_inplace 53.9710μs 17.1507μs 58.3065 KOps/s 58.6731 KOps/s $\color{#d91a1a}-0.62\%$
test_plain_set_stack_nested_inplace 40.3910μs 16.6524μs 60.0515 KOps/s 59.6245 KOps/s $\color{#35bf28}+0.72\%$
test_items 22.7500μs 6.0547μs 165.1603 KOps/s 165.7210 KOps/s $\color{#d91a1a}-0.34\%$
test_items_nested 0.5527ms 0.4623ms 2.1631 KOps/s 2.1546 KOps/s $\color{#35bf28}+0.40\%$
test_items_nested_locked 0.5153ms 0.4639ms 2.1555 KOps/s 2.1569 KOps/s $\color{#d91a1a}-0.07\%$
test_items_nested_leaf 0.1389ms 96.8024μs 10.3303 KOps/s 10.2960 KOps/s $\color{#35bf28}+0.33\%$
test_items_stack_nested 0.5113ms 0.4575ms 2.1857 KOps/s 2.1811 KOps/s $\color{#35bf28}+0.21\%$
test_items_stack_nested_leaf 0.1400ms 95.7430μs 10.4446 KOps/s 10.2897 KOps/s $\color{#35bf28}+1.51\%$
test_items_stack_nested_locked 0.5530ms 0.4629ms 2.1603 KOps/s 2.1586 KOps/s $\color{#35bf28}+0.08\%$
test_keys 25.7510μs 4.2666μs 234.3795 KOps/s 235.0767 KOps/s $\color{#d91a1a}-0.30\%$
test_keys_nested 0.1834ms 0.1309ms 7.6375 KOps/s 7.5862 KOps/s $\color{#35bf28}+0.68\%$
test_keys_nested_locked 0.7316ms 0.1391ms 7.1891 KOps/s 7.1467 KOps/s $\color{#35bf28}+0.59\%$
test_keys_nested_leaf 0.1635ms 0.1211ms 8.2552 KOps/s 8.2172 KOps/s $\color{#35bf28}+0.46\%$
test_keys_stack_nested 0.1761ms 0.1309ms 7.6386 KOps/s 7.6342 KOps/s $\color{#35bf28}+0.06\%$
test_keys_stack_nested_leaf 0.1738ms 0.1200ms 8.3360 KOps/s 8.2471 KOps/s $\color{#35bf28}+1.08\%$
test_keys_stack_nested_locked 0.1897ms 0.1391ms 7.1899 KOps/s 7.1937 KOps/s $\color{#d91a1a}-0.05\%$
test_values 5.9342μs 1.0335μs 967.5609 KOps/s 978.6690 KOps/s $\color{#d91a1a}-1.14\%$
test_values_nested 79.2820μs 52.3138μs 19.1154 KOps/s 18.9042 KOps/s $\color{#35bf28}+1.12\%$
test_values_nested_locked 83.7620μs 56.0338μs 17.8464 KOps/s 17.6989 KOps/s $\color{#35bf28}+0.83\%$
test_values_nested_leaf 88.2920μs 60.3431μs 16.5719 KOps/s 16.5360 KOps/s $\color{#35bf28}+0.22\%$
test_values_stack_nested 88.4820μs 52.5217μs 19.0397 KOps/s 19.0459 KOps/s $\color{#d91a1a}-0.03\%$
test_values_stack_nested_leaf 97.5430μs 59.7875μs 16.7259 KOps/s 16.6904 KOps/s $\color{#35bf28}+0.21\%$
test_values_stack_nested_locked 89.3020μs 56.4560μs 17.7129 KOps/s 17.7404 KOps/s $\color{#d91a1a}-0.15\%$
test_membership 5.4502μs 0.8467μs 1.1810 MOps/s 1.1807 MOps/s $\color{#35bf28}+0.03\%$
test_membership_nested 30.3610μs 3.0166μs 331.4936 KOps/s 329.2824 KOps/s $\color{#35bf28}+0.67\%$
test_membership_nested_leaf 33.7410μs 3.0111μs 332.1002 KOps/s 331.8637 KOps/s $\color{#35bf28}+0.07\%$
test_membership_stacked_nested 50.9810μs 2.8913μs 345.8636 KOps/s 342.5458 KOps/s $\color{#35bf28}+0.97\%$
test_membership_stacked_nested_leaf 29.6300μs 2.8860μs 346.4955 KOps/s 347.0495 KOps/s $\color{#d91a1a}-0.16\%$
test_membership_nested_last 47.0410μs 4.4724μs 223.5914 KOps/s 221.2125 KOps/s $\color{#35bf28}+1.08\%$
test_membership_nested_leaf_last 47.3210μs 4.4631μs 224.0599 KOps/s 221.9050 KOps/s $\color{#35bf28}+0.97\%$
test_membership_stacked_nested_last 43.1110μs 4.3285μs 231.0252 KOps/s 227.8101 KOps/s $\color{#35bf28}+1.41\%$
test_membership_stacked_nested_leaf_last 45.2910μs 4.3317μs 230.8587 KOps/s 228.3920 KOps/s $\color{#35bf28}+1.08\%$
test_nested_getleaf 51.1810μs 21.6355μs 46.2204 KOps/s 45.6793 KOps/s $\color{#35bf28}+1.18\%$
test_nested_get 52.9510μs 20.4361μs 48.9329 KOps/s 48.5297 KOps/s $\color{#35bf28}+0.83\%$
test_stacked_getleaf 49.5110μs 21.6456μs 46.1988 KOps/s 46.9825 KOps/s $\color{#d91a1a}-1.67\%$
test_stacked_get 61.1810μs 20.3944μs 49.0331 KOps/s 49.0188 KOps/s $\color{#35bf28}+0.03\%$
test_nested_getitemleaf 44.4110μs 22.1867μs 45.0720 KOps/s 45.6141 KOps/s $\color{#d91a1a}-1.19\%$
test_nested_getitem 47.2710μs 21.1057μs 47.3805 KOps/s 47.1400 KOps/s $\color{#35bf28}+0.51\%$
test_stacked_getitemleaf 49.2010μs 22.1506μs 45.1456 KOps/s 45.2886 KOps/s $\color{#d91a1a}-0.32\%$
test_stacked_getitem 51.0410μs 21.0368μs 47.5357 KOps/s 47.6055 KOps/s $\color{#d91a1a}-0.15\%$
test_lock_nested 4.9734ms 0.5188ms 1.9277 KOps/s 1.9750 KOps/s $\color{#d91a1a}-2.40\%$
test_lock_stack_nested 0.5835ms 0.5182ms 1.9296 KOps/s 1.9757 KOps/s $\color{#d91a1a}-2.33\%$
test_unlock_nested 0.4870ms 0.4056ms 2.4657 KOps/s 2.5355 KOps/s $\color{#d91a1a}-2.75\%$
test_unlock_stack_nested 0.5802ms 0.4130ms 2.4211 KOps/s 2.5119 KOps/s $\color{#d91a1a}-3.62\%$
test_flatten_speed 0.2009ms 0.1216ms 8.2261 KOps/s 8.1508 KOps/s $\color{#35bf28}+0.92\%$
test_unflatten_speed 0.6494ms 0.5721ms 1.7478 KOps/s 1.7310 KOps/s $\color{#35bf28}+0.97\%$
test_common_ops 0.1001s 1.7727ms 564.1104 Ops/s 765.8977 Ops/s $\textbf{\color{#d91a1a}-26.35\%}$
test_creation 75.2210μs 3.2075μs 311.7692 KOps/s 314.6896 KOps/s $\color{#d91a1a}-0.93\%$
test_creation_empty 44.4610μs 7.0951μs 140.9428 KOps/s 142.6817 KOps/s $\color{#d91a1a}-1.22\%$
test_creation_nested_1 45.7010μs 11.5958μs 86.2383 KOps/s 85.4067 KOps/s $\color{#35bf28}+0.97\%$
test_creation_nested_2 42.9110μs 13.6904μs 73.0439 KOps/s 72.6214 KOps/s $\color{#35bf28}+0.58\%$
test_creation_many_keys[10] 51.1010μs 21.4027μs 46.7231 KOps/s 46.8702 KOps/s $\color{#d91a1a}-0.31\%$
test_creation_many_keys[50] 0.1235ms 92.8101μs 10.7747 KOps/s 10.9605 KOps/s $\color{#d91a1a}-1.70\%$
test_creation_many_keys[100] 0.2117ms 0.1801ms 5.5528 KOps/s 5.6054 KOps/s $\color{#d91a1a}-0.94\%$
test_creation_nested_many_keys[10] 74.3020μs 46.1315μs 21.6772 KOps/s 21.8901 KOps/s $\color{#d91a1a}-0.97\%$
test_creation_nested_many_keys[50] 0.2588ms 0.1891ms 5.2872 KOps/s 5.3792 KOps/s $\color{#d91a1a}-1.71\%$
test_clone 40.7710μs 13.7033μs 72.9751 KOps/s 74.9485 KOps/s $\color{#d91a1a}-2.63\%$
test_getitem[int] 1.4552ms 15.7525μs 63.4821 KOps/s 66.3403 KOps/s $\color{#d91a1a}-4.31\%$
test_getitem[slice_int] 0.1446ms 25.4644μs 39.2705 KOps/s 40.7979 KOps/s $\color{#d91a1a}-3.74\%$
test_getitem[range] 0.1862ms 65.2221μs 15.3322 KOps/s 15.5972 KOps/s $\color{#d91a1a}-1.70\%$
test_getitem[tuple] 0.1471ms 25.2572μs 39.5926 KOps/s 41.4162 KOps/s $\color{#d91a1a}-4.40\%$
test_getitem[list] 0.1860ms 59.3651μs 16.8449 KOps/s 17.0912 KOps/s $\color{#d91a1a}-1.44\%$
test_setitem_dim[int] 49.6810μs 26.9539μs 37.1004 KOps/s 39.2083 KOps/s $\textbf{\color{#d91a1a}-5.38\%}$
test_setitem_dim[slice_int] 70.6420μs 43.0412μs 23.2335 KOps/s 22.5837 KOps/s $\color{#35bf28}+2.88\%$
test_setitem_dim[range] 0.1217ms 95.5970μs 10.4606 KOps/s 10.4038 KOps/s $\color{#35bf28}+0.55\%$
test_setitem_dim[tuple] 64.0910μs 39.7865μs 25.1342 KOps/s 25.8119 KOps/s $\color{#d91a1a}-2.63\%$
test_setitem 57.0510μs 18.4020μs 54.3418 KOps/s 56.6210 KOps/s $\color{#d91a1a}-4.03\%$
test_set 51.3610μs 17.5738μs 56.9028 KOps/s 59.3762 KOps/s $\color{#d91a1a}-4.17\%$
test_set_shared 0.4979ms 0.2113ms 4.7323 KOps/s 4.7486 KOps/s $\color{#d91a1a}-0.34\%$
test_update 0.2109ms 22.8815μs 43.7035 KOps/s 45.2874 KOps/s $\color{#d91a1a}-3.50\%$
test_update_nested 68.9110μs 34.7144μs 28.8065 KOps/s 29.8885 KOps/s $\color{#d91a1a}-3.62\%$
test_update__nested 0.5403ms 34.0135μs 29.4001 KOps/s 30.4314 KOps/s $\color{#d91a1a}-3.39\%$
test_set_nested 48.5310μs 19.5993μs 51.0221 KOps/s 52.9128 KOps/s $\color{#d91a1a}-3.57\%$
test_set_nested_new 51.7110μs 24.5528μs 40.7285 KOps/s 42.1721 KOps/s $\color{#d91a1a}-3.42\%$
test_select 76.7720μs 41.5529μs 24.0657 KOps/s 24.8812 KOps/s $\color{#d91a1a}-3.28\%$
test_select_nested 0.1123ms 74.7807μs 13.3724 KOps/s 13.3652 KOps/s $\color{#35bf28}+0.05\%$
test_exclude_nested 0.1226ms 92.6398μs 10.7945 KOps/s 10.8500 KOps/s $\color{#d91a1a}-0.51\%$
test_empty[True] 0.5386ms 0.3958ms 2.5265 KOps/s 2.4774 KOps/s $\color{#35bf28}+1.98\%$
test_empty[False] 7.0700μs 1.3936μs 717.5639 KOps/s 707.9276 KOps/s $\color{#35bf28}+1.36\%$
test_to 0.1163ms 74.9650μs 13.3396 KOps/s 14.1798 KOps/s $\textbf{\color{#d91a1a}-5.93\%}$
test_to_nonblocking 0.1172ms 74.2035μs 13.4765 KOps/s 14.5337 KOps/s $\textbf{\color{#d91a1a}-7.27\%}$
test_unbind_speed 2.0293ms 0.3686ms 2.7130 KOps/s 2.9532 KOps/s $\textbf{\color{#d91a1a}-8.13\%}$
test_unbind_speed_stack0 0.4302ms 0.3614ms 2.7672 KOps/s 3.0114 KOps/s $\textbf{\color{#d91a1a}-8.11\%}$
test_unbind_speed_stack1 0.1132s 0.8993ms 1.1120 KOps/s 1.1570 KOps/s $\color{#d91a1a}-3.89\%$
test_split 1.3038ms 1.1490ms 870.3134 Ops/s 715.2796 Ops/s $\textbf{\color{#35bf28}+21.67\%}$
test_chunk 0.1132s 1.2396ms 806.7135 Ops/s 912.7527 Ops/s $\textbf{\color{#d91a1a}-11.62\%}$
test_to_cpu_blocking 20.6838ms 20.4647ms 48.8647 Ops/s 49.5260 Ops/s $\color{#d91a1a}-1.34\%$
test_to_cpu_global_sync 12.1417ms 12.0078ms 83.2791 Ops/s 84.6722 Ops/s $\color{#d91a1a}-1.65\%$
test_to_cpu_event_sync 13.4184ms 13.0485ms 76.6371 Ops/s 68.9385 Ops/s $\textbf{\color{#35bf28}+11.17\%}$
test_to_cpu_default 14.3282ms 13.0688ms 76.5180 Ops/s 77.7143 Ops/s $\color{#d91a1a}-1.54\%$
test_consolidate[False-None] 4.3740ms 4.2472ms 235.4484 Ops/s 236.2498 Ops/s $\color{#d91a1a}-0.34\%$
test_consolidate[default-None] 3.6932ms 2.2471ms 445.0160 Ops/s 454.2560 Ops/s $\color{#d91a1a}-2.03\%$
test_consolidate[reduce-overhead-None] 3.7168ms 2.1641ms 462.0922 Ops/s 471.5849 Ops/s $\color{#d91a1a}-2.01\%$
