[BugFix] Preserve NonTensor entries in pad; fix NonTensorStack boolean-mask indexing

[vmoens]

Summary

tensordict.pad silently destroyed NonTensor entries: padding a TensorDict containing a NonTensorStack returned the entry as a stack of empty TensorDicts for all positions, valid ones included. Repro:

import torch
from tensordict import NonTensorStack, TensorDict, pad
td = TensorDict({"x": torch.arange(3).unsqueeze(-1), "instr": NonTensorStack("a", "b", "c")}, batch_size=[3])
p = pad(td, [0, 2])
p["instr"]  # before: stack of empty TensorDicts — 'a','b','c' gone;  after: ['a','b','c',None,None]
p["x"]      # correctly zero-padded

Root cause: NonTensorStack is a LazyStackedTensorDict subclass, so it fell into the generic lazy-stack branch and had its leaves filled with the numeric value.

What changed

pad (functional.py) now routes non-tensor entries to a dedicated path. Valid positions keep their original values; each pad slot holds a NonTensorData with data=None (a non-tensor value can't be filled with a numeric constant). Covered cases:

• top-level NonTensorData / NonTensorStack inputs and nested entries
• multi-dimensional padding and negative (cropping) pads
• broadcast NonTensorData (single value over a batch dim)
• unbatched (batch_size=[]) non-tensor metadata is carried through untouched
• inplace=True: NonTensorStack identity is preserved; the lazy-stack stack-dim grow path fills new slots with None for non-tensor leaves
• safe=True pre-flight validates non-tensor pad sizes too

Boolean-mask indexing (_lazy.py) — two related sharp edges surfaced by the same downstream code:

• Masking dim 0 of a 2-D NonTensorStack raised IndexError because a scalar bool inside a tuple index was treated like a scalar integer. Scalar bools now follow numpy new-axis semantics (True -> size-1 dim, False -> size-0 dim) in both __getitem__ and __setitem__.
• Indices selecting nothing (empty slice along the stack dim, all-False mask) raised "items cannot be empty" or produced a wrong batch size ([0, 0, ...]); they now return a correctly-shaped zero-length lazy stack that still interops with torch.cat.

LinkedList (utils.py) printed doubly-wrapped (LinkedList(LinkedList([...]))) because __str__ called super().__str__(), which dispatched back to __repr__. Dropped the redundant __str__.

Tests

Added to test/test_tensordict.py (extends existing pad / lazy-stack / non-tensor coverage):

• test_pad_nontensor, _2dim, _broadcast_data, _inplace, _top_level, _safe, _lazy_stack
• test_lazy_mask_nested_stack, test_lazy_scalar_bool_index, test_lazy_empty_selection_batch_size
• test_nontensorstack_boolean_mask, _2d, _scalar_bool_index, test_linked_list_str

Full test/test_tensordict.py (7716 passed), pad/index/mask subsets, test_compile.py subset, and test_tensorclass.py non-tensor/stack subset all pass.

Downstream

torchrl's collector trajs_per_batch path (torchrl/collectors/utils.py::_traj_emit) carries a workaround that pads non-tensor keys separately; once this lands that workaround can be removed.

🤖 Generated with Claude Code

[github-actions]

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

Total Benchmarks: 344. Improved: $\large\color{#35bf28}14$. Worsened: $\large\color{#d91a1a}12$.