test_consolidate_njt[False-None] 8.7142ms 8.6401ms 115.7391 Ops/s 114.9676 Ops/s $\color{#35bf28}+0.67\%$
test_to[False-False-None] 2.2939ms 2.2118ms 452.1275 Ops/s 452.2079 Ops/s $\color{#d91a1a}-0.02\%$
test_to[True-False-None] 2.2819ms 1.9745ms 506.4570 Ops/s 508.6057 Ops/s $\color{#d91a1a}-0.42\%$
test_to[within-False-None] 6.4275ms 6.3448ms 157.6106 Ops/s 158.4591 Ops/s $\color{#d91a1a}-0.54\%$
test_to[True-default-None] 11.2260ms 10.8531ms 92.1396 Ops/s 92.7769 Ops/s $\color{#d91a1a}-0.69\%$
test_to_njt[False-False-None] 8.8184ms 8.7207ms 114.6697 Ops/s 113.3964 Ops/s $\color{#35bf28}+1.12\%$
test_to_njt[True-False-None] 7.1280ms 6.9794ms 143.2797 Ops/s 139.3531 Ops/s $\color{#35bf28}+2.82\%$
test_to_njt[within-False-None] 15.9879ms 15.6048ms 64.0828 Ops/s 62.9889 Ops/s $\color{#35bf28}+1.74\%$
test_creation[device0] 0.3894ms 0.1191ms 8.3932 KOps/s 8.3076 KOps/s $\color{#35bf28}+1.03\%$
test_creation_from_tensor 0.5065ms 0.1153ms 8.6714 KOps/s 8.4131 KOps/s $\color{#35bf28}+3.07\%$
test_add_one[memmap_tensor0] 0.2696ms 6.8973μs 144.9846 KOps/s 151.1611 KOps/s $\color{#d91a1a}-4.09\%$
test_contiguous[memmap_tensor0] 29.2710μs 0.6737μs 1.4843 MOps/s 2.2306 MOps/s $\textbf{\color{#d91a1a}-33.46\%}$
test_stack[memmap_tensor0] 96.7120μs 4.4913μs 222.6547 KOps/s 239.2572 KOps/s $\textbf{\color{#d91a1a}-6.94\%}$
test_memmaptd_index 1.4700ms 0.2991ms 3.3437 KOps/s 3.6695 KOps/s $\textbf{\color{#d91a1a}-8.88\%}$
test_memmaptd_index_astensor 0.5541ms 0.4002ms 2.4989 KOps/s 2.6512 KOps/s $\textbf{\color{#d91a1a}-5.74\%}$
test_memmaptd_index_op 0.8147ms 0.6683ms 1.4963 KOps/s 1.5761 KOps/s $\textbf{\color{#d91a1a}-5.06\%}$
test_serialize_model 0.1392s 0.1351s 7.4017 Ops/s 7.4122 Ops/s $\color{#d91a1a}-0.14\%$
test_serialize_model_pickle 1.3609s 1.2126s 0.8247 Ops/s 0.8231 Ops/s $\color{#35bf28}+0.19\%$
test_serialize_weights 0.1365s 0.1339s 7.4685 Ops/s 7.4360 Ops/s $\color{#35bf28}+0.44\%$
test_serialize_weights_returnearly 0.4798s 90.4931ms 11.0506 Ops/s 10.7717 Ops/s $\color{#35bf28}+2.59\%$
test_serialize_weights_pickle 1.3801s 1.2149s 0.8231 Ops/s 0.8218 Ops/s $\color{#35bf28}+0.16\%$
test_reshape_pytree 0.2053ms 33.3351μs 29.9984 KOps/s 30.6182 KOps/s $\color{#d91a1a}-2.02\%$
test_reshape_td 79.4410μs 47.3980μs 21.0979 KOps/s 22.2313 KOps/s $\textbf{\color{#d91a1a}-5.10\%}$
test_view_pytree 0.2200ms 32.9148μs 30.3815 KOps/s 30.7772 KOps/s $\color{#d91a1a}-1.29\%$
test_view_td 0.1086ms 55.8434μs 17.9072 KOps/s 18.6818 KOps/s $\color{#d91a1a}-4.15\%$
test_unbind_pytree 0.2292ms 37.6836μs 26.5367 KOps/s 26.8078 KOps/s $\color{#d91a1a}-1.01\%$
test_unbind_td 76.1320μs 53.8312μs 18.5766 KOps/s 19.9276 KOps/s $\textbf{\color{#d91a1a}-6.78\%}$
test_split_pytree 0.2154ms 42.5832μs 23.4834 KOps/s 23.2704 KOps/s $\color{#35bf28}+0.92\%$
test_split_td 0.1938ms 65.6856μs 15.2240 KOps/s 15.4748 KOps/s $\color{#d91a1a}-1.62\%$
test_add_pytree 0.2317ms 42.8518μs 23.3363 KOps/s 23.6118 KOps/s $\color{#d91a1a}-1.17\%$
test_add_td 0.1019ms 54.9635μs 18.1939 KOps/s 18.4510 KOps/s $\color{#d91a1a}-1.39\%$
test_compile_add_one_nested[tensordict-compile] 0.3252ms 0.2427ms 4.1208 KOps/s 4.2505 KOps/s $\color{#d91a1a}-3.05\%$
test_compile_add_one_nested[tensordict-eager] 0.2903ms 0.2008ms 4.9813 KOps/s 5.0108 KOps/s $\color{#d91a1a}-0.59\%$
test_compile_add_one_nested[pytree-compile] 0.2755ms 0.1353ms 7.3936 KOps/s 7.0097 KOps/s $\textbf{\color{#35bf28}+5.48\%}$
test_compile_add_one_nested[pytree-eager] 0.4879ms 0.1854ms 5.3926 KOps/s 5.5457 KOps/s $\color{#d91a1a}-2.76\%$
test_compile_copy_nested[tensordict-compile] 84.6820μs 17.3669μs 57.5807 KOps/s 66.3231 KOps/s $\textbf{\color{#d91a1a}-13.18\%}$
test_compile_copy_nested[tensordict-eager] 0.4488ms 53.4043μs 18.7251 KOps/s 18.1704 KOps/s $\color{#35bf28}+3.05\%$
test_compile_copy_nested[pytree-compile] 0.1468ms 16.0626μs 62.2565 KOps/s 63.4588 KOps/s $\color{#d91a1a}-1.89\%$
test_compile_copy_nested[pytree-eager] 0.3864ms 67.3883μs 14.8394 KOps/s 14.4395 KOps/s $\color{#35bf28}+2.77\%$
test_compile_add_one_flat[tensordict-compile] 0.2966ms 0.2516ms 3.9743 KOps/s 3.7555 KOps/s $\textbf{\color{#35bf28}+5.83\%}$
test_compile_add_one_flat[tensordict-eager] 0.3172ms 0.2841ms 3.5200 KOps/s 3.4691 KOps/s $\color{#35bf28}+1.47\%$
test_compile_add_one_flat[tensorclass-compile] 0.2445ms 0.1866ms 5.3582 KOps/s 5.1833 KOps/s $\color{#35bf28}+3.38\%$
test_compile_add_one_flat[tensorclass-eager] 0.1118ms 75.3428μs 13.2727 KOps/s 13.1845 KOps/s $\color{#35bf28}+0.67\%$
test_compile_add_one_flat[pytree-compile] 0.4880ms 0.1937ms 5.1631 KOps/s 4.9618 KOps/s $\color{#35bf28}+4.06\%$
test_compile_add_one_flat[pytree-eager] 0.8189ms 0.5508ms 1.8154 KOps/s 1.9026 KOps/s $\color{#d91a1a}-4.58\%$
test_compile_add_self_flat[tensordict-eager] 0.4195ms 0.3322ms 3.0106 KOps/s 2.8865 KOps/s $\color{#35bf28}+4.30\%$
test_compile_add_self_flat[tensordict-compile] 0.3142ms 0.2533ms 3.9472 KOps/s 3.6044 KOps/s $\textbf{\color{#35bf28}+9.51\%}$
test_compile_add_self_flat[tensorclass-eager] 0.1411ms 91.9055μs 10.8807 KOps/s 10.7814 KOps/s $\color{#35bf28}+0.92\%$
test_compile_add_self_flat[tensorclass-compile] 0.3713ms 0.1886ms 5.3010 KOps/s 4.8992 KOps/s $\textbf{\color{#35bf28}+8.20\%}$
test_compile_add_self_flat[pytree-eager] 0.7004ms 0.4553ms 2.1962 KOps/s 2.2651 KOps/s $\color{#d91a1a}-3.04\%$
test_compile_add_self_flat[pytree-compile] 0.3645ms 0.1945ms 5.1403 KOps/s 4.8373 KOps/s $\textbf{\color{#35bf28}+6.26\%}$
test_compile_copy_flat[tensordict-compile] 46.1810μs 20.3901μs 49.0434 KOps/s 48.9827 KOps/s $\color{#35bf28}+0.12\%$
test_compile_copy_flat[tensordict-eager] 72.5710μs 41.6029μs 24.0368 KOps/s 24.3842 KOps/s $\color{#d91a1a}-1.42\%$
test_compile_copy_flat[pytree-compile] 0.1837ms 22.1751μs 45.0956 KOps/s 44.7269 KOps/s $\color{#35bf28}+0.82\%$
test_compile_copy_flat[pytree-eager] 0.3371ms 52.7080μs 18.9725 KOps/s 19.1070 KOps/s $\color{#d91a1a}-0.70\%$
test_compile_assign_and_add[tensordict-compile] 2.2050ms 0.2404ms 4.1593 KOps/s 3.7473 KOps/s $\textbf{\color{#35bf28}+10.99\%}$
test_compile_assign_and_add[tensordict-eager] 3.5106ms 3.3733ms 296.4464 Ops/s 301.3659 Ops/s $\color{#d91a1a}-1.63\%$
test_compile_assign_and_add[pytree-compile] 2.1050ms 0.1817ms 5.5050 KOps/s 5.4095 KOps/s $\color{#35bf28}+1.76\%$
test_compile_assign_and_add[pytree-eager] 3.1268ms 2.9197ms 342.4951 Ops/s 359.9428 Ops/s $\color{#d91a1a}-4.85\%$
test_compile_indexing[tensor-tensordict-compile] 0.2936ms 0.1360ms 7.3509 KOps/s 7.2109 KOps/s $\color{#35bf28}+1.94\%$
test_compile_indexing[tensor-tensordict-eager] 0.2193s 95.7205μs 10.4471 KOps/s 12.8241 KOps/s $\textbf{\color{#d91a1a}-18.54\%}$
test_compile_indexing[tensor-tensorclass-compile] 0.2699ms 0.1201ms 8.3264 KOps/s 8.2804 KOps/s $\color{#35bf28}+0.56\%$
test_compile_indexing[tensor-tensorclass-eager] 0.3048ms 46.3286μs 21.5849 KOps/s 20.5674 KOps/s $\color{#35bf28}+4.95\%$
test_compile_indexing[tensor-pytree-compile] 0.2921ms 0.1200ms 8.3350 KOps/s 8.0816 KOps/s $\color{#35bf28}+3.14\%$
test_compile_indexing[tensor-pytree-eager] 0.2468ms 46.8230μs 21.3570 KOps/s 20.4408 KOps/s $\color{#35bf28}+4.48\%$
test_compile_indexing[slice-tensordict-compile] 0.2653ms 85.6045μs 11.6816 KOps/s 11.6786 KOps/s $\color{#35bf28}+0.03\%$
test_compile_indexing[slice-tensordict-eager] 0.2233ms 28.6180μs 34.9430 KOps/s 36.2187 KOps/s $\color{#d91a1a}-3.52\%$
test_compile_indexing[slice-tensorclass-compile] 0.2033ms 71.5159μs 13.9829 KOps/s 13.9738 KOps/s $\color{#35bf28}+0.07\%$
test_compile_indexing[slice-tensorclass-eager] 0.2615ms 22.7371μs 43.9809 KOps/s 44.3062 KOps/s $\color{#d91a1a}-0.73\%$
test_compile_indexing[slice-pytree-compile] 0.1076ms 72.3404μs 13.8235 KOps/s 14.0182 KOps/s $\color{#d91a1a}-1.39\%$
test_compile_indexing[slice-pytree-eager] 0.2452ms 22.6678μs 44.1154 KOps/s 44.4336 KOps/s $\color{#d91a1a}-0.72\%$
test_compile_indexing[int-tensordict-compile] 0.1292ms 85.7209μs 11.6658 KOps/s 11.5406 KOps/s $\color{#35bf28}+1.08\%$
test_compile_indexing[int-tensordict-eager] 0.2290ms 28.3080μs 35.3257 KOps/s 35.7137 KOps/s $\color{#d91a1a}-1.09\%$
test_compile_indexing[int-tensorclass-compile] 0.1177ms 72.2555μs 13.8398 KOps/s 14.0118 KOps/s $\color{#d91a1a}-1.23\%$
test_compile_indexing[int-tensorclass-eager] 0.2526ms 22.9006μs 43.6671 KOps/s 44.3550 KOps/s $\color{#d91a1a}-1.55\%$
test_compile_indexing[int-pytree-compile] 0.1169ms 72.6375μs 13.7670 KOps/s 13.9188 KOps/s $\color{#d91a1a}-1.09\%$
test_compile_indexing[int-pytree-eager] 0.2784ms 22.7319μs 43.9911 KOps/s 44.4790 KOps/s $\color{#d91a1a}-1.10\%$
test_compile_replace[single-eager] 99.3820μs 49.7808μs 20.0881 KOps/s 20.2444 KOps/s $\color{#d91a1a}-0.77\%$
test_compile_replace[single-compile] 0.2113ms 0.1310ms 7.6348 KOps/s 7.5280 KOps/s $\color{#35bf28}+1.42\%$
test_compile_replace[multi-eager] 0.6852ms 0.5751ms 1.7388 KOps/s 1.7677 KOps/s $\color{#d91a1a}-1.63\%$
test_compile_replace[multi-compile] 0.3289ms 0.1395ms 7.1710 KOps/s 7.1555 KOps/s $\color{#35bf28}+0.22\%$
test_compile_tc_getattr_20[eager] 0.2395ms 0.1803ms 5.5473 KOps/s 6.0389 KOps/s $\textbf{\color{#d91a1a}-8.14\%}$