Expand to view detailed results

Name	Max	Mean	Ops	Ops on Repo HEAD	Change
test_collector_preallocated_write[stack_out-cpu-32-2]	0.2117ms	0.1724ms	5.8020 KOps/s	5.7581 KOps/s	$\color{#35bf28}+0.76\%$
test_collector_preallocated_write[stack_out-cpu-32-8]	0.3345ms	0.2974ms	3.3622 KOps/s	3.3492 KOps/s	$\color{#35bf28}+0.39\%$
test_collector_preallocated_write[stack_out-cpu-32-32]	0.9160ms	0.7778ms	1.2858 KOps/s	1.2727 KOps/s	$\color{#35bf28}+1.03\%$
test_collector_preallocated_write[stack_out-cpu-4096-2]	0.4088ms	0.3435ms	2.9111 KOps/s	2.9029 KOps/s	$\color{#35bf28}+0.28\%$
test_collector_preallocated_write[stack_out-cpu-4096-8]	1.3212ms	1.1667ms	857.1306 Ops/s	917.1108 Ops/s	$\textbf{\color{#d91a1a}-6.54\%}$
test_collector_preallocated_write[stack_out-cpu-4096-32]	7.7489ms	7.3001ms	136.9847 Ops/s	135.0006 Ops/s	$\color{#35bf28}+1.47\%$
test_collector_preallocated_write[stack_out-cuda-32-2]	0.3317ms	0.2521ms	3.9671 KOps/s	3.9980 KOps/s	$\color{#d91a1a}-0.77\%$
test_collector_preallocated_write[stack_out-cuda-32-8]	0.4360ms	0.3745ms	2.6706 KOps/s	2.6646 KOps/s	$\color{#35bf28}+0.23\%$
test_collector_preallocated_write[stack_out-cuda-32-32]	0.9138ms	0.8264ms	1.2100 KOps/s	1.2087 KOps/s	$\color{#35bf28}+0.11\%$
test_collector_preallocated_write[stack_out-cuda-4096-2]	0.3922ms	0.2520ms	3.9682 KOps/s	3.8512 KOps/s	$\color{#35bf28}+3.04\%$
test_collector_preallocated_write[stack_out-cuda-4096-8]	0.5216ms	0.3757ms	2.6619 KOps/s	2.6620 KOps/s	$-0.00\%$
test_collector_preallocated_write[stack_out-cuda-4096-32]	0.9744ms	0.8874ms	1.1269 KOps/s	1.1241 KOps/s	$\color{#35bf28}+0.25\%$
test_collector_preallocated_write[set_at_loop-cpu-32-2]	0.2936ms	0.2192ms	4.5629 KOps/s	4.5541 KOps/s	$\color{#35bf28}+0.19\%$
test_collector_preallocated_write[set_at_loop-cpu-32-8]	0.9879ms	0.8416ms	1.1882 KOps/s	1.1867 KOps/s	$\color{#35bf28}+0.13\%$
test_collector_preallocated_write[set_at_loop-cpu-32-32]	3.4832ms	3.3682ms	296.8935 Ops/s	298.1007 Ops/s	$\color{#d91a1a}-0.40\%$
test_collector_preallocated_write[set_at_loop-cpu-4096-2]	0.7402ms	0.3704ms	2.6996 KOps/s	2.5910 KOps/s	$\color{#35bf28}+4.19\%$
test_collector_preallocated_write[set_at_loop-cpu-4096-8]	2.1846ms	1.7596ms	568.3058 Ops/s	595.9025 Ops/s	$\color{#d91a1a}-4.63\%$
test_collector_preallocated_write[set_at_loop-cpu-4096-32]	10.6057ms	10.3662ms	96.4673 Ops/s	96.6172 Ops/s	$\color{#d91a1a}-0.16\%$
test_collector_preallocated_write[set_at_loop-cuda-32-2]	0.4489ms	0.3963ms	2.5233 KOps/s	2.4966 KOps/s	$\color{#35bf28}+1.07\%$
test_collector_preallocated_write[set_at_loop-cuda-32-8]	1.6275ms	1.4741ms	678.3958 Ops/s	670.6027 Ops/s	$\color{#35bf28}+1.16\%$
test_collector_preallocated_write[set_at_loop-cuda-32-32]	5.9445ms	5.7999ms	172.4174 Ops/s	170.7234 Ops/s	$\color{#35bf28}+0.99\%$
test_collector_preallocated_write[set_at_loop-cuda-4096-2]	0.4943ms	0.3972ms	2.5175 KOps/s	2.5014 KOps/s	$\color{#35bf28}+0.64\%$
test_collector_preallocated_write[set_at_loop-cuda-4096-8]	1.5952ms	1.4940ms	669.3324 Ops/s	669.0193 Ops/s	$\color{#35bf28}+0.05\%$
test_collector_preallocated_write[set_at_loop-cuda-4096-32]	5.9102ms	5.8520ms	170.8824 Ops/s	170.5918 Ops/s	$\color{#35bf28}+0.17\%$
test_collector_preallocated_write[copy_at-cpu-32-2]	0.3114ms	0.1747ms	5.7225 KOps/s	5.7316 KOps/s	$\color{#d91a1a}-0.16\%$
test_collector_preallocated_write[copy_at-cpu-32-8]	0.7608ms	0.6883ms	1.4528 KOps/s	1.4726 KOps/s	$\color{#d91a1a}-1.35\%$
test_collector_preallocated_write[copy_at-cpu-32-32]	2.8156ms	2.7196ms	367.6950 Ops/s	367.4117 Ops/s	$\color{#35bf28}+0.08\%$
test_collector_preallocated_write[copy_at-cpu-4096-2]	0.6508ms	0.3143ms	3.1816 KOps/s	3.0869 KOps/s	$\color{#35bf28}+3.07\%$
test_collector_preallocated_write[copy_at-cpu-4096-8]	1.7604ms	1.5254ms	655.5589 Ops/s	688.0825 Ops/s	$\color{#d91a1a}-4.73\%$
test_collector_preallocated_write[copy_at-cpu-4096-32]	9.7669ms	9.5933ms	104.2399 Ops/s	103.5920 Ops/s	$\color{#35bf28}+0.63\%$
test_collector_preallocated_write[copy_at-cuda-32-2]	0.4438ms	0.3303ms	3.0273 KOps/s	3.0362 KOps/s	$\color{#d91a1a}-0.29\%$
test_collector_preallocated_write[copy_at-cuda-32-8]	1.3482ms	1.2361ms	809.0062 Ops/s	820.4173 Ops/s	$\color{#d91a1a}-1.39\%$
test_collector_preallocated_write[copy_at-cuda-32-32]	4.9990ms	4.8226ms	207.3571 Ops/s	209.6160 Ops/s	$\color{#d91a1a}-1.08\%$
test_collector_preallocated_write[copy_at-cuda-4096-2]	0.4069ms	0.3322ms	3.0105 KOps/s	3.0309 KOps/s	$\color{#d91a1a}-0.67\%$
test_collector_preallocated_write[copy_at-cuda-4096-8]	1.3026ms	1.2380ms	807.7456 Ops/s	813.7471 Ops/s	$\color{#d91a1a}-0.74\%$
test_collector_preallocated_write[copy_at-cuda-4096-32]	4.9588ms	4.8429ms	206.4889 Ops/s	207.6582 Ops/s	$\color{#d91a1a}-0.56\%$
test_collector_preallocated_write[manual_leaf_copy-cpu-32-2]	0.1623ms	0.1250ms	7.9974 KOps/s	8.1758 KOps/s	$\color{#d91a1a}-2.18\%$
test_collector_preallocated_write[manual_leaf_copy-cpu-32-8]	0.5352ms	0.4895ms	2.0428 KOps/s	2.1072 KOps/s	$\color{#d91a1a}-3.05\%$
test_collector_preallocated_write[manual_leaf_copy-cpu-32-32]	2.0150ms	1.9411ms	515.1744 Ops/s	530.1983 Ops/s	$\color{#d91a1a}-2.83\%$
test_collector_preallocated_write[manual_leaf_copy-cpu-4096-2]	0.3214ms	0.2616ms	3.8226 KOps/s	3.8407 KOps/s	$\color{#d91a1a}-0.47\%$
test_collector_preallocated_write[manual_leaf_copy-cpu-4096-8]	1.7983ms	1.3209ms	757.0461 Ops/s	756.1471 Ops/s	$\color{#35bf28}+0.12\%$
test_collector_preallocated_write[manual_leaf_copy-cpu-4096-32]	9.1243ms	8.6867ms	115.1190 Ops/s	115.4082 Ops/s	$\color{#d91a1a}-0.25\%$
test_collector_preallocated_write[manual_leaf_copy-cuda-32-2]	0.4159ms	0.2881ms	3.4713 KOps/s	3.6249 KOps/s	$\color{#d91a1a}-4.24\%$
test_collector_preallocated_write[manual_leaf_copy-cuda-32-8]	1.3376ms	1.0629ms	940.8450 Ops/s	968.8408 Ops/s	$\color{#d91a1a}-2.89\%$
test_collector_preallocated_write[manual_leaf_copy-cuda-32-32]	4.4266ms	4.1724ms	239.6685 Ops/s	247.7170 Ops/s	$\color{#d91a1a}-3.25\%$
test_collector_preallocated_write[manual_leaf_copy-cuda-4096-2]	0.3520ms	0.2870ms	3.4843 KOps/s	3.6273 KOps/s	$\color{#d91a1a}-3.94\%$
test_collector_preallocated_write[manual_leaf_copy-cuda-4096-8]	1.2169ms	1.0612ms	942.3415 Ops/s	971.2180 Ops/s	$\color{#d91a1a}-2.97\%$
test_collector_preallocated_write[manual_leaf_copy-cuda-4096-32]	4.3314ms	4.1755ms	239.4945 Ops/s	246.1724 Ops/s	$\color{#d91a1a}-2.71\%$