test_compile_tc_getattr_20[compile] 0.3939ms 0.1488ms 6.7209 KOps/s 6.4913 KOps/s $\color{#35bf28}+3.54\%$
test_compile_clone_shallow[20-eager] 43.2410μs 19.4482μs 51.4187 KOps/s 51.8500 KOps/s $\color{#d91a1a}-0.83\%$
test_compile_clone_shallow[20-compile] 74.0410μs 17.4477μs 57.3143 KOps/s 58.7254 KOps/s $\color{#d91a1a}-2.40\%$
test_compile_clone_shallow[40-eager] 67.5320μs 34.1752μs 29.2610 KOps/s 29.1243 KOps/s $\color{#35bf28}+0.47\%$
test_compile_clone_shallow[40-compile] 83.2620μs 18.8176μs 53.1416 KOps/s 53.5741 KOps/s $\color{#d91a1a}-0.81\%$
test_compile_clone_shallow[80-eager] 97.6520μs 63.2663μs 15.8062 KOps/s 15.9084 KOps/s $\color{#d91a1a}-0.64\%$
test_compile_clone_shallow[80-compile] 76.2220μs 22.5170μs 44.4109 KOps/s 44.7681 KOps/s $\color{#d91a1a}-0.80\%$
test_compile_update_inplace[eager] 0.2402ms 62.1136μs 16.0995 KOps/s 16.4107 KOps/s $\color{#d91a1a}-1.90\%$
test_compile_update_inplace[compile] 0.2471ms 0.1477ms 6.7720 KOps/s 6.6546 KOps/s $\color{#35bf28}+1.76\%$
test_mod_add[eager] 0.2891ms 0.2289ms 4.3679 KOps/s 4.4238 KOps/s $\color{#d91a1a}-1.26\%$
test_mod_add[compile] 0.3425ms 0.1363ms 7.3382 KOps/s 7.4126 KOps/s $\color{#d91a1a}-1.00\%$
test_mod_add[compile-overhead] 0.3439ms 0.1795ms 5.5707 KOps/s 5.5254 KOps/s $\color{#35bf28}+0.82\%$
test_mod_wrap[eager] 0.5784ms 0.4740ms 2.1098 KOps/s 2.0710 KOps/s $\color{#35bf28}+1.87\%$
test_mod_wrap[compile] 0.5041ms 0.4057ms 2.4652 KOps/s 2.4851 KOps/s $\color{#d91a1a}-0.80\%$
test_mod_wrap[compile-overhead] 8.5485ms 4.6751ms 213.8984 Ops/s 214.9685 Ops/s $\color{#d91a1a}-0.50\%$
test_mod_wrap_and_backward[eager] 1.8646ms 1.7258ms 579.4394 Ops/s 588.1434 Ops/s $\color{#d91a1a}-1.48\%$
test_mod_wrap_and_backward[compile] 2.5456ms 1.6401ms 609.7159 Ops/s 644.1992 Ops/s $\textbf{\color{#d91a1a}-5.35\%}$
test_mod_wrap_and_backward[compile-overhead] 1.3440ms 0.9506ms 1.0520 KOps/s 1.0451 KOps/s $\color{#35bf28}+0.66\%$
test_seq_add[eager] 0.5836ms 0.5107ms 1.9579 KOps/s 1.9451 KOps/s $\color{#35bf28}+0.66\%$
test_seq_add[compile] 0.2734ms 0.1430ms 6.9933 KOps/s 6.7995 KOps/s $\color{#35bf28}+2.85\%$
test_seq_add[compile-overhead] 0.2477ms 0.1838ms 5.4397 KOps/s 5.2712 KOps/s $\color{#35bf28}+3.20\%$
test_seq_wrap[eager] 1.5289ms 1.4267ms 700.9221 Ops/s 711.6621 Ops/s $\color{#d91a1a}-1.51\%$
test_seq_wrap[compile] 0.5000ms 0.4161ms 2.4031 KOps/s 2.3825 KOps/s $\color{#35bf28}+0.86\%$
test_seq_wrap[compile-overhead] 0.4267ms 0.3061ms 3.2665 KOps/s 3.2241 KOps/s $\color{#35bf28}+1.32\%$
test_func_call_runtime[False-eager] 0.8990ms 0.8140ms 1.2285 KOps/s 1.2513 KOps/s $\color{#d91a1a}-1.83\%$
test_func_call_runtime[False-compile] 1.1258ms 0.9964ms 1.0037 KOps/s 1.0138 KOps/s $\color{#d91a1a}-1.00\%$
test_func_call_runtime[False-compile-overhead] 0.6065ms 0.5166ms 1.9358 KOps/s 1.9439 KOps/s $\color{#d91a1a}-0.42\%$
test_func_call_runtime[True-eager] 1.1864ms 1.0778ms 927.7895 Ops/s 928.8850 Ops/s $\color{#d91a1a}-0.12\%$
test_func_call_runtime[True-compile] 1.1006ms 1.0222ms 978.2797 Ops/s 993.6474 Ops/s $\color{#d91a1a}-1.55\%$
test_func_call_runtime[True-compile-overhead] 0.6638ms 0.5294ms 1.8888 KOps/s 1.8614 KOps/s $\color{#35bf28}+1.47\%$
test_func_call_cm_runtime[False-eager] 0.9404ms 0.8716ms 1.1474 KOps/s 1.2519 KOps/s $\textbf{\color{#d91a1a}-8.35\%}$
test_func_call_cm_runtime[False-compile] 1.1193ms 0.9851ms 1.0151 KOps/s 1.0326 KOps/s $\color{#d91a1a}-1.70\%$
test_func_call_cm_runtime[False-compile-overhead] 0.6888ms 0.4964ms 2.0145 KOps/s 1.9908 KOps/s $\color{#35bf28}+1.19\%$
test_func_call_cm_runtime[True-eager] 1.3670ms 1.2219ms 818.4057 Ops/s 820.0473 Ops/s $\color{#d91a1a}-0.20\%$
test_func_call_cm_runtime[True-compile] 1.1200ms 1.0411ms 960.5664 Ops/s 971.5938 Ops/s $\color{#d91a1a}-1.13\%$
test_func_call_cm_runtime[True-compile-overhead] 0.6270ms 0.5547ms 1.8029 KOps/s 1.7766 KOps/s $\color{#35bf28}+1.48\%$
test_vmap_func_call_cm_runtime[eager] 2.9217ms 2.4369ms 410.3594 Ops/s 412.4851 Ops/s $\color{#d91a1a}-0.52\%$
test_vmap_func_call_cm_runtime[compile] 3.1136ms 2.6157ms 382.3094 Ops/s 383.4911 Ops/s $\color{#d91a1a}-0.31\%$
test_vmap_func_call_cm_runtime[compile-overhead] 10.4496ms 2.6329ms 379.8065 Ops/s 383.1714 Ops/s $\color{#d91a1a}-0.88\%$
test_distributed 0.7139ms 0.1538ms 6.5038 KOps/s 6.5532 KOps/s $\color{#d91a1a}-0.75\%$
test_tdmodule 0.4251ms 0.1924ms 5.1968 KOps/s 5.1337 KOps/s $\color{#35bf28}+1.23\%$
test_tdmodule_dispatch 0.4284ms 0.3738ms 2.6756 KOps/s 2.6791 KOps/s $\color{#d91a1a}-0.13\%$
test_tdseq 49.0010μs 26.8888μs 37.1901 KOps/s 37.1935 KOps/s $-0.01\%$
test_tdseq_dispatch 0.2579ms 0.2185ms 4.5763 KOps/s 4.6026 KOps/s $\color{#d91a1a}-0.57\%$
test_instantiation_functorch 2.1739ms 2.0874ms 479.0718 Ops/s 482.5644 Ops/s $\color{#d91a1a}-0.72\%$
test_exec_functorch 0.2297ms 0.1860ms 5.3772 KOps/s 5.4926 KOps/s $\color{#d91a1a}-2.10\%$
test_exec_functional_call 0.2280ms 0.1639ms 6.1001 KOps/s 6.2991 KOps/s $\color{#d91a1a}-3.16\%$
test_exec_td_decorator 0.5084ms 0.2534ms 3.9465 KOps/s 4.0078 KOps/s $\color{#d91a1a}-1.53\%$
test_vmap_mlp_speed_decorator[True-True] 1.2710ms 1.0478ms 954.4030 Ops/s 957.9544 Ops/s $\color{#d91a1a}-0.37\%$
test_vmap_mlp_speed_decorator[True-False] 1.2918ms 1.0470ms 955.1039 Ops/s 961.1748 Ops/s $\color{#d91a1a}-0.63\%$
test_vmap_mlp_speed_decorator[False-True] 1.0450ms 0.7514ms 1.3309 KOps/s 1.3516 KOps/s $\color{#d91a1a}-1.54\%$
test_vmap_mlp_speed_decorator[False-False] 0.9364ms 0.7464ms 1.3397 KOps/s 1.3498 KOps/s $\color{#d91a1a}-0.75\%$
test_vmap_transformer_speed_decorator[True-True] 22.2760ms 21.4692ms 46.5784 Ops/s 47.1116 Ops/s $\color{#d91a1a}-1.13\%$
test_vmap_transformer_speed_decorator[True-False] 22.0953ms 21.4902ms 46.5327 Ops/s 47.1548 Ops/s $\color{#d91a1a}-1.32\%$
test_vmap_transformer_speed_decorator[False-True] 21.8187ms 21.0149ms 47.5852 Ops/s 48.0147 Ops/s $\color{#d91a1a}-0.89\%$
test_vmap_transformer_speed_decorator[False-False] 21.8121ms 21.0337ms 47.5427 Ops/s 48.0220 Ops/s $\color{#d91a1a}-1.00\%$
test_to_module_speed[True] 2.2691ms 1.6382ms 610.4103 Ops/s 609.3993 Ops/s $\color{#35bf28}+0.17\%$
test_to_module_speed[False] 2.1669ms 1.6085ms 621.6854 Ops/s 614.4869 Ops/s $\color{#35bf28}+1.17\%$
test_to_module_plain_tensor_speed[None] 2.3042ms 1.7134ms 583.6244 Ops/s 571.0762 Ops/s $\color{#35bf28}+2.20\%$
test_to_module_plain_tensor_speed[False] 2.2107ms 1.5670ms 638.1613 Ops/s 628.1989 Ops/s $\color{#35bf28}+1.59\%$
test_to_module_plain_tensor_speed[True] 3.5127ms 3.4103ms 293.2288 Ops/s 286.3760 Ops/s $\color{#35bf28}+2.39\%$
test_tc_init 71.2910μs 43.4003μs 23.0413 KOps/s 22.9803 KOps/s $\color{#35bf28}+0.27\%$
test_tc_init_tensor_only 32.5310μs 9.0106μs 110.9808 KOps/s 107.2373 KOps/s $\color{#35bf28}+3.49\%$
test_tc_init_nested 0.1742ms 86.2286μs 11.5971 KOps/s 11.3749 KOps/s $\color{#35bf28}+1.95\%$
test_tc_init_many_fields 41.8910μs 14.7986μs 67.5739 KOps/s 65.5237 KOps/s $\color{#35bf28}+3.13\%$
test_tc_first_layer_tensor 19.7600μs 1.8554μs 538.9764 KOps/s 543.3154 KOps/s $\color{#d91a1a}-0.80\%$
test_tc_first_layer_tensor_only 2.8301μs 0.3938μs 2.5392 MOps/s 2.5600 MOps/s $\color{#d91a1a}-0.81\%$
test_tc_first_layer_tensor_set 22.6400μs 3.9587μs 252.6052 KOps/s 253.6021 KOps/s $\color{#d91a1a}-0.39\%$
test_tc_first_layer_tensor_only_set 25.5210μs 3.3112μs 302.0041 KOps/s 303.9552 KOps/s $\color{#d91a1a}-0.64\%$
test_tc_first_layer_nontensor 28.2500μs 6.1528μs 162.5288 KOps/s 159.8711 KOps/s $\color{#35bf28}+1.66\%$
test_tc_second_layer_tensor 32.5610μs 4.4312μs 225.6746 KOps/s 223.8779 KOps/s $\color{#35bf28}+0.80\%$
test_tc_second_layer_nontensor 30.3800μs 8.7344μs 114.4903 KOps/s 114.8563 KOps/s $\color{#d91a1a}-0.32\%$
test_unbind 0.2756s 13.7710ms 72.6164 Ops/s 65.1512 Ops/s $\textbf{\color{#35bf28}+11.46\%}$
test_full_like 4.4862ms 3.9856ms 250.9010 Ops/s 228.4718 Ops/s $\textbf{\color{#35bf28}+9.82\%}$
test_zeros_like 4.5027ms 4.3793ms 228.3445 Ops/s 229.1404 Ops/s $\color{#d91a1a}-0.35\%$
test_ones_like 4.4878ms 4.3904ms 227.7698 Ops/s 228.4050 Ops/s $\color{#d91a1a}-0.28\%$
test_clone 6.8486ms 6.5986ms 151.5478 Ops/s 153.2385 Ops/s $\color{#d91a1a}-1.10\%$
test_squeeze 74.2310μs 15.8027μs 63.2802 KOps/s 68.7601 KOps/s $\textbf{\color{#d91a1a}-7.97\%}$
test_unsqueeze 0.1735ms 0.1177ms 8.4987 KOps/s 8.9243 KOps/s $\color{#d91a1a}-4.77\%$
test_split 0.3941ms 0.1899ms 5.2659 KOps/s 5.5602 KOps/s $\textbf{\color{#d91a1a}-5.29\%}$
test_permute 0.2735ms 0.2189ms 4.5690 KOps/s 4.8328 KOps/s $\textbf{\color{#d91a1a}-5.46\%}$
test_stack 19.1012ms 18.4746ms 54.1283 Ops/s 54.3190 Ops/s $\color{#d91a1a}-0.35\%$
test_cat 27.4332ms 26.9544ms 37.0997 Ops/s 54.6516 Ops/s $\textbf{\color{#d91a1a}-32.12\%}$
test_sequential_tensordict 0.7568ms 0.5801ms 1.7239 KOps/s 1.7364 KOps/s $\color{#d91a1a}-0.72\%$
test_sequential_graph_module 0.2728ms 0.1226ms 8.1569 KOps/s 8.1214 KOps/s $\color{#35bf28}+0.44\%$
test_nested_tensordict 1.7730ms 0.8560ms 1.1682 KOps/s 1.1964 KOps/s $\color{#d91a1a}-2.36\%$
test_nested_graph_module 0.2066ms 0.1349ms 7.4138 KOps/s 7.8923 KOps/s $\textbf{\color{#d91a1a}-6.06\%}$