test_wide_nested_preallocated_write[stack_out-cpu-128-128]	16.3758ms	16.2645ms	61.4835 Ops/s	61.4138 Ops/s	$\color{#35bf28}+0.11\%$
test_wide_nested_preallocated_write[stack_out-cpu-128-256]	33.0329ms	32.8961ms	30.3987 Ops/s	30.3972 Ops/s	$+0.01\%$
test_wide_nested_preallocated_write[stack_out-cpu-4096-128]	85.4823ms	84.7651ms	11.7973 Ops/s	11.6748 Ops/s	$\color{#35bf28}+1.05\%$
test_wide_nested_preallocated_write[stack_out-cpu-4096-256]	0.2059s	0.2036s	4.9120 Ops/s	5.0980 Ops/s	$\color{#d91a1a}-3.65\%$
test_wide_nested_preallocated_write[stack_out-cuda-128-128]	13.8812ms	13.7823ms	72.5569 Ops/s	72.1432 Ops/s	$\color{#35bf28}+0.57\%$
test_wide_nested_preallocated_write[stack_out-cuda-128-256]	27.1731ms	26.9063ms	37.1660 Ops/s	36.8503 Ops/s	$\color{#35bf28}+0.86\%$
test_wide_nested_preallocated_write[stack_out-cuda-4096-128]	14.1683ms	13.9338ms	71.7679 Ops/s	71.2383 Ops/s	$\color{#35bf28}+0.74\%$
test_wide_nested_preallocated_write[stack_out-cuda-4096-256]	27.4991ms	27.3148ms	36.6102 Ops/s	36.1742 Ops/s	$\color{#35bf28}+1.21\%$
test_wide_nested_preallocated_write[set_at_loop-cpu-128-128]	54.4846ms	53.0131ms	18.8633 Ops/s	18.9364 Ops/s	$\color{#d91a1a}-0.39\%$
test_wide_nested_preallocated_write[set_at_loop-cpu-128-256]	0.1074s	0.1064s	9.3949 Ops/s	9.3904 Ops/s	$\color{#35bf28}+0.05\%$
test_wide_nested_preallocated_write[set_at_loop-cpu-4096-128]	0.1536s	0.1487s	6.7271 Ops/s	6.8877 Ops/s	$\color{#d91a1a}-2.33\%$
test_wide_nested_preallocated_write[set_at_loop-cpu-4096-256]	0.3319s	0.3300s	3.0305 Ops/s	3.1415 Ops/s	$\color{#d91a1a}-3.53\%$
test_wide_nested_preallocated_write[set_at_loop-cuda-128-128]	76.5883ms	75.3276ms	13.2753 Ops/s	13.2168 Ops/s	$\color{#35bf28}+0.44\%$
test_wide_nested_preallocated_write[set_at_loop-cuda-128-256]	0.1507s	0.1491s	6.7047 Ops/s	6.6467 Ops/s	$\color{#35bf28}+0.87\%$
test_wide_nested_preallocated_write[set_at_loop-cuda-4096-128]	76.0996ms	75.2126ms	13.2956 Ops/s	13.1616 Ops/s	$\color{#35bf28}+1.02\%$
test_wide_nested_preallocated_write[set_at_loop-cuda-4096-256]	0.1532s	0.1521s	6.5739 Ops/s	6.5655 Ops/s	$\color{#35bf28}+0.13\%$
test_wide_nested_preallocated_write[copy_at-cpu-128-128]	31.2590ms	30.9958ms	32.2625 Ops/s	32.9139 Ops/s	$\color{#d91a1a}-1.98\%$
test_wide_nested_preallocated_write[copy_at-cpu-128-256]	63.2221ms	62.6315ms	15.9664 Ops/s	16.2892 Ops/s	$\color{#d91a1a}-1.98\%$
test_wide_nested_preallocated_write[copy_at-cpu-4096-128]	0.1228s	0.1213s	8.2438 Ops/s	8.5091 Ops/s	$\color{#d91a1a}-3.12\%$
test_wide_nested_preallocated_write[copy_at-cpu-4096-256]	0.2745s	0.2722s	3.6732 Ops/s	3.8113 Ops/s	$\color{#d91a1a}-3.62\%$
test_wide_nested_preallocated_write[copy_at-cuda-128-128]	44.4975ms	43.9895ms	22.7327 Ops/s	22.9544 Ops/s	$\color{#d91a1a}-0.97\%$
test_wide_nested_preallocated_write[copy_at-cuda-128-256]	89.1768ms	88.3416ms	11.3197 Ops/s	11.5068 Ops/s	$\color{#d91a1a}-1.63\%$
test_wide_nested_preallocated_write[copy_at-cuda-4096-128]	44.5177ms	44.2921ms	22.5774 Ops/s	22.7570 Ops/s	$\color{#d91a1a}-0.79\%$
test_wide_nested_preallocated_write[copy_at-cuda-4096-256]	88.9564ms	88.6202ms	11.2841 Ops/s	11.3900 Ops/s	$\color{#d91a1a}-0.93\%$
test_wide_nested_preallocated_write[manual_leaf_copy-cpu-128-128]	21.7689ms	21.2417ms	47.0773 Ops/s	48.9619 Ops/s	$\color{#d91a1a}-3.85\%$
test_wide_nested_preallocated_write[manual_leaf_copy-cpu-128-256]	43.4922ms	43.3208ms	23.0836 Ops/s	23.9930 Ops/s	$\color{#d91a1a}-3.79\%$
test_wide_nested_preallocated_write[manual_leaf_copy-cpu-4096-128]	0.1117s	0.1100s	9.0931 Ops/s	9.2588 Ops/s	$\color{#d91a1a}-1.79\%$
test_wide_nested_preallocated_write[manual_leaf_copy-cpu-4096-256]	0.2522s	0.2497s	4.0051 Ops/s	4.2074 Ops/s	$\color{#d91a1a}-4.81\%$
test_wide_nested_preallocated_write[manual_leaf_copy-cuda-128-128]	38.5233ms	38.1312ms	26.2252 Ops/s	27.1444 Ops/s	$\color{#d91a1a}-3.39\%$
test_wide_nested_preallocated_write[manual_leaf_copy-cuda-128-256]	76.6262ms	75.7859ms	13.1951 Ops/s	13.5231 Ops/s	$\color{#d91a1a}-2.43\%$
test_wide_nested_preallocated_write[manual_leaf_copy-cuda-4096-128]	39.2507ms	37.9618ms	26.3423 Ops/s	27.1251 Ops/s	$\color{#d91a1a}-2.89\%$
test_wide_nested_preallocated_write[manual_leaf_copy-cuda-4096-256]	76.0720ms	75.3934ms	13.2638 Ops/s	13.5426 Ops/s	$\color{#d91a1a}-2.06\%$
test_plain_set_nested	44.6610μs	15.3972μs	64.9469 KOps/s	64.6605 KOps/s	$\color{#35bf28}+0.44\%$
test_plain_set_stack_nested	47.2310μs	15.3968μs	64.9486 KOps/s	65.9646 KOps/s	$\color{#d91a1a}-1.54\%$
test_plain_set_nested_inplace	44.2710μs	16.9536μs	58.9844 KOps/s	58.3620 KOps/s	$\color{#35bf28}+1.07\%$
test_plain_set_stack_nested_inplace	49.7020μs	16.6319μs	60.1255 KOps/s	60.2335 KOps/s	$\color{#d91a1a}-0.18\%$
test_items	24.2900μs	6.0555μs	165.1384 KOps/s	163.8474 KOps/s	$\color{#35bf28}+0.79\%$
test_items_nested	0.5215ms	0.4656ms	2.1479 KOps/s	2.1199 KOps/s	$\color{#35bf28}+1.32\%$
test_items_nested_locked	0.5505ms	0.4713ms	2.1216 KOps/s	2.1148 KOps/s	$\color{#35bf28}+0.32\%$
test_items_nested_leaf	0.1508ms	96.4586μs	10.3671 KOps/s	10.1425 KOps/s	$\color{#35bf28}+2.22\%$
test_items_stack_nested	0.6044ms	0.4696ms	2.1293 KOps/s	2.1152 KOps/s	$\color{#35bf28}+0.66\%$
test_items_stack_nested_leaf	0.1501ms	97.3910μs	10.2679 KOps/s	10.2354 KOps/s	$\color{#35bf28}+0.32\%$
test_items_stack_nested_locked	0.5546ms	0.4718ms	2.1197 KOps/s	2.1006 KOps/s	$\color{#35bf28}+0.91\%$
test_keys	25.8300μs	4.2273μs	236.5593 KOps/s	236.6119 KOps/s	$\color{#d91a1a}-0.02\%$
test_keys_nested	0.2042ms	0.1303ms	7.6772 KOps/s	7.6473 KOps/s	$\color{#35bf28}+0.39\%$
test_keys_nested_locked	0.7511ms	0.1393ms	7.1768 KOps/s	7.1164 KOps/s	$\color{#35bf28}+0.85\%$
test_keys_nested_leaf	0.1552ms	0.1215ms	8.2337 KOps/s	8.2110 KOps/s	$\color{#35bf28}+0.28\%$
test_keys_stack_nested	0.1867ms	0.1298ms	7.7051 KOps/s	7.6358 KOps/s	$\color{#35bf28}+0.91\%$
test_keys_stack_nested_leaf	0.1717ms	0.1201ms	8.3251 KOps/s	8.1727 KOps/s	$\color{#35bf28}+1.86\%$
test_keys_stack_nested_locked	0.1811ms	0.1380ms	7.2457 KOps/s	7.1294 KOps/s	$\color{#35bf28}+1.63\%$
test_values	5.8362μs	1.0302μs	970.7052 KOps/s	963.5396 KOps/s	$\color{#35bf28}+0.74\%$
test_values_nested	81.7620μs	52.7374μs	18.9619 KOps/s	18.9956 KOps/s	$\color{#d91a1a}-0.18\%$
test_values_nested_locked	86.4220μs	55.7672μs	17.9317 KOps/s	17.8047 KOps/s	$\color{#35bf28}+0.71\%$
test_values_nested_leaf	96.4020μs	60.3534μs	16.5691 KOps/s	16.6273 KOps/s	$\color{#d91a1a}-0.35\%$
test_values_stack_nested	85.5820μs	52.2890μs	19.1245 KOps/s	19.0884 KOps/s	$\color{#35bf28}+0.19\%$
test_values_stack_nested_leaf	91.5520μs	59.8242μs	16.7156 KOps/s	16.6537 KOps/s	$\color{#35bf28}+0.37\%$
test_values_stack_nested_locked	91.3120μs	55.6171μs	17.9801 KOps/s	17.7224 KOps/s	$\color{#35bf28}+1.45\%$
test_membership	4.9318μs	0.8324μs	1.2013 MOps/s	1.1682 MOps/s	$\color{#35bf28}+2.83\%$
test_membership_nested	37.9710μs	2.9851μs	334.9926 KOps/s	329.9386 KOps/s	$\color{#35bf28}+1.53\%$
test_membership_nested_leaf	26.5410μs	2.9646μs	337.3113 KOps/s	333.7542 KOps/s	$\color{#35bf28}+1.07\%$
test_membership_stacked_nested	29.3000μs	2.8964μs	345.2552 KOps/s	344.3658 KOps/s	$\color{#35bf28}+0.26\%$
test_membership_stacked_nested_leaf	28.8710μs	2.8683μs	348.6383 KOps/s	348.7552 KOps/s	$\color{#d91a1a}-0.03\%$
test_membership_nested_last	28.8510μs	4.4338μs	225.5381 KOps/s	222.0697 KOps/s	$\color{#35bf28}+1.56\%$
test_membership_nested_leaf_last	35.8810μs	4.4000μs	227.2719 KOps/s	223.5712 KOps/s	$\color{#35bf28}+1.66\%$
test_membership_stacked_nested_last	39.9510μs	4.3723μs	228.7140 KOps/s	229.4365 KOps/s	$\color{#d91a1a}-0.31\%$
test_membership_stacked_nested_leaf_last	36.5410μs	4.3446μs	230.1706 KOps/s	232.2457 KOps/s	$\color{#d91a1a}-0.89\%$
test_nested_getleaf	51.2210μs	22.0209μs	45.4113 KOps/s	45.7203 KOps/s	$\color{#d91a1a}-0.68\%$
test_nested_get	45.9110μs	20.6503μs	48.4255 KOps/s	48.0756 KOps/s	$\color{#35bf28}+0.73\%$
test_stacked_getleaf	48.3110μs	22.1313μs	45.1850 KOps/s	46.2523 KOps/s	$\color{#d91a1a}-2.31\%$
test_stacked_get	45.0110μs	20.6959μs	48.3187 KOps/s	48.5306 KOps/s	$\color{#d91a1a}-0.44\%$
test_nested_getitemleaf	50.7810μs	22.4863μs	44.4716 KOps/s	44.2023 KOps/s	$\color{#35bf28}+0.61\%$
test_nested_getitem	48.8410μs	21.0780μs	47.4427 KOps/s	46.9830 KOps/s	$\color{#35bf28}+0.98\%$
test_stacked_getitemleaf	48.0110μs	22.1647μs	45.1167 KOps/s	45.4969 KOps/s	$\color{#d91a1a}-0.84\%$
test_stacked_getitem	57.5320μs	21.4331μs	46.6567 KOps/s	46.9103 KOps/s	$\color{#d91a1a}-0.54\%$
test_lock_nested	7.7425ms	0.5164ms	1.9366 KOps/s	1.9665 KOps/s	$\color{#d91a1a}-1.52\%$
test_lock_stack_nested	0.6363ms	0.5173ms	1.9332 KOps/s	1.9570 KOps/s	$\color{#d91a1a}-1.22\%$
test_unlock_nested	0.4849ms	0.4029ms	2.4820 KOps/s	2.4813 KOps/s	$\color{#35bf28}+0.03\%$
test_unlock_stack_nested	0.5354ms	0.4087ms	2.4469 KOps/s	2.4644 KOps/s	$\color{#d91a1a}-0.71\%$
test_flatten_speed	0.2084ms	0.1224ms	8.1717 KOps/s	8.1771 KOps/s	$\color{#d91a1a}-0.07\%$
test_unflatten_speed	0.6879ms	0.5753ms	1.7384 KOps/s	1.7259 KOps/s	$\color{#35bf28}+0.72\%$
test_common_ops	0.1007s	1.8073ms	553.3102 Ops/s	640.0617 Ops/s	$\textbf{\color{#d91a1a}-13.55\%}$
test_creation	83.6020μs	3.2172μs	310.8328 KOps/s	313.5400 KOps/s	$\color{#d91a1a}-0.86\%$
test_creation_empty	38.2810μs	7.1104μs	140.6391 KOps/s	140.7905 KOps/s	$\color{#d91a1a}-0.11\%$
test_creation_nested_1	36.4610μs	11.6155μs	86.0920 KOps/s	85.3876 KOps/s	$\color{#35bf28}+0.82\%$
test_creation_nested_2	42.4010μs	13.7738μs	72.6014 KOps/s	72.6969 KOps/s	$\color{#d91a1a}-0.13\%$
test_creation_many_keys[10]	52.0410μs	21.1589μs	47.2615 KOps/s	46.4264 KOps/s	$\color{#35bf28}+1.80\%$
test_creation_many_keys[50]	0.1230ms	90.5307μs	11.0460 KOps/s	10.7434 KOps/s	$\color{#35bf28}+2.82\%$
test_creation_many_keys[100]	0.2748ms	0.1775ms	5.6345 KOps/s	5.5328 KOps/s	$\color{#35bf28}+1.84\%$
test_creation_nested_many_keys[10]	91.3820μs	45.8335μs	21.8181 KOps/s	21.5796 KOps/s	$\color{#35bf28}+1.11\%$
test_creation_nested_many_keys[50]	0.2479ms	0.1860ms	5.3762 KOps/s	5.2831 KOps/s	$\color{#35bf28}+1.76\%$
test_clone	31.8310μs	13.7387μs	72.7873 KOps/s	73.4010 KOps/s	$\color{#d91a1a}-0.84\%$
test_getitem[int]	1.5926ms	15.5870μs	64.1562 KOps/s	64.2830 KOps/s	$\color{#d91a1a}-0.20\%$
test_getitem[slice_int]	0.1410ms	25.1863μs	39.7042 KOps/s	40.6153 KOps/s	$\color{#d91a1a}-2.24\%$
test_getitem[range]	0.1932ms	65.9495μs	15.1631 KOps/s	15.2953 KOps/s	$\color{#d91a1a}-0.86\%$
test_getitem[tuple]	0.1441ms	25.1576μs	39.7494 KOps/s	40.8242 KOps/s	$\color{#d91a1a}-2.63\%$
test_getitem[list]	0.1896ms	60.3493μs	16.5702 KOps/s	16.3609 KOps/s	$\color{#35bf28}+1.28\%$
test_setitem_dim[int]	51.1810μs	27.0479μs	36.9715 KOps/s	37.9085 KOps/s	$\color{#d91a1a}-2.47\%$
test_setitem_dim[slice_int]	66.5610μs	43.9522μs	22.7520 KOps/s	23.0041 KOps/s	$\color{#d91a1a}-1.10\%$
test_setitem_dim[range]	0.1333ms	98.9461μs	10.1065 KOps/s	10.1657 KOps/s	$\color{#d91a1a}-0.58\%$
test_setitem_dim[tuple]	83.8220μs	42.5886μs	23.4804 KOps/s	24.7570 KOps/s	$\textbf{\color{#d91a1a}-5.16\%}$
test_setitem	47.9410μs	18.3918μs	54.3722 KOps/s	55.6572 KOps/s	$\color{#d91a1a}-2.31\%$
test_set	45.6110μs	17.3823μs	57.5296 KOps/s	58.5769 KOps/s	$\color{#d91a1a}-1.79\%$
test_set_shared	0.5565ms	0.2179ms	4.5890 KOps/s	4.7719 KOps/s	$\color{#d91a1a}-3.83\%$
test_update	0.3360ms	22.9790μs	43.5180 KOps/s	44.9900 KOps/s	$\color{#d91a1a}-3.27\%$
test_update_nested	68.7110μs	34.8105μs	28.7270 KOps/s	29.0788 KOps/s	$\color{#d91a1a}-1.21\%$
test_update__nested	0.4644ms	35.2430μs	28.3744 KOps/s	29.2564 KOps/s	$\color{#d91a1a}-3.01\%$
test_set_nested	62.0910μs	19.4754μs	51.3467 KOps/s	51.7322 KOps/s	$\color{#d91a1a}-0.75\%$
test_set_nested_new	63.0710μs	24.3138μs	41.1289 KOps/s	41.2670 KOps/s	$\color{#d91a1a}-0.33\%$
test_select	74.1620μs	41.5445μs	24.0705 KOps/s	24.6612 KOps/s	$\color{#d91a1a}-2.40\%$
test_select_nested	0.1036ms	74.3451μs	13.4508 KOps/s	13.3258 KOps/s	$\color{#35bf28}+0.94\%$
test_exclude_nested	0.1221ms	91.8367μs	10.8889 KOps/s	10.9105 KOps/s	$\color{#d91a1a}-0.20\%$
test_empty[True]	0.4836ms	0.3985ms	2.5092 KOps/s	2.4715 KOps/s	$\color{#35bf28}+1.53\%$
test_empty[False]	6.4850μs	1.3781μs	725.6532 KOps/s	719.2747 KOps/s	$\color{#35bf28}+0.89\%$
test_to	97.1420μs	70.6528μs	14.1537 KOps/s	14.2667 KOps/s	$\color{#d91a1a}-0.79\%$
test_to_nonblocking	0.1220ms	71.7530μs	13.9367 KOps/s	14.3639 KOps/s	$\color{#d91a1a}-2.97\%$
test_unbind_speed	0.4261ms	0.3467ms	2.8844 KOps/s	2.9455 KOps/s	$\color{#d91a1a}-2.07\%$
test_unbind_speed_stack0	0.4003ms	0.3419ms	2.9252 KOps/s	2.9722 KOps/s	$\color{#d91a1a}-1.58\%$
test_unbind_speed_stack1	0.1146s	0.8663ms	1.1543 KOps/s	1.1615 KOps/s	$\color{#d91a1a}-0.62\%$
test_split	0.1141s	1.2921ms	773.9196 Ops/s	718.8363 Ops/s	$\textbf{\color{#35bf28}+7.66\%}$