@github-actions

github-actions Bot commented Jun 15, 2026 •
edited
Loading

Copy link
Copy Markdown
Contributor

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

Total Benchmarks: 344. Improved: $\large\color{#35bf28}32$. Worsened: $\large\color{#d91a1a}11$.

Expand to view detailed results
Name Max Mean Ops Ops on Repo HEAD Change
test_collector_preallocated_write[stack_out-cpu-32-2] 0.2188ms 0.1717ms 5.8248 KOps/s 5.7841 KOps/s $\color{#35bf28}+0.70\%$
test_collector_preallocated_write[stack_out-cpu-32-8] 0.3616ms 0.2975ms 3.3613 KOps/s 3.3609 KOps/s $\color{#35bf28}+0.01\%$
test_collector_preallocated_write[stack_out-cpu-32-32] 0.8739ms 0.7808ms 1.2807 KOps/s 1.2739 KOps/s $\color{#35bf28}+0.53\%$
test_collector_preallocated_write[stack_out-cpu-4096-2] 0.5769ms 0.3395ms 2.9451 KOps/s 2.9450 KOps/s $+0.00\%$
test_collector_preallocated_write[stack_out-cpu-4096-8] 1.3270ms 1.1023ms 907.1971 Ops/s 925.9948 Ops/s $\color{#d91a1a}-2.03\%$
test_collector_preallocated_write[stack_out-cpu-4096-32] 7.7106ms 7.2932ms 137.1142 Ops/s 137.5222 Ops/s $\color{#d91a1a}-0.30\%$
test_collector_preallocated_write[stack_out-cuda-32-2] 0.3087ms 0.2501ms 3.9976 KOps/s 3.9941 KOps/s $\color{#35bf28}+0.09\%$
test_collector_preallocated_write[stack_out-cuda-32-8] 0.4508ms 0.3717ms 2.6902 KOps/s 2.7201 KOps/s $\color{#d91a1a}-1.10\%$
test_collector_preallocated_write[stack_out-cuda-32-32] 0.8889ms 0.8286ms 1.2068 KOps/s 1.2167 KOps/s $\color{#d91a1a}-0.81\%$
test_collector_preallocated_write[stack_out-cuda-4096-2] 0.3066ms 0.2499ms 4.0008 KOps/s 4.0851 KOps/s $\color{#d91a1a}-2.06\%$
test_collector_preallocated_write[stack_out-cuda-4096-8] 0.4590ms 0.3741ms 2.6729 KOps/s 2.6979 KOps/s $\color{#d91a1a}-0.93\%$
test_collector_preallocated_write[stack_out-cuda-4096-32] 0.9946ms 0.8888ms 1.1251 KOps/s 1.1281 KOps/s $\color{#d91a1a}-0.27\%$
test_collector_preallocated_write[set_at_loop-cpu-32-2] 0.2640ms 0.2191ms 4.5645 KOps/s 4.6831 KOps/s $\color{#d91a1a}-2.53\%$
test_collector_preallocated_write[set_at_loop-cpu-32-8] 0.9367ms 0.8355ms 1.1969 KOps/s 1.1971 KOps/s $\color{#d91a1a}-0.02\%$
test_collector_preallocated_write[set_at_loop-cpu-32-32] 3.5121ms 3.3327ms 300.0606 Ops/s 298.8587 Ops/s $\color{#35bf28}+0.40\%$
test_collector_preallocated_write[set_at_loop-cpu-4096-2] 0.7121ms 0.3662ms 2.7304 KOps/s 2.7103 KOps/s $\color{#35bf28}+0.74\%$
test_collector_preallocated_write[set_at_loop-cpu-4096-8] 2.0992ms 1.7558ms 569.5455 Ops/s 609.5302 Ops/s $\textbf{\color{#d91a1a}-6.56\%}$
test_collector_preallocated_write[set_at_loop-cpu-4096-32] 10.5773ms 10.3389ms 96.7219 Ops/s 96.2918 Ops/s $\color{#35bf28}+0.45\%$
test_collector_preallocated_write[set_at_loop-cuda-32-2] 0.4506ms 0.3985ms 2.5093 KOps/s 2.5249 KOps/s $\color{#d91a1a}-0.62\%$
test_collector_preallocated_write[set_at_loop-cuda-32-8] 1.5887ms 1.4774ms 676.8576 Ops/s 663.6264 Ops/s $\color{#35bf28}+1.99\%$
test_collector_preallocated_write[set_at_loop-cuda-32-32] 6.2873ms 5.8247ms 171.6823 Ops/s 172.0372 Ops/s $\color{#d91a1a}-0.21\%$
test_collector_preallocated_write[set_at_loop-cuda-4096-2] 0.5421ms 0.3952ms 2.5302 KOps/s 2.4667 KOps/s $\color{#35bf28}+2.58\%$
test_collector_preallocated_write[set_at_loop-cuda-4096-8] 1.5943ms 1.4835ms 674.0770 Ops/s 662.1500 Ops/s $\color{#35bf28}+1.80\%$
test_collector_preallocated_write[set_at_loop-cuda-4096-32] 5.9382ms 5.7835ms 172.9069 Ops/s 171.3907 Ops/s $\color{#35bf28}+0.88\%$
test_collector_preallocated_write[copy_at-cpu-32-2] 0.5930ms 0.1754ms 5.7028 KOps/s 5.7155 KOps/s $\color{#d91a1a}-0.22\%$
test_collector_preallocated_write[copy_at-cpu-32-8] 0.7708ms 0.6799ms 1.4709 KOps/s 1.4785 KOps/s $\color{#d91a1a}-0.51\%$
test_collector_preallocated_write[copy_at-cpu-32-32] 2.8685ms 2.7252ms 366.9495 Ops/s 371.3381 Ops/s $\color{#d91a1a}-1.18\%$
test_collector_preallocated_write[copy_at-cpu-4096-2] 1.8544ms 0.3230ms 3.0963 KOps/s 3.1735 KOps/s $\color{#d91a1a}-2.43\%$
test_collector_preallocated_write[copy_at-cpu-4096-8] 2.1225ms 1.5292ms 653.9472 Ops/s 667.1057 Ops/s $\color{#d91a1a}-1.97\%$
test_collector_preallocated_write[copy_at-cpu-4096-32] 9.9673ms 9.5361ms 104.8650 Ops/s 105.9464 Ops/s $\color{#d91a1a}-1.02\%$
test_collector_preallocated_write[copy_at-cuda-32-2] 0.4064ms 0.3298ms 3.0323 KOps/s 3.0813 KOps/s $\color{#d91a1a}-1.59\%$
test_collector_preallocated_write[copy_at-cuda-32-8] 1.3129ms 1.2327ms 811.2266 Ops/s 821.1719 Ops/s $\color{#d91a1a}-1.21\%$
test_collector_preallocated_write[copy_at-cuda-32-32] 5.0295ms 4.8335ms 206.8915 Ops/s 210.3429 Ops/s $\color{#d91a1a}-1.64\%$
test_collector_preallocated_write[copy_at-cuda-4096-2] 0.3937ms 0.3318ms 3.0137 KOps/s 3.0652 KOps/s $\color{#d91a1a}-1.68\%$
test_collector_preallocated_write[copy_at-cuda-4096-8] 1.6774ms 1.2350ms 809.7075 Ops/s 820.1635 Ops/s $\color{#d91a1a}-1.27\%$
test_collector_preallocated_write[copy_at-cuda-4096-32] 4.9828ms 4.8393ms 206.6413 Ops/s 207.4262 Ops/s $\color{#d91a1a}-0.38\%$
test_collector_preallocated_write[manual_leaf_copy-cpu-32-2] 0.1769ms 0.1249ms 8.0075 KOps/s 8.3072 KOps/s $\color{#d91a1a}-3.61\%$
test_collector_preallocated_write[manual_leaf_copy-cpu-32-8] 0.5473ms 0.4882ms 2.0483 KOps/s 2.1221 KOps/s $\color{#d91a1a}-3.48\%$
test_collector_preallocated_write[manual_leaf_copy-cpu-32-32] 2.0475ms 1.9488ms 513.1347 Ops/s 528.8450 Ops/s $\color{#d91a1a}-2.97\%$
test_collector_preallocated_write[manual_leaf_copy-cpu-4096-2] 0.6666ms 0.2606ms 3.8372 KOps/s 3.8153 KOps/s $\color{#35bf28}+0.57\%$
test_collector_preallocated_write[manual_leaf_copy-cpu-4096-8] 1.4342ms 1.2908ms 774.7105 Ops/s 781.0072 Ops/s $\color{#d91a1a}-0.81\%$
test_collector_preallocated_write[manual_leaf_copy-cpu-4096-32] 9.0608ms 8.6833ms 115.1634 Ops/s 117.7469 Ops/s $\color{#d91a1a}-2.19\%$
test_collector_preallocated_write[manual_leaf_copy-cuda-32-2] 0.5064ms 0.2875ms 3.4785 KOps/s 3.2866 KOps/s $\textbf{\color{#35bf28}+5.84\%}$
test_collector_preallocated_write[manual_leaf_copy-cuda-32-8] 1.1541ms 1.0718ms 932.9956 Ops/s 878.1431 Ops/s $\textbf{\color{#35bf28}+6.25\%}$
test_collector_preallocated_write[manual_leaf_copy-cuda-32-32] 4.2833ms 4.1811ms 239.1725 Ops/s 225.3989 Ops/s $\textbf{\color{#35bf28}+6.11\%}$
test_collector_preallocated_write[manual_leaf_copy-cuda-4096-2] 0.3274ms 0.2836ms 3.5255 KOps/s 3.4140 KOps/s $\color{#35bf28}+3.27\%$
test_collector_preallocated_write[manual_leaf_copy-cuda-4096-8] 1.1500ms 1.0627ms 941.0146 Ops/s 960.0774 Ops/s $\color{#d91a1a}-1.99\%$
test_collector_preallocated_write[manual_leaf_copy-cuda-4096-32] 4.6305ms 4.1936ms 238.4580 Ops/s 244.2014 Ops/s $\color{#d91a1a}-2.35\%$
test_wide_nested_preallocated_write[stack_out-cpu-128-128] 16.9385ms 16.6325ms 60.1232 Ops/s 60.3456 Ops/s $\color{#d91a1a}-0.37\%$
test_wide_nested_preallocated_write[stack_out-cpu-128-256] 33.6479ms 33.3742ms 29.9632 Ops/s 30.1201 Ops/s $\color{#d91a1a}-0.52\%$
test_wide_nested_preallocated_write[stack_out-cpu-4096-128] 85.2473ms 84.5810ms 11.8230 Ops/s 11.9914 Ops/s $\color{#d91a1a}-1.40\%$
test_wide_nested_preallocated_write[stack_out-cpu-4096-256] 0.2034s 0.2027s 4.9329 Ops/s 5.0821 Ops/s $\color{#d91a1a}-2.94\%$
test_wide_nested_preallocated_write[stack_out-cuda-128-128] 14.1934ms 14.0899ms 70.9730 Ops/s 71.8081 Ops/s $\color{#d91a1a}-1.16\%$
test_wide_nested_preallocated_write[stack_out-cuda-128-256] 27.7993ms 27.3607ms 36.5488 Ops/s 36.9237 Ops/s $\color{#d91a1a}-1.02\%$
test_wide_nested_preallocated_write[stack_out-cuda-4096-128] 14.4119ms 14.2424ms 70.2130 Ops/s 70.8166 Ops/s $\color{#d91a1a}-0.85\%$
test_wide_nested_preallocated_write[stack_out-cuda-4096-256] 28.2518ms 28.0035ms 35.7099 Ops/s 36.4371 Ops/s $\color{#d91a1a}-2.00\%$
test_wide_nested_preallocated_write[set_at_loop-cpu-128-128] 53.9496ms 53.3810ms 18.7332 Ops/s 18.9968 Ops/s $\color{#d91a1a}-1.39\%$
test_wide_nested_preallocated_write[set_at_loop-cpu-128-256] 0.1090s 0.1079s 9.2680 Ops/s 9.4249 Ops/s $\color{#d91a1a}-1.66\%$
test_wide_nested_preallocated_write[set_at_loop-cpu-4096-128] 0.1473s 0.1465s 6.8253 Ops/s 6.8831 Ops/s $\color{#d91a1a}-0.84\%$
test_wide_nested_preallocated_write[set_at_loop-cpu-4096-256] 0.3238s 0.3217s 3.1081 Ops/s 3.1824 Ops/s $\color{#d91a1a}-2.34\%$
test_wide_nested_preallocated_write[set_at_loop-cuda-128-128] 80.0615ms 78.9617ms 12.6644 Ops/s 13.3687 Ops/s $\textbf{\color{#d91a1a}-5.27\%}$
test_wide_nested_preallocated_write[set_at_loop-cuda-128-256] 0.1594s 0.1578s 6.3372 Ops/s 6.2161 Ops/s $\color{#35bf28}+1.95\%$
test_wide_nested_preallocated_write[set_at_loop-cuda-4096-128] 80.8863ms 79.4145ms 12.5922 Ops/s 12.3945 Ops/s $\color{#35bf28}+1.59\%$
test_wide_nested_preallocated_write[set_at_loop-cuda-4096-256] 0.1606s 0.1584s 6.3134 Ops/s 6.1854 Ops/s $\color{#35bf28}+2.07\%$
test_wide_nested_preallocated_write[copy_at-cpu-128-128] 31.1892ms 30.9909ms 32.2676 Ops/s 32.7660 Ops/s $\color{#d91a1a}-1.52\%$
test_wide_nested_preallocated_write[copy_at-cpu-128-256] 63.1678ms 62.6026ms 15.9738 Ops/s 16.2093 Ops/s $\color{#d91a1a}-1.45\%$
test_wide_nested_preallocated_write[copy_at-cpu-4096-128] 0.1205s 0.1190s 8.4065 Ops/s 8.5371 Ops/s $\color{#d91a1a}-1.53\%$
test_wide_nested_preallocated_write[copy_at-cpu-4096-256] 0.2660s 0.2654s 3.7673 Ops/s 3.8146 Ops/s $\color{#d91a1a}-1.24\%$
test_wide_nested_preallocated_write[copy_at-cuda-128-128] 44.4866ms 44.2803ms 22.5834 Ops/s 23.0728 Ops/s $\color{#d91a1a}-2.12\%$
test_wide_nested_preallocated_write[copy_at-cuda-128-256] 88.8404ms 88.4078ms 11.3112 Ops/s 11.5685 Ops/s $\color{#d91a1a}-2.22\%$
test_wide_nested_preallocated_write[copy_at-cuda-4096-128] 44.6308ms 44.3936ms 22.5258 Ops/s 22.9201 Ops/s $\color{#d91a1a}-1.72\%$