test_chunk	1.2821ms	1.0931ms	914.8215 Ops/s	910.5765 Ops/s	$\color{#35bf28}+0.47\%$
test_to_cpu_blocking	0.1340s	22.6305ms	44.1882 Ops/s	49.3516 Ops/s	$\textbf{\color{#d91a1a}-10.46\%}$
test_to_cpu_global_sync	12.2369ms	12.0240ms	83.1668 Ops/s	83.9766 Ops/s	$\color{#d91a1a}-0.96\%$
test_to_cpu_event_sync	13.3856ms	13.0495ms	76.6313 Ops/s	68.8217 Ops/s	$\textbf{\color{#35bf28}+11.35\%}$
test_to_cpu_default	13.5156ms	13.0382ms	76.6977 Ops/s	77.4831 Ops/s	$\color{#d91a1a}-1.01\%$
test_consolidate[False-None]	4.5798ms	4.3102ms	232.0057 Ops/s	235.3607 Ops/s	$\color{#d91a1a}-1.43\%$
test_consolidate[default-None]	3.6872ms	2.2522ms	444.0159 Ops/s	456.4075 Ops/s	$\color{#d91a1a}-2.72\%$
test_consolidate[reduce-overhead-None]	3.6117ms	2.1450ms	466.1949 Ops/s	468.1262 Ops/s	$\color{#d91a1a}-0.41\%$
test_consolidate_njt[False-None]	8.9044ms	8.7140ms	114.7576 Ops/s	114.3064 Ops/s	$\color{#35bf28}+0.39\%$
test_to[False-False-None]	2.2954ms	2.2181ms	450.8427 Ops/s	456.4907 Ops/s	$\color{#d91a1a}-1.24\%$
test_to[True-False-None]	2.2926ms	1.9656ms	508.7588 Ops/s	527.6962 Ops/s	$\color{#d91a1a}-3.59\%$
test_to[within-False-None]	6.4539ms	6.3448ms	157.6095 Ops/s	159.5672 Ops/s	$\color{#d91a1a}-1.23\%$
test_to[True-default-None]	11.0441ms	10.8848ms	91.8711 Ops/s	90.7943 Ops/s	$\color{#35bf28}+1.19\%$
test_to_njt[False-False-None]	8.7744ms	8.6958ms	114.9984 Ops/s	114.9898 Ops/s	$+0.01\%$
test_to_njt[True-False-None]	7.1306ms	7.0000ms	142.8567 Ops/s	145.1472 Ops/s	$\color{#d91a1a}-1.58\%$
test_to_njt[within-False-None]	15.8213ms	15.6536ms	63.8831 Ops/s	64.6138 Ops/s	$\color{#d91a1a}-1.13\%$
test_creation[device0]	0.3975ms	0.1187ms	8.4232 KOps/s	8.2401 KOps/s	$\color{#35bf28}+2.22\%$
test_creation_from_tensor	0.4498ms	0.1142ms	8.7577 KOps/s	8.4914 KOps/s	$\color{#35bf28}+3.14\%$
test_add_one[memmap_tensor0]	0.4115ms	7.0084μs	142.6861 KOps/s	146.9257 KOps/s	$\color{#d91a1a}-2.89\%$
test_contiguous[memmap_tensor0]	20.9010μs	0.6575μs	1.5208 MOps/s	2.2239 MOps/s	$\textbf{\color{#d91a1a}-31.61\%}$
test_stack[memmap_tensor0]	99.2820μs	4.4404μs	225.2068 KOps/s	235.7303 KOps/s	$\color{#d91a1a}-4.46\%$
test_memmaptd_index	1.2617ms	0.2939ms	3.4028 KOps/s	3.6092 KOps/s	$\textbf{\color{#d91a1a}-5.72\%}$
test_memmaptd_index_astensor	0.5553ms	0.3972ms	2.5176 KOps/s	2.6183 KOps/s	$\color{#d91a1a}-3.84\%$
test_memmaptd_index_op	0.8100ms	0.6634ms	1.5074 KOps/s	1.5387 KOps/s	$\color{#d91a1a}-2.04\%$
test_serialize_model	0.1379s	0.1347s	7.4225 Ops/s	7.4617 Ops/s	$\color{#d91a1a}-0.53\%$
test_serialize_model_pickle	1.3680s	1.2141s	0.8237 Ops/s	0.8264 Ops/s	$\color{#d91a1a}-0.33\%$
test_serialize_weights	0.1377s	0.1342s	7.4530 Ops/s	7.4020 Ops/s	$\color{#35bf28}+0.69\%$
test_serialize_weights_returnearly	85.8874ms	65.5820ms	15.2481 Ops/s	10.7557 Ops/s	$\textbf{\color{#35bf28}+41.77\%}$
test_serialize_weights_pickle	1.3686s	1.2136s	0.8240 Ops/s	0.8227 Ops/s	$\color{#35bf28}+0.15\%$
test_reshape_pytree	0.2077ms	33.8304μs	29.5592 KOps/s	30.7834 KOps/s	$\color{#d91a1a}-3.98\%$
test_reshape_td	71.2720μs	45.6749μs	21.8939 KOps/s	22.0019 KOps/s	$\color{#d91a1a}-0.49\%$
test_view_pytree	0.2229ms	33.1542μs	30.1621 KOps/s	30.9716 KOps/s	$\color{#d91a1a}-2.61\%$
test_view_td	86.6820μs	55.7796μs	17.9277 KOps/s	18.1816 KOps/s	$\color{#d91a1a}-1.40\%$
test_unbind_pytree	0.2394ms	37.9088μs	26.3791 KOps/s	26.7411 KOps/s	$\color{#d91a1a}-1.35\%$
test_unbind_td	0.1546ms	51.8628μs	19.2816 KOps/s	19.5777 KOps/s	$\color{#d91a1a}-1.51\%$
test_split_pytree	0.2467ms	43.8532μs	22.8034 KOps/s	23.0069 KOps/s	$\color{#d91a1a}-0.88\%$
test_split_td	0.2301ms	67.0932μs	14.9046 KOps/s	15.3398 KOps/s	$\color{#d91a1a}-2.84\%$
test_add_pytree	0.2432ms	43.7862μs	22.8382 KOps/s	23.2633 KOps/s	$\color{#d91a1a}-1.83\%$
test_add_td	96.0520μs	60.6558μs	16.4865 KOps/s	17.7250 KOps/s	$\textbf{\color{#d91a1a}-6.99\%}$
test_compile_add_one_nested[tensordict-compile]	0.5687ms	0.2435ms	4.1065 KOps/s	4.1722 KOps/s	$\color{#d91a1a}-1.58\%$
test_compile_add_one_nested[tensordict-eager]	0.6202ms	0.2039ms	4.9033 KOps/s	5.0253 KOps/s	$\color{#d91a1a}-2.43\%$
test_compile_add_one_nested[pytree-compile]	0.5634ms	0.1400ms	7.1406 KOps/s	7.1555 KOps/s	$\color{#d91a1a}-0.21\%$
test_compile_add_one_nested[pytree-eager]	0.6260ms	0.1853ms	5.3962 KOps/s	5.3908 KOps/s	$\color{#35bf28}+0.10\%$
test_compile_copy_nested[tensordict-compile]	0.3010ms	20.3520μs	49.1353 KOps/s	65.9871 KOps/s	$\textbf{\color{#d91a1a}-25.54\%}$
test_compile_copy_nested[tensordict-eager]	0.5811ms	53.7297μs	18.6117 KOps/s	18.3632 KOps/s	$\color{#35bf28}+1.35\%$
test_compile_copy_nested[pytree-compile]	0.1211ms	16.1878μs	61.7749 KOps/s	62.9370 KOps/s	$\color{#d91a1a}-1.85\%$
test_compile_copy_nested[pytree-eager]	0.4933ms	69.7199μs	14.3431 KOps/s	14.5740 KOps/s	$\color{#d91a1a}-1.58\%$
test_compile_add_one_flat[tensordict-compile]	0.3367ms	0.2532ms	3.9495 KOps/s	3.8281 KOps/s	$\color{#35bf28}+3.17\%$
test_compile_add_one_flat[tensordict-eager]	0.7179ms	0.2830ms	3.5341 KOps/s	3.5262 KOps/s	$\color{#35bf28}+0.23\%$
test_compile_add_one_flat[tensorclass-compile]	0.6361ms	0.1911ms	5.2326 KOps/s	5.1877 KOps/s	$\color{#35bf28}+0.87\%$
test_compile_add_one_flat[tensorclass-eager]	0.5071ms	76.9734μs	12.9915 KOps/s	13.0035 KOps/s	$\color{#d91a1a}-0.09\%$
test_compile_add_one_flat[pytree-compile]	0.3317ms	0.1934ms	5.1705 KOps/s	4.9227 KOps/s	$\textbf{\color{#35bf28}+5.03\%}$
test_compile_add_one_flat[pytree-eager]	0.9730ms	0.5401ms	1.8516 KOps/s	1.7791 KOps/s	$\color{#35bf28}+4.08\%$
test_compile_add_self_flat[tensordict-eager]	0.7725ms	0.3321ms	3.0116 KOps/s	2.9415 KOps/s	$\color{#35bf28}+2.38\%$
test_compile_add_self_flat[tensordict-compile]	0.3073ms	0.2549ms	3.9228 KOps/s	3.6661 KOps/s	$\textbf{\color{#35bf28}+7.00\%}$
test_compile_add_self_flat[tensorclass-eager]	0.5377ms	93.2384μs	10.7252 KOps/s	10.6415 KOps/s	$\color{#35bf28}+0.79\%$
test_compile_add_self_flat[tensorclass-compile]	0.3838ms	0.1908ms	5.2414 KOps/s	4.9514 KOps/s	$\textbf{\color{#35bf28}+5.86\%}$
test_compile_add_self_flat[pytree-eager]	0.8696ms	0.4415ms	2.2652 KOps/s	2.3171 KOps/s	$\color{#d91a1a}-2.24\%$
test_compile_add_self_flat[pytree-compile]	0.6331ms	0.1963ms	5.0946 KOps/s	4.9869 KOps/s	$\color{#35bf28}+2.16\%$
test_compile_copy_flat[tensordict-compile]	74.9820μs	20.0635μs	49.8418 KOps/s	49.1417 KOps/s	$\color{#35bf28}+1.42\%$
test_compile_copy_flat[tensordict-eager]	0.4803ms	41.6101μs	24.0326 KOps/s	23.6656 KOps/s	$\color{#35bf28}+1.55\%$
test_compile_copy_flat[pytree-compile]	0.4558ms	21.7729μs	45.9287 KOps/s	45.2379 KOps/s	$\color{#35bf28}+1.53\%$
test_compile_copy_flat[pytree-eager]	0.4834ms	52.6289μs	19.0010 KOps/s	18.8772 KOps/s	$\color{#35bf28}+0.66\%$
test_compile_assign_and_add[tensordict-compile]	2.3187ms	0.2423ms	4.1266 KOps/s	3.8650 KOps/s	$\textbf{\color{#35bf28}+6.77\%}$
test_compile_assign_and_add[tensordict-eager]	3.7852ms	3.3366ms	299.7050 Ops/s	298.9851 Ops/s	$\color{#35bf28}+0.24\%$
test_compile_assign_and_add[pytree-compile]	2.1232ms	0.1824ms	5.4820 KOps/s	5.4025 KOps/s	$\color{#35bf28}+1.47\%$
test_compile_assign_and_add[pytree-eager]	3.3366ms	2.8796ms	347.2742 Ops/s	354.8617 Ops/s	$\color{#d91a1a}-2.14\%$
test_compile_indexing[tensor-tensordict-compile]	0.5715ms	0.1365ms	7.3250 KOps/s	6.9071 KOps/s	$\textbf{\color{#35bf28}+6.05\%}$
test_compile_indexing[tensor-tensordict-eager]	0.2204s	95.1938μs	10.5049 KOps/s	12.4864 KOps/s	$\textbf{\color{#d91a1a}-15.87\%}$
test_compile_indexing[tensor-tensorclass-compile]	0.2631ms	0.1196ms	8.3596 KOps/s	8.3151 KOps/s	$\color{#35bf28}+0.53\%$
test_compile_indexing[tensor-tensorclass-eager]	0.4754ms	46.0467μs	21.7171 KOps/s	20.3102 KOps/s	$\textbf{\color{#35bf28}+6.93\%}$
test_compile_indexing[tensor-pytree-compile]	0.5506ms	0.1206ms	8.2937 KOps/s	8.2948 KOps/s	$\color{#d91a1a}-0.01\%$
test_compile_indexing[tensor-pytree-eager]	0.4837ms	45.9250μs	21.7746 KOps/s	20.9302 KOps/s	$\color{#35bf28}+4.03\%$
test_compile_indexing[slice-tensordict-compile]	0.1753ms	86.8894μs	11.5089 KOps/s	11.6732 KOps/s	$\color{#d91a1a}-1.41\%$
test_compile_indexing[slice-tensordict-eager]	0.2201ms	29.1839μs	34.2655 KOps/s	35.0977 KOps/s	$\color{#d91a1a}-2.37\%$
test_compile_indexing[slice-tensorclass-compile]	0.2206ms	70.9810μs	14.0883 KOps/s	14.1462 KOps/s	$\color{#d91a1a}-0.41\%$
test_compile_indexing[slice-tensorclass-eager]	0.2462ms	22.9574μs	43.5590 KOps/s	43.8460 KOps/s	$\color{#d91a1a}-0.65\%$
test_compile_indexing[slice-pytree-compile]	0.1014ms	72.0615μs	13.8770 KOps/s	13.8584 KOps/s	$\color{#35bf28}+0.13\%$
test_compile_indexing[slice-pytree-eager]	0.2745ms	22.6537μs	44.1430 KOps/s	44.2534 KOps/s	$\color{#d91a1a}-0.25\%$
test_compile_indexing[int-tensordict-compile]	0.1230ms	86.2723μs	11.5912 KOps/s	11.6834 KOps/s	$\color{#d91a1a}-0.79\%$
test_compile_indexing[int-tensordict-eager]	0.2911ms	28.5106μs	35.0747 KOps/s	34.0975 KOps/s	$\color{#35bf28}+2.87\%$
test_compile_indexing[int-tensorclass-compile]	0.1093ms	73.5216μs	13.6014 KOps/s	13.9759 KOps/s	$\color{#d91a1a}-2.68\%$
test_compile_indexing[int-tensorclass-eager]	0.2473ms	22.9382μs	43.5953 KOps/s	44.5585 KOps/s	$\color{#d91a1a}-2.16\%$
test_compile_indexing[int-pytree-compile]	0.1157ms	72.3203μs	13.8274 KOps/s	13.3783 KOps/s	$\color{#35bf28}+3.36\%$
test_compile_indexing[int-pytree-eager]	0.2778ms	22.7712μs	43.9151 KOps/s	44.3267 KOps/s	$\color{#d91a1a}-0.93\%$
test_compile_replace[single-eager]	80.4920μs	48.6336μs	20.5619 KOps/s	20.0121 KOps/s	$\color{#35bf28}+2.75\%$
test_compile_replace[single-compile]	0.2934ms	0.1307ms	7.6500 KOps/s	7.1201 KOps/s	$\textbf{\color{#35bf28}+7.44\%}$
test_compile_replace[multi-eager]	0.6801ms	0.5716ms	1.7494 KOps/s	1.6922 KOps/s	$\color{#35bf28}+3.38\%$
test_compile_replace[multi-compile]	0.3383ms	0.1387ms	7.2123 KOps/s	7.1541 KOps/s	$\color{#35bf28}+0.81\%$
test_compile_tc_getattr_20[eager]	0.2085ms	0.1738ms	5.7523 KOps/s	6.0776 KOps/s	$\textbf{\color{#d91a1a}-5.35\%}$
test_compile_tc_getattr_20[compile]	0.3187ms	0.1498ms	6.6751 KOps/s	6.7388 KOps/s	$\color{#d91a1a}-0.95\%$
test_compile_clone_shallow[20-eager]	48.6410μs	19.4737μs	51.3514 KOps/s	51.6781 KOps/s	$\color{#d91a1a}-0.63\%$
test_compile_clone_shallow[20-compile]	76.7720μs	17.2683μs	57.9095 KOps/s	57.8729 KOps/s	$\color{#35bf28}+0.06\%$
test_compile_clone_shallow[40-eager]	65.5820μs	34.0523μs	29.3666 KOps/s	29.5451 KOps/s	$\color{#d91a1a}-0.60\%$
test_compile_clone_shallow[40-compile]	91.3930μs	19.0505μs	52.4921 KOps/s	53.6587 KOps/s	$\color{#d91a1a}-2.17\%$
test_compile_clone_shallow[80-eager]	92.1820μs	63.2933μs	15.7995 KOps/s	15.8097 KOps/s	$\color{#d91a1a}-0.06\%$
test_compile_clone_shallow[80-compile]	86.7520μs	22.2082μs	45.0284 KOps/s	44.2408 KOps/s	$\color{#35bf28}+1.78\%$
test_compile_update_inplace[eager]	0.2358ms	61.3323μs	16.3046 KOps/s	16.5779 KOps/s	$\color{#d91a1a}-1.65\%$
test_compile_update_inplace[compile]	0.2163ms	0.1487ms	6.7234 KOps/s	6.6030 KOps/s	$\color{#35bf28}+1.82\%$
test_mod_add[eager]	0.2915ms	0.2287ms	4.3722 KOps/s	4.3736 KOps/s	$\color{#d91a1a}-0.03\%$
test_mod_add[compile]	0.2180ms	0.1329ms	7.5237 KOps/s	7.3005 KOps/s	$\color{#35bf28}+3.06\%$
test_mod_add[compile-overhead]	0.4169ms	0.1795ms	5.5720 KOps/s	5.5271 KOps/s	$\color{#35bf28}+0.81\%$
test_mod_wrap[eager]	0.5624ms	0.4704ms	2.1259 KOps/s	2.1214 KOps/s	$\color{#35bf28}+0.21\%$
test_mod_wrap[compile]	0.5248ms	0.4043ms	2.4732 KOps/s	2.4666 KOps/s	$\color{#35bf28}+0.27\%$
test_mod_wrap[compile-overhead]	8.4186ms	4.5453ms	220.0093 Ops/s	216.3506 Ops/s	$\color{#35bf28}+1.69\%$
test_mod_wrap_and_backward[eager]	2.0574ms	1.7054ms	586.3665 Ops/s	554.6397 Ops/s	$\textbf{\color{#35bf28}+5.72\%}$
test_mod_wrap_and_backward[compile]	1.7123ms	1.5546ms	643.2416 Ops/s	594.9936 Ops/s	$\textbf{\color{#35bf28}+8.11\%}$
test_mod_wrap_and_backward[compile-overhead]	1.4651ms	0.9497ms	1.0530 KOps/s	926.3807 Ops/s	$\textbf{\color{#35bf28}+13.67\%}$
test_seq_add[eager]	0.6055ms	0.5132ms	1.9484 KOps/s	1.9538 KOps/s	$\color{#d91a1a}-0.27\%$
test_seq_add[compile]	0.2318ms	0.1425ms	7.0170 KOps/s	6.8170 KOps/s	$\color{#35bf28}+2.93\%$
test_seq_add[compile-overhead]	0.3127ms	0.1836ms	5.4479 KOps/s	5.2662 KOps/s	$\color{#35bf28}+3.45\%$
test_seq_wrap[eager]	1.5230ms	1.4211ms	703.6634 Ops/s	699.4685 Ops/s	$\color{#35bf28}+0.60\%$
test_seq_wrap[compile]	0.4859ms	0.4141ms	2.4152 KOps/s	2.3650 KOps/s	$\color{#35bf28}+2.12\%$
test_seq_wrap[compile-overhead]	0.4414ms	0.3071ms	3.2561 KOps/s	3.2309 KOps/s	$\color{#35bf28}+0.78\%$
test_func_call_runtime[False-eager]	0.9313ms	0.8489ms	1.1780 KOps/s	1.2536 KOps/s	$\textbf{\color{#d91a1a}-6.03\%}$
test_func_call_runtime[False-compile]	1.2271ms	0.9897ms	1.0104 KOps/s	994.6268 Ops/s	$\color{#35bf28}+1.59\%$