test_wide_nested_preallocated_write[copy_at-cuda-4096-256] 89.6497ms 89.0994ms 11.2234 Ops/s 11.4513 Ops/s $\color{#d91a1a}-1.99\%$
test_wide_nested_preallocated_write[manual_leaf_copy-cpu-128-128] 21.7459ms 21.3859ms 46.7598 Ops/s 48.9841 Ops/s $\color{#d91a1a}-4.54\%$
test_wide_nested_preallocated_write[manual_leaf_copy-cpu-128-256] 43.6361ms 43.3858ms 23.0490 Ops/s 23.7316 Ops/s $\color{#d91a1a}-2.88\%$
test_wide_nested_preallocated_write[manual_leaf_copy-cpu-4096-128] 0.1099s 0.1079s 9.2721 Ops/s 9.5470 Ops/s $\color{#d91a1a}-2.88\%$
test_wide_nested_preallocated_write[manual_leaf_copy-cpu-4096-256] 0.2436s 0.2418s 4.1361 Ops/s 4.2022 Ops/s $\color{#d91a1a}-1.57\%$
test_wide_nested_preallocated_write[manual_leaf_copy-cuda-128-128] 38.9101ms 38.3570ms 26.0708 Ops/s 26.7536 Ops/s $\color{#d91a1a}-2.55\%$
test_wide_nested_preallocated_write[manual_leaf_copy-cuda-128-256] 77.5206ms 76.5747ms 13.0592 Ops/s 13.4059 Ops/s $\color{#d91a1a}-2.59\%$
test_wide_nested_preallocated_write[manual_leaf_copy-cuda-4096-128] 38.6591ms 38.2396ms 26.1509 Ops/s 26.8353 Ops/s $\color{#d91a1a}-2.55\%$
test_wide_nested_preallocated_write[manual_leaf_copy-cuda-4096-256] 76.7498ms 76.2445ms 13.1157 Ops/s 13.4299 Ops/s $\color{#d91a1a}-2.34\%$
test_plain_set_nested 35.5410μs 15.6566μs 63.8710 KOps/s 64.7342 KOps/s $\color{#d91a1a}-1.33\%$
test_plain_set_stack_nested 42.7010μs 15.2952μs 65.3801 KOps/s 65.4627 KOps/s $\color{#d91a1a}-0.13\%$
test_plain_set_nested_inplace 50.7710μs 16.9083μs 59.1427 KOps/s 58.4808 KOps/s $\color{#35bf28}+1.13\%$
test_plain_set_stack_nested_inplace 44.5710μs 16.7325μs 59.7639 KOps/s 60.0305 KOps/s $\color{#d91a1a}-0.44\%$
test_items 29.2610μs 6.0069μs 166.4762 KOps/s 165.3220 KOps/s $\color{#35bf28}+0.70\%$
test_items_nested 0.5243ms 0.4718ms 2.1194 KOps/s 2.1128 KOps/s $\color{#35bf28}+0.31\%$
test_items_nested_locked 0.5069ms 0.4709ms 2.1235 KOps/s 2.0913 KOps/s $\color{#35bf28}+1.54\%$
test_items_nested_leaf 0.1248ms 96.3364μs 10.3803 KOps/s 10.2185 KOps/s $\color{#35bf28}+1.58\%$
test_items_stack_nested 0.5655ms 0.4716ms 2.1204 KOps/s 2.0965 KOps/s $\color{#35bf28}+1.14\%$
test_items_stack_nested_leaf 0.1332ms 96.7911μs 10.3315 KOps/s 10.3985 KOps/s $\color{#d91a1a}-0.64\%$
test_items_stack_nested_locked 0.5327ms 0.4725ms 2.1163 KOps/s 2.0995 KOps/s $\color{#35bf28}+0.80\%$
test_keys 29.6700μs 4.2060μs 237.7584 KOps/s 236.8694 KOps/s $\color{#35bf28}+0.38\%$
test_keys_nested 0.1681ms 0.1311ms 7.6254 KOps/s 7.6316 KOps/s $\color{#d91a1a}-0.08\%$
test_keys_nested_locked 0.7614ms 0.1403ms 7.1276 KOps/s 7.1721 KOps/s $\color{#d91a1a}-0.62\%$
test_keys_nested_leaf 0.1637ms 0.1231ms 8.1218 KOps/s 8.2404 KOps/s $\color{#d91a1a}-1.44\%$
test_keys_stack_nested 0.1632ms 0.1307ms 7.6524 KOps/s 7.6046 KOps/s $\color{#35bf28}+0.63\%$
test_keys_stack_nested_leaf 0.1565ms 0.1217ms 8.2164 KOps/s 8.2733 KOps/s $\color{#d91a1a}-0.69\%$
test_keys_stack_nested_locked 0.1938ms 0.1389ms 7.2011 KOps/s 7.2197 KOps/s $\color{#d91a1a}-0.26\%$
test_values 7.2762μs 1.0280μs 972.7395 KOps/s 980.3590 KOps/s $\color{#d91a1a}-0.78\%$
test_values_nested 82.0620μs 52.6809μs 18.9822 KOps/s 18.9470 KOps/s $\color{#35bf28}+0.19\%$
test_values_nested_locked 86.0120μs 55.9637μs 17.8687 KOps/s 17.9031 KOps/s $\color{#d91a1a}-0.19\%$
test_values_nested_leaf 0.1237ms 59.8504μs 16.7083 KOps/s 16.5571 KOps/s $\color{#35bf28}+0.91\%$
test_values_stack_nested 81.6720μs 52.3369μs 19.1070 KOps/s 19.1762 KOps/s $\color{#d91a1a}-0.36\%$
test_values_stack_nested_leaf 85.2220μs 60.4488μs 16.5429 KOps/s 16.6068 KOps/s $\color{#d91a1a}-0.38\%$
test_values_stack_nested_locked 93.7420μs 56.2381μs 17.7816 KOps/s 17.8733 KOps/s $\color{#d91a1a}-0.51\%$
test_membership 4.9252μs 0.8527μs 1.1728 MOps/s 1.1866 MOps/s $\color{#d91a1a}-1.16\%$
test_membership_nested 33.6100μs 3.1376μs 318.7108 KOps/s 332.0419 KOps/s $\color{#d91a1a}-4.01\%$
test_membership_nested_leaf 48.2710μs 2.9735μs 336.3073 KOps/s 335.4173 KOps/s $\color{#35bf28}+0.27\%$
test_membership_stacked_nested 42.4310μs 2.9317μs 341.0954 KOps/s 340.6736 KOps/s $\color{#35bf28}+0.12\%$
test_membership_stacked_nested_leaf 21.9500μs 2.9185μs 342.6449 KOps/s 345.1553 KOps/s $\color{#d91a1a}-0.73\%$
test_membership_nested_last 30.6800μs 4.5737μs 218.6417 KOps/s 222.3533 KOps/s $\color{#d91a1a}-1.67\%$
test_membership_nested_leaf_last 41.3110μs 4.5379μs 220.3667 KOps/s 221.1836 KOps/s $\color{#d91a1a}-0.37\%$
test_membership_stacked_nested_last 52.0010μs 4.3953μs 227.5153 KOps/s 228.2134 KOps/s $\color{#d91a1a}-0.31\%$
test_membership_stacked_nested_leaf_last 28.0000μs 4.3952μs 227.5203 KOps/s 231.0136 KOps/s $\color{#d91a1a}-1.51\%$
test_nested_getleaf 73.5710μs 20.9290μs 47.7806 KOps/s 45.9016 KOps/s $\color{#35bf28}+4.09\%$
test_nested_get 47.7510μs 20.5834μs 48.5828 KOps/s 48.5743 KOps/s $\color{#35bf28}+0.02\%$
test_stacked_getleaf 77.5210μs 21.3344μs 46.8727 KOps/s 46.6533 KOps/s $\color{#35bf28}+0.47\%$
test_stacked_get 49.0610μs 20.4636μs 48.8673 KOps/s 48.5925 KOps/s $\color{#35bf28}+0.57\%$
test_nested_getitemleaf 46.6010μs 22.0298μs 45.3930 KOps/s 44.8421 KOps/s $\color{#35bf28}+1.23\%$
test_nested_getitem 47.8910μs 20.9060μs 47.8331 KOps/s 47.2945 KOps/s $\color{#35bf28}+1.14\%$
test_stacked_getitemleaf 55.7410μs 21.8639μs 45.7376 KOps/s 45.5802 KOps/s $\color{#35bf28}+0.35\%$
test_stacked_getitem 45.8410μs 20.8286μs 48.0109 KOps/s 47.8431 KOps/s $\color{#35bf28}+0.35\%$
test_lock_nested 5.1809ms 0.5092ms 1.9638 KOps/s 1.9873 KOps/s $\color{#d91a1a}-1.18\%$
test_lock_stack_nested 0.5658ms 0.5142ms 1.9449 KOps/s 1.9715 KOps/s $\color{#d91a1a}-1.35\%$
test_unlock_nested 0.5107ms 0.4014ms 2.4910 KOps/s 2.5290 KOps/s $\color{#d91a1a}-1.50\%$
test_unlock_stack_nested 0.4518ms 0.4050ms 2.4693 KOps/s 2.5092 KOps/s $\color{#d91a1a}-1.59\%$
test_flatten_speed 0.1679ms 0.1214ms 8.2359 KOps/s 8.1574 KOps/s $\color{#35bf28}+0.96\%$
test_unflatten_speed 0.6252ms 0.5634ms 1.7750 KOps/s 1.7509 KOps/s $\color{#35bf28}+1.37\%$
test_common_ops 0.1573s 1.6219ms 616.5504 Ops/s 214.6796 Ops/s $\textbf{\color{#35bf28}+187.20\%}$
test_creation 0.1191ms 3.1671μs 315.7511 KOps/s 315.8453 KOps/s $\color{#d91a1a}-0.03\%$
test_creation_empty 32.5800μs 6.9949μs 142.9610 KOps/s 143.3048 KOps/s $\color{#d91a1a}-0.24\%$
test_creation_nested_1 41.2710μs 11.4618μs 87.2460 KOps/s 87.4920 KOps/s $\color{#d91a1a}-0.28\%$
test_creation_nested_2 39.4010μs 13.6183μs 73.4307 KOps/s 74.1505 KOps/s $\color{#d91a1a}-0.97\%$
test_creation_many_keys[10] 85.1820μs 20.4140μs 48.9859 KOps/s 47.6320 KOps/s $\color{#35bf28}+2.84\%$
test_creation_many_keys[50] 0.1200ms 90.7640μs 11.0176 KOps/s 11.1823 KOps/s $\color{#d91a1a}-1.47\%$
test_creation_many_keys[100] 0.2233ms 0.1788ms 5.5931 KOps/s 5.7329 KOps/s $\color{#d91a1a}-2.44\%$
test_creation_nested_many_keys[10] 74.9720μs 45.5240μs 21.9665 KOps/s 21.9272 KOps/s $\color{#35bf28}+0.18\%$
test_creation_nested_many_keys[50] 0.2480ms 0.1862ms 5.3716 KOps/s 5.4819 KOps/s $\color{#d91a1a}-2.01\%$
test_clone 31.9300μs 13.3615μs 74.8419 KOps/s 74.6876 KOps/s $\color{#35bf28}+0.21\%$
test_getitem[int] 1.4214ms 15.4521μs 64.7160 KOps/s 65.8504 KOps/s $\color{#d91a1a}-1.72\%$
test_getitem[slice_int] 0.1509ms 24.9970μs 40.0049 KOps/s 40.7398 KOps/s $\color{#d91a1a}-1.80\%$
test_getitem[range] 0.1855ms 64.3675μs 15.5358 KOps/s 15.3487 KOps/s $\color{#35bf28}+1.22\%$
test_getitem[tuple] 0.1455ms 24.7930μs 40.3339 KOps/s 41.5791 KOps/s $\color{#d91a1a}-2.99\%$
test_getitem[list] 0.1921ms 59.6064μs 16.7767 KOps/s 15.8327 KOps/s $\textbf{\color{#35bf28}+5.96\%}$
test_setitem_dim[int] 47.2110μs 25.8402μs 38.6994 KOps/s 36.7449 KOps/s $\textbf{\color{#35bf28}+5.32\%}$
test_setitem_dim[slice_int] 77.9110μs 42.5538μs 23.4997 KOps/s 22.2300 KOps/s $\textbf{\color{#35bf28}+5.71\%}$
test_setitem_dim[range] 0.1301ms 95.7615μs 10.4426 KOps/s 9.8340 KOps/s $\textbf{\color{#35bf28}+6.19\%}$
test_setitem_dim[tuple] 71.0520μs 39.5842μs 25.2626 KOps/s 23.8624 KOps/s $\textbf{\color{#35bf28}+5.87\%}$
test_setitem 41.1710μs 17.7255μs 56.4158 KOps/s 52.1753 KOps/s $\textbf{\color{#35bf28}+8.13\%}$
test_set 48.3410μs 17.1730μs 58.2308 KOps/s 57.3179 KOps/s $\color{#35bf28}+1.59\%$
test_set_shared 0.6531ms 0.2113ms 4.7333 KOps/s 4.4674 KOps/s $\textbf{\color{#35bf28}+5.95\%}$
test_update 0.3241ms 22.3182μs 44.8065 KOps/s 42.5957 KOps/s $\textbf{\color{#35bf28}+5.19\%}$
test_update_nested 84.4420μs 33.8816μs 29.5146 KOps/s 28.8410 KOps/s $\color{#35bf28}+2.34\%$
test_update__nested 0.4601ms 33.6141μs 29.7494 KOps/s 28.5375 KOps/s $\color{#35bf28}+4.25\%$
test_set_nested 54.2710μs 18.9945μs 52.6469 KOps/s 48.5321 KOps/s $\textbf{\color{#35bf28}+8.48\%}$
test_set_nested_new 60.9810μs 23.9154μs 41.8141 KOps/s 38.8418 KOps/s $\textbf{\color{#35bf28}+7.65\%}$
test_select 67.5620μs 40.8437μs 24.4836 KOps/s 24.0323 KOps/s $\color{#35bf28}+1.88\%$
test_select_nested 0.1162ms 74.4011μs 13.4407 KOps/s 13.3274 KOps/s $\color{#35bf28}+0.85\%$
test_exclude_nested 0.1353ms 91.6383μs 10.9125 KOps/s 10.9415 KOps/s $\color{#d91a1a}-0.27\%$
test_empty[True] 0.5628ms 0.3981ms 2.5117 KOps/s 2.4929 KOps/s $\color{#35bf28}+0.76\%$
test_empty[False] 11.7803μs 1.3912μs 718.8145 KOps/s 709.9003 KOps/s $\color{#35bf28}+1.26\%$
test_to 97.4120μs 67.6807μs 14.7753 KOps/s 13.5874 KOps/s $\textbf{\color{#35bf28}+8.74\%}$
test_to_nonblocking 0.1089ms 68.3665μs 14.6270 KOps/s 13.7813 KOps/s $\textbf{\color{#35bf28}+6.14\%}$
test_unbind_speed 0.4496ms 0.3434ms 2.9124 KOps/s 2.9706 KOps/s $\color{#d91a1a}-1.96\%$
test_unbind_speed_stack0 0.4296ms 0.3392ms 2.9484 KOps/s 3.0268 KOps/s $\color{#d91a1a}-2.59\%$
test_unbind_speed_stack1 0.1155s 0.8613ms 1.1610 KOps/s 1.1765 KOps/s $\color{#d91a1a}-1.31\%$
test_split 1.2357ms 1.1465ms 872.1988 Ops/s 715.3990 Ops/s $\textbf{\color{#35bf28}+21.92\%}$
test_chunk 0.1152s 1.2207ms 819.2168 Ops/s 916.9244 Ops/s $\textbf{\color{#d91a1a}-10.66\%}$
test_to_cpu_blocking 20.6407ms 20.1588ms 49.6061 Ops/s 49.4836 Ops/s $\color{#35bf28}+0.25\%$
test_to_cpu_global_sync 12.1639ms 11.8439ms 84.4316 Ops/s 84.6561 Ops/s $\color{#d91a1a}-0.27\%$
test_to_cpu_event_sync 13.2314ms 12.8547ms 77.7923 Ops/s 69.6377 Ops/s $\textbf{\color{#35bf28}+11.71\%}$
test_to_cpu_default 13.3426ms 12.8742ms 77.6749 Ops/s 77.6000 Ops/s $\color{#35bf28}+0.10\%$
test_consolidate[False-None] 0.1449s 4.8168ms 207.6088 Ops/s 238.4634 Ops/s $\textbf{\color{#d91a1a}-12.94\%}$
test_consolidate[default-None] 3.5485ms 2.2166ms 451.1366 Ops/s 460.2098 Ops/s $\color{#d91a1a}-1.97\%$
test_consolidate[reduce-overhead-None] 0.2050s 2.5828ms 387.1797 Ops/s 475.8197 Ops/s $\textbf{\color{#d91a1a}-18.63\%}$
test_consolidate_njt[False-None] 9.3467ms 8.5606ms 116.8139 Ops/s 115.5922 Ops/s $\color{#35bf28}+1.06\%$
test_to[False-False-None] 2.2615ms 2.1686ms 461.1371 Ops/s 455.5667 Ops/s $\color{#35bf28}+1.22\%$
test_to[True-False-None] 2.2201ms 1.9504ms 512.7052 Ops/s 519.6722 Ops/s $\color{#d91a1a}-1.34\%$
test_to[within-False-None] 6.5102ms 6.2583ms 159.7883 Ops/s 160.8094 Ops/s $\color{#d91a1a}-0.63\%$
test_to[True-default-None] 11.0517ms 10.8606ms 92.0756 Ops/s 92.7206 Ops/s $\color{#d91a1a}-0.70\%$
test_to_njt[False-False-None] 8.8117ms 8.6537ms 115.5570 Ops/s 113.4308 Ops/s $\color{#35bf28}+1.87\%$
test_to_njt[True-False-None] 7.2364ms 7.0538ms 141.7669 Ops/s 144.0290 Ops/s $\color{#d91a1a}-1.57\%$
test_to_njt[within-False-None] 0.2058s 18.7619ms 53.2995 Ops/s 32.9579 Ops/s $\textbf{\color{#35bf28}+61.72\%}$
test_creation[device0] 0.4519ms 0.1188ms 8.4142 KOps/s 8.4536 KOps/s $\color{#d91a1a}-0.47\%$
test_creation_from_tensor 0.5011ms 0.1154ms 8.6620 KOps/s 8.7166 KOps/s $\color{#d91a1a}-0.63\%$
test_add_one[memmap_tensor0] 0.3006ms 6.9194μs 144.5212 KOps/s 146.5749 KOps/s $\color{#d91a1a}-1.40\%$
test_contiguous[memmap_tensor0] 22.4310μs 0.6649μs 1.5040 MOps/s 2.0518 MOps/s $\textbf{\color{#d91a1a}-26.70\%}$
test_stack[memmap_tensor0] 28.4800μs 4.4865μs 222.8923 KOps/s 233.6836 KOps/s $\color{#d91a1a}-4.62\%$
test_memmaptd_index 1.3250ms 0.2887ms 3.4636 KOps/s 3.6715 KOps/s $\textbf{\color{#d91a1a}-5.66\%}$
test_memmaptd_index_astensor 0.5569ms 0.3940ms 2.5382 KOps/s 2.6267 KOps/s $\color{#d91a1a}-3.37\%$
test_memmaptd_index_op 0.8393ms 0.6722ms 1.4877 KOps/s 1.5555 KOps/s $\color{#d91a1a}-4.36\%$
test_serialize_model 0.1381s 0.1355s 7.3827 Ops/s 7.4169 Ops/s $\color{#d91a1a}-0.46\%$
test_serialize_model_pickle 1.3630s 1.2126s 0.8247 Ops/s 0.8226 Ops/s $\color{#35bf28}+0.25\%$
test_serialize_weights 0.1364s 0.1335s 7.4907 Ops/s 7.4221 Ops/s $\color{#35bf28}+0.92\%$
test_serialize_weights_returnearly 0.4854s 92.3405ms 10.8295 Ops/s 10.6912 Ops/s $\color{#35bf28}+1.29\%$
test_serialize_weights_pickle 1.3799s 1.2145s 0.8234 Ops/s 0.8219 Ops/s $\color{#35bf28}+0.18\%$
test_reshape_pytree 0.2101ms 33.5117μs 29.8403 KOps/s 30.5613 KOps/s $\color{#d91a1a}-2.36\%$
test_reshape_td 72.8810μs 45.1721μs 22.1376 KOps/s 22.5684 KOps/s $\color{#d91a1a}-1.91\%$
test_view_pytree 0.2237ms 33.0086μs 30.2951 KOps/s 31.1785 KOps/s $\color{#d91a1a}-2.83\%$
test_view_td 0.1040ms 54.4045μs 18.3808 KOps/s 18.8292 KOps/s $\color{#d91a1a}-2.38\%$
test_unbind_pytree 0.2413ms 37.6307μs 26.5740 KOps/s 26.8220 KOps/s $\color{#d91a1a}-0.92\%$
test_unbind_td 0.1236ms 51.0740μs 19.5794 KOps/s 20.0112 KOps/s $\color{#d91a1a}-2.16\%$
test_split_pytree 0.2519ms 43.4824μs 22.9978 KOps/s 23.2153 KOps/s $\color{#d91a1a}-0.94\%$
test_split_td 0.1510ms 66.1726μs 15.1120 KOps/s 15.5485 KOps/s $\color{#d91a1a}-2.81\%$
test_add_pytree 0.2324ms 42.3435μs 23.6164 KOps/s 23.0822 KOps/s $\color{#35bf28}+2.31\%$
test_add_td 99.8720μs 56.3444μs 17.7480 KOps/s 18.0187 KOps/s $\color{#d91a1a}-1.50\%$
test_compile_add_one_nested[tensordict-compile] 0.3361ms 0.2408ms 4.1529 KOps/s 4.2316 KOps/s $\color{#d91a1a}-1.86\%$
test_compile_add_one_nested[tensordict-eager] 0.2683ms 0.2009ms 4.9771 KOps/s 4.9516 KOps/s $\color{#35bf28}+0.51\%$
test_compile_add_one_nested[pytree-compile] 0.2645ms 0.1339ms 7.4657 KOps/s 6.9952 KOps/s $\textbf{\color{#35bf28}+6.73\%}$
test_compile_add_one_nested[pytree-eager] 0.4870ms 0.1784ms 5.6057 KOps/s 5.4979 KOps/s $\color{#35bf28}+1.96\%$
test_compile_copy_nested[tensordict-compile] 75.9010μs 15.8099μs 63.2515 KOps/s 64.9021 KOps/s $\color{#d91a1a}-2.54\%$
test_compile_copy_nested[tensordict-eager] 0.3464ms 53.4291μs 18.7164 KOps/s 18.2483 KOps/s $\color{#35bf28}+2.56\%$
test_compile_copy_nested[pytree-compile] 0.1466ms 15.8738μs 62.9968 KOps/s 64.1544 KOps/s $\color{#d91a1a}-1.80\%$
test_compile_copy_nested[pytree-eager] 0.3949ms 67.4830μs 14.8185 KOps/s 14.7776 KOps/s $\color{#35bf28}+0.28\%$
test_compile_add_one_flat[tensordict-compile] 0.6900ms 0.2567ms 3.8952 KOps/s 3.8697 KOps/s $\color{#35bf28}+0.66\%$
test_compile_add_one_flat[tensordict-eager] 0.3554ms 0.2821ms 3.5443 KOps/s 3.4966 KOps/s $\color{#35bf28}+1.36\%$
test_compile_add_one_flat[tensorclass-compile] 0.3653ms 0.1949ms 5.1315 KOps/s 5.1741 KOps/s $\color{#d91a1a}-0.82\%$
test_compile_add_one_flat[tensorclass-eager] 0.1142ms 74.1845μs 13.4799 KOps/s 13.3186 KOps/s $\color{#35bf28}+1.21\%$
test_compile_add_one_flat[pytree-compile] 0.5438ms 0.1920ms 5.2086 KOps/s 5.0024 KOps/s $\color{#35bf28}+4.12\%$
test_compile_add_one_flat[pytree-eager] 0.7949ms 0.5267ms 1.8986 KOps/s 1.8841 KOps/s $\color{#35bf28}+0.77\%$
test_compile_add_self_flat[tensordict-eager] 0.5104ms 0.3317ms 3.0143 KOps/s 2.9862 KOps/s $\color{#35bf28}+0.94\%$
test_compile_add_self_flat[tensordict-compile] 0.5241ms 0.2554ms 3.9160 KOps/s 3.7149 KOps/s $\textbf{\color{#35bf28}+5.41\%}$
test_compile_add_self_flat[tensorclass-eager] 0.1372ms 90.8585μs 11.0061 KOps/s 11.0443 KOps/s $\color{#d91a1a}-0.35\%$
test_compile_add_self_flat[tensorclass-compile] 0.4623ms 0.1923ms 5.1994 KOps/s 5.0090 KOps/s $\color{#35bf28}+3.80\%$
test_compile_add_self_flat[pytree-eager] 0.7029ms 0.4381ms 2.2828 KOps/s 2.2933 KOps/s $\color{#d91a1a}-0.46\%$
test_compile_add_self_flat[pytree-compile] 0.5257ms 0.1926ms 5.1925 KOps/s 4.8583 KOps/s $\textbf{\color{#35bf28}+6.88\%}$
test_compile_copy_flat[tensordict-compile] 63.2710μs 19.5190μs 51.2322 KOps/s 50.6521 KOps/s $\color{#35bf28}+1.15\%$
test_compile_copy_flat[tensordict-eager] 0.1432ms 41.4095μs 24.1490 KOps/s 24.2933 KOps/s $\color{#d91a1a}-0.59\%$
test_compile_copy_flat[pytree-compile] 0.2026ms 20.8703μs 47.9149 KOps/s 44.3917 KOps/s $\textbf{\color{#35bf28}+7.94\%}$
test_compile_copy_flat[pytree-eager] 0.4831ms 52.2196μs 19.1499 KOps/s 19.0698 KOps/s $\color{#35bf28}+0.42\%$
test_compile_assign_and_add[tensordict-compile] 2.2087ms 0.2410ms 4.1496 KOps/s 3.7444 KOps/s $\textbf{\color{#35bf28}+10.82\%}$
test_compile_assign_and_add[tensordict-eager] 3.5214ms 3.2507ms 307.6302 Ops/s 307.0877 Ops/s $\color{#35bf28}+0.18\%$
test_compile_assign_and_add[pytree-compile] 2.1037ms 0.1840ms 5.4341 KOps/s 5.3889 KOps/s $\color{#35bf28}+0.84\%$
test_compile_assign_and_add[pytree-eager] 2.9570ms 2.8106ms 355.7904 Ops/s 355.7194 Ops/s $\color{#35bf28}+0.02\%$
test_compile_indexing[tensor-tensordict-compile] 0.2825ms 0.1353ms 7.3902 KOps/s 7.2001 KOps/s $\color{#35bf28}+2.64\%$
test_compile_indexing[tensor-tensordict-eager] 0.3958ms 74.8789μs 13.3549 KOps/s 12.8553 KOps/s $\color{#35bf28}+3.89\%$
test_compile_indexing[tensor-tensorclass-compile] 0.2623ms 0.1185ms 8.4412 KOps/s 8.1487 KOps/s $\color{#35bf28}+3.59\%$
test_compile_indexing[tensor-tensorclass-eager] 0.3114ms 45.4131μs 22.0201 KOps/s 21.3154 KOps/s $\color{#35bf28}+3.31\%$
test_compile_indexing[tensor-pytree-compile] 0.1838ms 0.1186ms 8.4346 KOps/s 7.9498 KOps/s $\textbf{\color{#35bf28}+6.10\%}$
test_compile_indexing[tensor-pytree-eager] 0.2607ms 45.3381μs 22.0565 KOps/s 21.5911 KOps/s $\color{#35bf28}+2.16\%$
test_compile_indexing[slice-tensordict-compile] 0.2572ms 86.2872μs 11.5892 KOps/s 11.6615 KOps/s $\color{#d91a1a}-0.62\%$
test_compile_indexing[slice-tensordict-eager] 0.2068ms 30.9276μs 32.3336 KOps/s 36.3939 KOps/s $\textbf{\color{#d91a1a}-11.16\%}$
test_compile_indexing[slice-tensorclass-compile] 0.1960ms 70.3435μs 14.2160 KOps/s 13.9271 KOps/s $\color{#35bf28}+2.07\%$
test_compile_indexing[slice-tensorclass-eager] 0.2481ms 22.5028μs 44.4389 KOps/s 44.7124 KOps/s $\color{#d91a1a}-0.61\%$
test_compile_indexing[slice-pytree-compile] 0.1245ms 71.3089μs 14.0235 KOps/s 13.6981 KOps/s $\color{#35bf28}+2.38\%$
test_compile_indexing[slice-pytree-eager] 0.2356ms 22.5375μs 44.3706 KOps/s 45.3117 KOps/s $\color{#d91a1a}-2.08\%$
test_compile_indexing[int-tensordict-compile] 0.1284ms 85.6319μs 11.6779 KOps/s 11.1086 KOps/s $\textbf{\color{#35bf28}+5.12\%}$
test_compile_indexing[int-tensordict-eager] 0.2635ms 27.6419μs 36.1770 KOps/s 36.5278 KOps/s $\color{#d91a1a}-0.96\%$
test_compile_indexing[int-tensorclass-compile] 0.1065ms 70.8376μs 14.1168 KOps/s 14.1385 KOps/s $\color{#d91a1a}-0.15\%$
test_compile_indexing[int-tensorclass-eager] 0.2419ms 22.2882μs 44.8667 KOps/s 44.8111 KOps/s $\color{#35bf28}+0.12\%$
test_compile_indexing[int-pytree-compile] 0.1258ms 71.6616μs 13.9545 KOps/s 13.7328 KOps/s $\color{#35bf28}+1.61\%$
test_compile_indexing[int-pytree-eager] 0.2496ms 22.4901μs 44.4641 KOps/s 44.8506 KOps/s $\color{#d91a1a}-0.86\%$