test_func_call_runtime[False-compile-overhead]	0.5976ms	0.5179ms	1.9310 KOps/s	1.8981 KOps/s	$\color{#35bf28}+1.73\%$
test_func_call_runtime[True-eager]	1.1745ms	1.0711ms	933.6421 Ops/s	936.0890 Ops/s	$\color{#d91a1a}-0.26\%$
test_func_call_runtime[True-compile]	1.0961ms	1.0070ms	993.0512 Ops/s	983.8260 Ops/s	$\color{#35bf28}+0.94\%$
test_func_call_runtime[True-compile-overhead]	0.6404ms	0.5343ms	1.8718 KOps/s	1.8650 KOps/s	$\color{#35bf28}+0.36\%$
test_func_call_cm_runtime[False-eager]	1.1937ms	0.8445ms	1.1842 KOps/s	1.2406 KOps/s	$\color{#d91a1a}-4.55\%$
test_func_call_cm_runtime[False-compile]	1.5160ms	1.0242ms	976.3855 Ops/s	1.0106 KOps/s	$\color{#d91a1a}-3.38\%$
test_func_call_cm_runtime[False-compile-overhead]	0.7097ms	0.4964ms	2.0146 KOps/s	1.9618 KOps/s	$\color{#35bf28}+2.69\%$
test_func_call_cm_runtime[True-eager]	1.3372ms	1.2141ms	823.6558 Ops/s	816.2485 Ops/s	$\color{#35bf28}+0.91\%$
test_func_call_cm_runtime[True-compile]	1.1484ms	1.0282ms	972.6054 Ops/s	961.2466 Ops/s	$\color{#35bf28}+1.18\%$
test_func_call_cm_runtime[True-compile-overhead]	0.6260ms	0.5539ms	1.8052 KOps/s	1.7804 KOps/s	$\color{#35bf28}+1.39\%$
test_vmap_func_call_cm_runtime[eager]	2.9238ms	2.4325ms	411.0943 Ops/s	410.3262 Ops/s	$\color{#35bf28}+0.19\%$
test_vmap_func_call_cm_runtime[compile]	3.1322ms	2.6144ms	382.4964 Ops/s	381.0886 Ops/s	$\color{#35bf28}+0.37\%$
test_vmap_func_call_cm_runtime[compile-overhead]	3.0692ms	2.6037ms	384.0733 Ops/s	380.0681 Ops/s	$\color{#35bf28}+1.05\%$
test_distributed	0.8382ms	0.1525ms	6.5580 KOps/s	6.4491 KOps/s	$\color{#35bf28}+1.69\%$
test_tdmodule	0.3907ms	0.1921ms	5.2057 KOps/s	5.2043 KOps/s	$\color{#35bf28}+0.03\%$
test_tdmodule_dispatch	0.4268ms	0.3749ms	2.6676 KOps/s	2.6684 KOps/s	$\color{#d91a1a}-0.03\%$
test_tdseq	45.9910μs	26.8243μs	37.2796 KOps/s	37.3790 KOps/s	$\color{#d91a1a}-0.27\%$
test_tdseq_dispatch	0.2600ms	0.2169ms	4.6104 KOps/s	4.6144 KOps/s	$\color{#d91a1a}-0.09\%$
test_instantiation_functorch	2.1961ms	2.0852ms	479.5767 Ops/s	485.0473 Ops/s	$\color{#d91a1a}-1.13\%$
test_exec_functorch	0.2538ms	0.1843ms	5.4270 KOps/s	5.4558 KOps/s	$\color{#d91a1a}-0.53\%$
test_exec_functional_call	0.2188ms	0.1663ms	6.0136 KOps/s	6.1411 KOps/s	$\color{#d91a1a}-2.08\%$
test_exec_td_decorator	0.4992ms	0.2545ms	3.9297 KOps/s	3.9225 KOps/s	$\color{#35bf28}+0.18\%$
test_vmap_mlp_speed_decorator[True-True]	1.2407ms	1.0443ms	957.5339 Ops/s	949.6735 Ops/s	$\color{#35bf28}+0.83\%$
test_vmap_mlp_speed_decorator[True-False]	1.2322ms	1.0374ms	963.9254 Ops/s	950.3738 Ops/s	$\color{#35bf28}+1.43\%$
test_vmap_mlp_speed_decorator[False-True]	0.9610ms	0.7401ms	1.3512 KOps/s	1.3101 KOps/s	$\color{#35bf28}+3.14\%$
test_vmap_mlp_speed_decorator[False-False]	0.9585ms	0.7383ms	1.3545 KOps/s	1.3115 KOps/s	$\color{#35bf28}+3.28\%$
test_vmap_transformer_speed_decorator[True-True]	21.6100ms	21.2575ms	47.0423 Ops/s	46.6624 Ops/s	$\color{#35bf28}+0.81\%$
test_vmap_transformer_speed_decorator[True-False]	21.9422ms	21.2510ms	47.0566 Ops/s	46.9642 Ops/s	$\color{#35bf28}+0.20\%$
test_vmap_transformer_speed_decorator[False-True]	21.6382ms	20.8827ms	47.8865 Ops/s	47.6570 Ops/s	$\color{#35bf28}+0.48\%$
test_vmap_transformer_speed_decorator[False-False]	21.3722ms	21.0352ms	47.5393 Ops/s	46.5556 Ops/s	$\color{#35bf28}+2.11\%$
test_to_module_speed[True]	2.3217ms	1.6680ms	599.5099 Ops/s	603.7415 Ops/s	$\color{#d91a1a}-0.70\%$
test_to_module_speed[False]	2.1919ms	1.6439ms	608.2968 Ops/s	610.7162 Ops/s	$\color{#d91a1a}-0.40\%$
test_to_module_plain_tensor_speed[None]	2.4315ms	1.7491ms	571.7296 Ops/s	565.1409 Ops/s	$\color{#35bf28}+1.17\%$
test_to_module_plain_tensor_speed[False]	2.2324ms	1.5893ms	629.2218 Ops/s	625.0654 Ops/s	$\color{#35bf28}+0.66\%$
test_to_module_plain_tensor_speed[True]	3.5392ms	3.4125ms	293.0427 Ops/s	286.3566 Ops/s	$\color{#35bf28}+2.33\%$
test_tc_init	80.2620μs	43.5703μs	22.9514 KOps/s	22.2393 KOps/s	$\color{#35bf28}+3.20\%$
test_tc_init_tensor_only	48.9410μs	9.2966μs	107.5664 KOps/s	105.4155 KOps/s	$\color{#35bf28}+2.04\%$
test_tc_init_nested	0.1279ms	85.7315μs	11.6643 KOps/s	11.6732 KOps/s	$\color{#d91a1a}-0.08\%$
test_tc_init_many_fields	48.1010μs	14.7610μs	67.7461 KOps/s	66.9648 KOps/s	$\color{#35bf28}+1.17\%$
test_tc_first_layer_tensor	50.4720μs	1.7710μs	564.6560 KOps/s	545.5887 KOps/s	$\color{#35bf28}+3.49\%$
test_tc_first_layer_tensor_only	2.9771μs	0.3883μs	2.5751 MOps/s	2.5398 MOps/s	$\color{#35bf28}+1.39\%$
test_tc_first_layer_tensor_set	25.2910μs	3.8908μs	257.0171 KOps/s	252.3604 KOps/s	$\color{#35bf28}+1.85\%$
test_tc_first_layer_tensor_only_set	35.0910μs	3.2605μs	306.7059 KOps/s	303.1341 KOps/s	$\color{#35bf28}+1.18\%$
test_tc_first_layer_nontensor	48.7810μs	6.1683μs	162.1198 KOps/s	154.9751 KOps/s	$\color{#35bf28}+4.61\%$
test_tc_second_layer_tensor	22.2410μs	4.4728μs	223.5730 KOps/s	226.5197 KOps/s	$\color{#d91a1a}-1.30\%$
test_tc_second_layer_nontensor	41.0810μs	8.7759μs	113.9483 KOps/s	111.6715 KOps/s	$\color{#35bf28}+2.04\%$
test_unbind	0.2741s	13.7215ms	72.8785 Ops/s	64.8198 Ops/s	$\textbf{\color{#35bf28}+12.43\%}$
test_full_like	4.4878ms	4.3795ms	228.3381 Ops/s	226.2470 Ops/s	$\color{#35bf28}+0.92\%$
test_zeros_like	4.4919ms	4.3726ms	228.6944 Ops/s	227.2528 Ops/s	$\color{#35bf28}+0.63\%$
test_ones_like	4.7051ms	4.3871ms	227.9433 Ops/s	227.8069 Ops/s	$\color{#35bf28}+0.06\%$
test_clone	7.1568ms	6.6119ms	151.2427 Ops/s	153.1915 Ops/s	$\color{#d91a1a}-1.27\%$
test_squeeze	69.4010μs	14.4008μs	69.4405 KOps/s	67.9444 KOps/s	$\color{#35bf28}+2.20\%$
test_unsqueeze	0.1722ms	0.1164ms	8.5900 KOps/s	8.9857 KOps/s	$\color{#d91a1a}-4.40\%$
test_split	0.3908ms	0.1862ms	5.3708 KOps/s	5.4233 KOps/s	$\color{#d91a1a}-0.97\%$
test_permute	0.2820ms	0.2160ms	4.6302 KOps/s	4.6416 KOps/s	$\color{#d91a1a}-0.24\%$
test_stack	18.4555ms	18.2234ms	54.8744 Ops/s	54.4331 Ops/s	$\color{#35bf28}+0.81\%$
test_cat	27.3353ms	26.2353ms	38.1166 Ops/s	54.5241 Ops/s	$\textbf{\color{#d91a1a}-30.09\%}$
test_sequential_tensordict	0.7513ms	0.5726ms	1.7466 KOps/s	1.7084 KOps/s	$\color{#35bf28}+2.23\%$
test_sequential_graph_module	0.1960ms	0.1235ms	8.0976 KOps/s	7.8797 KOps/s	$\color{#35bf28}+2.77\%$
test_nested_tensordict	1.0948ms	0.8194ms	1.2203 KOps/s	1.2116 KOps/s	$\color{#35bf28}+0.72\%$
test_nested_graph_module	0.5346ms	0.1337ms	7.4768 KOps/s	7.4954 KOps/s	$\color{#d91a1a}-0.25\%$