test_compile_replace[single-eager] 99.0820μs 48.2222μs 20.7373 KOps/s 20.1072 KOps/s $\color{#35bf28}+3.13\%$
test_compile_replace[single-compile] 0.2427ms 0.1353ms 7.3897 KOps/s 7.2468 KOps/s $\color{#35bf28}+1.97\%$
test_compile_replace[multi-eager] 0.6451ms 0.5713ms 1.7505 KOps/s 1.6953 KOps/s $\color{#35bf28}+3.26\%$
test_compile_replace[multi-compile] 0.3352ms 0.1406ms 7.1133 KOps/s 6.7851 KOps/s $\color{#35bf28}+4.84\%$
test_compile_tc_getattr_20[eager] 0.2103ms 0.1702ms 5.8767 KOps/s 6.0674 KOps/s $\color{#d91a1a}-3.14\%$
test_compile_tc_getattr_20[compile] 0.4009ms 0.1473ms 6.7890 KOps/s 6.5061 KOps/s $\color{#35bf28}+4.35\%$
test_compile_clone_shallow[20-eager] 49.5810μs 19.2314μs 51.9984 KOps/s 52.1292 KOps/s $\color{#d91a1a}-0.25\%$
test_compile_clone_shallow[20-compile] 89.0320μs 17.1019μs 58.4732 KOps/s 59.3478 KOps/s $\color{#d91a1a}-1.47\%$
test_compile_clone_shallow[40-eager] 64.7210μs 33.8847μs 29.5119 KOps/s 29.7381 KOps/s $\color{#d91a1a}-0.76\%$
test_compile_clone_shallow[40-compile] 90.4820μs 18.5861μs 53.8036 KOps/s 55.2394 KOps/s $\color{#d91a1a}-2.60\%$
test_compile_clone_shallow[80-eager] 0.1005ms 62.4432μs 16.0145 KOps/s 16.1341 KOps/s $\color{#d91a1a}-0.74\%$
test_compile_clone_shallow[80-compile] 61.6920μs 23.4378μs 42.6661 KOps/s 45.4050 KOps/s $\textbf{\color{#d91a1a}-6.03\%}$
test_compile_update_inplace[eager] 0.1043ms 61.8813μs 16.1600 KOps/s 16.2199 KOps/s $\color{#d91a1a}-0.37\%$
test_compile_update_inplace[compile] 0.4219ms 0.1514ms 6.6043 KOps/s 6.4550 KOps/s $\color{#35bf28}+2.31\%$
test_mod_add[eager] 0.2869ms 0.2284ms 4.3791 KOps/s 4.3849 KOps/s $\color{#d91a1a}-0.13\%$
test_mod_add[compile] 0.2427ms 0.1332ms 7.5102 KOps/s 7.0233 KOps/s $\textbf{\color{#35bf28}+6.93\%}$
test_mod_add[compile-overhead] 0.3988ms 0.1801ms 5.5519 KOps/s 5.2538 KOps/s $\textbf{\color{#35bf28}+5.67\%}$
test_mod_wrap[eager] 0.6016ms 0.4677ms 2.1380 KOps/s 2.0350 KOps/s $\textbf{\color{#35bf28}+5.06\%}$
test_mod_wrap[compile] 0.4668ms 0.3981ms 2.5122 KOps/s 2.4749 KOps/s $\color{#35bf28}+1.51\%$
test_mod_wrap[compile-overhead] 8.4796ms 4.6458ms 215.2483 Ops/s 218.2713 Ops/s $\color{#d91a1a}-1.38\%$
test_mod_wrap_and_backward[eager] 1.8697ms 1.7071ms 585.8012 Ops/s 574.9745 Ops/s $\color{#35bf28}+1.88\%$
test_mod_wrap_and_backward[compile] 2.7991ms 1.5910ms 628.5382 Ops/s 601.8766 Ops/s $\color{#35bf28}+4.43\%$
test_mod_wrap_and_backward[compile-overhead] 1.3583ms 0.9482ms 1.0546 KOps/s 1.0300 KOps/s $\color{#35bf28}+2.39\%$
test_seq_add[eager] 0.5947ms 0.5123ms 1.9521 KOps/s 1.9594 KOps/s $\color{#d91a1a}-0.37\%$
test_seq_add[compile] 0.2642ms 0.1412ms 7.0808 KOps/s 6.8501 KOps/s $\color{#35bf28}+3.37\%$
test_seq_add[compile-overhead] 0.2419ms 0.1837ms 5.4448 KOps/s 5.1565 KOps/s $\textbf{\color{#35bf28}+5.59\%}$
test_seq_wrap[eager] 1.5182ms 1.4179ms 705.2450 Ops/s 707.8276 Ops/s $\color{#d91a1a}-0.36\%$
test_seq_wrap[compile] 0.5164ms 0.4136ms 2.4178 KOps/s 2.3419 KOps/s $\color{#35bf28}+3.24\%$
test_seq_wrap[compile-overhead] 0.4284ms 0.3069ms 3.2580 KOps/s 3.1740 KOps/s $\color{#35bf28}+2.65\%$
test_func_call_runtime[False-eager] 0.8604ms 0.8022ms 1.2466 KOps/s 1.1820 KOps/s $\textbf{\color{#35bf28}+5.46\%}$
test_func_call_runtime[False-compile] 1.0819ms 0.9924ms 1.0077 KOps/s 1.0107 KOps/s $\color{#d91a1a}-0.30\%$
test_func_call_runtime[False-compile-overhead] 0.5896ms 0.5157ms 1.9392 KOps/s 1.9292 KOps/s $\color{#35bf28}+0.52\%$
test_func_call_runtime[True-eager] 1.1812ms 1.0660ms 938.0845 Ops/s 936.9085 Ops/s $\color{#35bf28}+0.13\%$
test_func_call_runtime[True-compile] 1.1073ms 1.0027ms 997.3456 Ops/s 1.0037 KOps/s $\color{#d91a1a}-0.64\%$
test_func_call_runtime[True-compile-overhead] 0.5721ms 0.5295ms 1.8886 KOps/s 1.8659 KOps/s $\color{#35bf28}+1.22\%$
test_func_call_cm_runtime[False-eager] 0.8856ms 0.7983ms 1.2527 KOps/s 1.2066 KOps/s $\color{#35bf28}+3.82\%$
test_func_call_cm_runtime[False-compile] 1.0634ms 0.9704ms 1.0305 KOps/s 1.0290 KOps/s $\color{#35bf28}+0.15\%$
test_func_call_cm_runtime[False-compile-overhead] 0.6986ms 0.4928ms 2.0292 KOps/s 2.0082 KOps/s $\color{#35bf28}+1.04\%$
test_func_call_cm_runtime[True-eager] 1.3431ms 1.2110ms 825.7895 Ops/s 820.0357 Ops/s $\color{#35bf28}+0.70\%$
test_func_call_cm_runtime[True-compile] 1.1546ms 1.0291ms 971.7376 Ops/s 973.2565 Ops/s $\color{#d91a1a}-0.16\%$
test_func_call_cm_runtime[True-compile-overhead] 0.6428ms 0.5527ms 1.8092 KOps/s 1.7905 KOps/s $\color{#35bf28}+1.04\%$
test_vmap_func_call_cm_runtime[eager] 2.9090ms 2.4202ms 413.1848 Ops/s 408.2434 Ops/s $\color{#35bf28}+1.21\%$
test_vmap_func_call_cm_runtime[compile] 3.0464ms 2.6015ms 384.3970 Ops/s 380.2296 Ops/s $\color{#35bf28}+1.10\%$
test_vmap_func_call_cm_runtime[compile-overhead] 3.1016ms 2.6032ms 384.1358 Ops/s 372.2686 Ops/s $\color{#35bf28}+3.19\%$
test_distributed 0.6603ms 0.1515ms 6.6007 KOps/s 6.6104 KOps/s $\color{#d91a1a}-0.15\%$
test_tdmodule 0.3960ms 0.1912ms 5.2296 KOps/s 5.1071 KOps/s $\color{#35bf28}+2.40\%$
test_tdmodule_dispatch 0.4178ms 0.3718ms 2.6899 KOps/s 2.7029 KOps/s $\color{#d91a1a}-0.48\%$
test_tdseq 50.7510μs 27.0012μs 37.0354 KOps/s 37.0147 KOps/s $\color{#35bf28}+0.06\%$
test_tdseq_dispatch 0.2782ms 0.2160ms 4.6293 KOps/s 4.6015 KOps/s $\color{#35bf28}+0.60\%$
test_instantiation_functorch 2.2027ms 2.0838ms 479.8860 Ops/s 478.8459 Ops/s $\color{#35bf28}+0.22\%$
test_exec_functorch 0.2544ms 0.1821ms 5.4904 KOps/s 5.4441 KOps/s $\color{#35bf28}+0.85\%$
test_exec_functional_call 0.2103ms 0.1628ms 6.1409 KOps/s 6.1830 KOps/s $\color{#d91a1a}-0.68\%$
test_exec_td_decorator 0.5032ms 0.2528ms 3.9560 KOps/s 3.9634 KOps/s $\color{#d91a1a}-0.19\%$
test_vmap_mlp_speed_decorator[True-True] 1.2291ms 1.0376ms 963.7218 Ops/s 938.5061 Ops/s $\color{#35bf28}+2.69\%$
test_vmap_mlp_speed_decorator[True-False] 1.2560ms 1.0348ms 966.3460 Ops/s 949.0065 Ops/s $\color{#35bf28}+1.83\%$
test_vmap_mlp_speed_decorator[False-True] 0.9171ms 0.7364ms 1.3579 KOps/s 1.3348 KOps/s $\color{#35bf28}+1.73\%$
test_vmap_mlp_speed_decorator[False-False] 0.9306ms 0.7361ms 1.3585 KOps/s 1.3289 KOps/s $\color{#35bf28}+2.22\%$
test_vmap_transformer_speed_decorator[True-True] 22.3057ms 21.3159ms 46.9133 Ops/s 46.2073 Ops/s $\color{#35bf28}+1.53\%$
test_vmap_transformer_speed_decorator[True-False] 21.3871ms 21.2565ms 47.0444 Ops/s 46.0909 Ops/s $\color{#35bf28}+2.07\%$
test_vmap_transformer_speed_decorator[False-True] 21.0501ms 20.9001ms 47.8467 Ops/s 46.6734 Ops/s $\color{#35bf28}+2.51\%$
test_vmap_transformer_speed_decorator[False-False] 21.5165ms 20.8793ms 47.8944 Ops/s 47.1507 Ops/s $\color{#35bf28}+1.58\%$
test_to_module_speed[True] 1.7569ms 1.6686ms 599.3006 Ops/s 609.6106 Ops/s $\color{#d91a1a}-1.69\%$
test_to_module_speed[False] 1.7476ms 1.6100ms 621.1026 Ops/s 620.3911 Ops/s $\color{#35bf28}+0.11\%$
test_to_module_plain_tensor_speed[None] 1.8593ms 1.7428ms 573.7871 Ops/s 579.5549 Ops/s $\color{#d91a1a}-1.00\%$
test_to_module_plain_tensor_speed[False] 1.7198ms 1.5943ms 627.2311 Ops/s 640.0245 Ops/s $\color{#d91a1a}-2.00\%$
test_to_module_plain_tensor_speed[True] 3.5489ms 3.4419ms 290.5363 Ops/s 292.6954 Ops/s $\color{#d91a1a}-0.74\%$
test_tc_init 65.5410μs 42.2237μs 23.6834 KOps/s 22.9423 KOps/s $\color{#35bf28}+3.23\%$
test_tc_init_tensor_only 36.4710μs 9.3099μs 107.4126 KOps/s 110.8024 KOps/s $\color{#d91a1a}-3.06\%$
test_tc_init_nested 0.1276ms 83.2981μs 12.0051 KOps/s 11.8615 KOps/s $\color{#35bf28}+1.21\%$
test_tc_init_many_fields 38.2410μs 14.9357μs 66.9536 KOps/s 67.5822 KOps/s $\color{#d91a1a}-0.93\%$
test_tc_first_layer_tensor 17.6210μs 1.8216μs 548.9643 KOps/s 551.8025 KOps/s $\color{#d91a1a}-0.51\%$
test_tc_first_layer_tensor_only 2.0343μs 0.4004μs 2.4975 MOps/s 2.5537 MOps/s $\color{#d91a1a}-2.20\%$
test_tc_first_layer_tensor_set 34.5700μs 3.9234μs 254.8805 KOps/s 259.6474 KOps/s $\color{#d91a1a}-1.84\%$
test_tc_first_layer_tensor_only_set 29.1810μs 3.2379μs 308.8441 KOps/s 307.9470 KOps/s $\color{#35bf28}+0.29\%$
test_tc_first_layer_nontensor 28.5510μs 6.1922μs 161.4926 KOps/s 165.9260 KOps/s $\color{#d91a1a}-2.67\%$
test_tc_second_layer_tensor 26.4700μs 4.4259μs 225.9448 KOps/s 229.9201 KOps/s $\color{#d91a1a}-1.73\%$
test_tc_second_layer_nontensor 46.0710μs 8.6870μs 115.1140 KOps/s 117.9550 KOps/s $\color{#d91a1a}-2.41\%$
test_unbind 0.2710s 14.2267ms 70.2905 Ops/s 66.5598 Ops/s $\textbf{\color{#35bf28}+5.60\%}$
test_full_like 4.5188ms 4.3806ms 228.2801 Ops/s 227.8569 Ops/s $\color{#35bf28}+0.19\%$
test_zeros_like 4.5381ms 4.3679ms 228.9445 Ops/s 228.6068 Ops/s $\color{#35bf28}+0.15\%$
test_ones_like 4.5286ms 4.3767ms 228.4809 Ops/s 228.4648 Ops/s $+0.01\%$
test_clone 6.7387ms 6.4623ms 154.7434 Ops/s 155.3426 Ops/s $\color{#d91a1a}-0.39\%$
test_squeeze 87.1620μs 14.1810μs 70.5170 KOps/s 69.7208 KOps/s $\color{#35bf28}+1.14\%$
test_unsqueeze 0.1707ms 0.1138ms 8.7897 KOps/s 9.0870 KOps/s $\color{#d91a1a}-3.27\%$
test_split 0.3803ms 0.1834ms 5.4536 KOps/s 5.4975 KOps/s $\color{#d91a1a}-0.80\%$
test_permute 0.2751ms 0.2089ms 4.7873 KOps/s 4.8145 KOps/s $\color{#d91a1a}-0.56\%$
test_stack 51.9512ms 51.6115ms 19.3755 Ops/s 55.4593 Ops/s $\textbf{\color{#d91a1a}-65.06\%}$
test_cat 51.9074ms 51.5770ms 19.3885 Ops/s 55.7381 Ops/s $\textbf{\color{#d91a1a}-65.22\%}$
test_sequential_tensordict 0.6823ms 0.5726ms 1.7465 KOps/s 1.7236 KOps/s $\color{#35bf28}+1.33\%$
test_sequential_graph_module 0.5167ms 0.1226ms 8.1596 KOps/s 8.0431 KOps/s $\color{#35bf28}+1.45\%$
test_nested_tensordict 0.9476ms 0.8157ms 1.2259 KOps/s 1.2073 KOps/s $\color{#35bf28}+1.54\%$
test_nested_graph_module 0.5577ms 0.1340ms 7.4636 KOps/s 7.4031 KOps/s $\color{#35bf28}+0.82\%$

vmoens and others added 2 commits June 15, 2026 10:57
@vmoens @claude
Merge origin/main into claude/friendly-jang-a55564
d965d1c 
Resolves the modify/delete conflict on test/test_tensordict.py from
#1725 (NonTensor pad / boolean-mask bugfix). That commit added 14 test
methods to the monolith, which this branch had split into test/tensordict/.
The new methods were rerouted into the corresponding split files,
preserving their original classes:
  - 7 test_pad_nontensor* -> test/tensordict/test_generic.py (TestGeneric)
  - 3 test_lazy_*         -> test/tensordict/test_lazy.py (TestLazyStackedTensorDict)
  - test_linked_list_str + 3 test_nontensorstack_* -> test/tensordict/test_nontensor.py (TestNonTensorData)

Verified the split union equals origin/main's test_tensordict.py exactly
(25 classes, 728 methods). The #1725 test_distributed.py change merged
onto its renamed path test/distributed/test_distributed.py.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
@vmoens @claude
[Test] Fix resource paths and a flaky test after the subdir reorg
7610a9c 
Moving test files one level deeper (test/<area>/) broke resources that
were located relative to __file__:

- test/tensorclass/test_tensorclass.py: tensorclass.pyi -> parents[2]
- test/tensordict/test_generic.py: artifacts/ -> Path(__file__).parent.parent
- test/utils/test_setup.py: repo root (version.txt, pyproject.toml) -> parents[2]

Also makes TestLazyStackedTensorDict.test_arithmetic_ops deterministic:
with ragged=True it drew three random middle-dim sizes from randint(1, 5),
which collided ~6% of the time and left the stack non-ragged, so norm() did
not raise. The reorg changed test execution order (hence RNG state) and
surfaced this. Use distinct sizes (2, 3, 4) so the ragged case is always
ragged.

Co-Authored-By: Claude Opus 4.8 (1M context) <noreply@anthropic.com>
@vmoens
vmoens merged commit 78d14a0 into main Jun 15, 2026
71 of 73 checks passed
@vmoens
vmoens deleted the claude/friendly-jang-a55564 branch June 15, 2026 15:00
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

CI CLA Signed This label is managed by the Facebook bot. Authors need to sign the CLA before a PR can be reviewed. Test

Projects

None yet

Milestone

No milestone

Development

Successfully merging this pull request may close these issues.

1 participant

@vmoens