[github-actions]

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

Total Benchmarks: 344. Improved: $\large\color{#35bf28}18$. Worsened: $\large\color{#d91a1a}10$.

Expand to view detailed results

Name	Max	Mean	Ops	Ops on Repo HEAD	Change
test_collector_preallocated_write[stack_out-cpu-32-2]	0.2101ms	0.1649ms	6.0656 KOps/s	5.9037 KOps/s	$\color{#35bf28}+2.74\%$
test_collector_preallocated_write[stack_out-cpu-32-8]	0.3453ms	0.2943ms	3.3977 KOps/s	3.3720 KOps/s	$\color{#35bf28}+0.76\%$
test_collector_preallocated_write[stack_out-cpu-32-32]	0.8177ms	0.7745ms	1.2912 KOps/s	1.2786 KOps/s	$\color{#35bf28}+0.98\%$
test_collector_preallocated_write[stack_out-cpu-4096-2]	0.4575ms	0.3449ms	2.8997 KOps/s	2.9799 KOps/s	$\color{#d91a1a}-2.69\%$
test_collector_preallocated_write[stack_out-cpu-4096-8]	1.4357ms	1.1091ms	901.6379 Ops/s	921.4535 Ops/s	$\color{#d91a1a}-2.15\%$
test_collector_preallocated_write[stack_out-cpu-4096-32]	7.8688ms	7.4211ms	134.7516 Ops/s	133.4950 Ops/s	$\color{#35bf28}+0.94\%$
test_collector_preallocated_write[stack_out-cuda-32-2]	0.2903ms	0.2460ms	4.0651 KOps/s	4.0571 KOps/s	$\color{#35bf28}+0.20\%$
test_collector_preallocated_write[stack_out-cuda-32-8]	0.4724ms	0.3719ms	2.6888 KOps/s	2.7058 KOps/s	$\color{#d91a1a}-0.63\%$
test_collector_preallocated_write[stack_out-cuda-32-32]	0.9026ms	0.8217ms	1.2170 KOps/s	1.2164 KOps/s	$\color{#35bf28}+0.05\%$
test_collector_preallocated_write[stack_out-cuda-4096-2]	0.3272ms	0.2526ms	3.9586 KOps/s	4.0508 KOps/s	$\color{#d91a1a}-2.28\%$
test_collector_preallocated_write[stack_out-cuda-4096-8]	0.4513ms	0.3730ms	2.6812 KOps/s	2.6934 KOps/s	$\color{#d91a1a}-0.45\%$
test_collector_preallocated_write[stack_out-cuda-4096-32]	1.0284ms	0.8837ms	1.1316 KOps/s	1.1255 KOps/s	$\color{#35bf28}+0.54\%$
test_collector_preallocated_write[set_at_loop-cpu-32-2]	0.2843ms	0.2136ms	4.6827 KOps/s	4.6488 KOps/s	$\color{#35bf28}+0.73\%$
test_collector_preallocated_write[set_at_loop-cpu-32-8]	0.9151ms	0.8213ms	1.2176 KOps/s	1.1994 KOps/s	$\color{#35bf28}+1.52\%$
test_collector_preallocated_write[set_at_loop-cpu-32-32]	3.5251ms	3.3254ms	300.7178 Ops/s	301.6993 Ops/s	$\color{#d91a1a}-0.33\%$
test_collector_preallocated_write[set_at_loop-cpu-4096-2]	0.4440ms	0.3680ms	2.7175 KOps/s	2.6727 KOps/s	$\color{#35bf28}+1.67\%$
test_collector_preallocated_write[set_at_loop-cpu-4096-8]	2.2829ms	1.8582ms	538.1456 Ops/s	601.6935 Ops/s	$\textbf{\color{#d91a1a}-10.56\%}$
test_collector_preallocated_write[set_at_loop-cpu-4096-32]	10.5793ms	10.3272ms	96.8319 Ops/s	95.4043 Ops/s	$\color{#35bf28}+1.50\%$
test_collector_preallocated_write[set_at_loop-cuda-32-2]	0.4776ms	0.3863ms	2.5890 KOps/s	2.5928 KOps/s	$\color{#d91a1a}-0.15\%$
test_collector_preallocated_write[set_at_loop-cuda-32-8]	1.7610ms	1.4810ms	675.2342 Ops/s	680.9649 Ops/s	$\color{#d91a1a}-0.84\%$
test_collector_preallocated_write[set_at_loop-cuda-32-32]	5.9778ms	5.7058ms	175.2588 Ops/s	168.3350 Ops/s	$\color{#35bf28}+4.11\%$
test_collector_preallocated_write[set_at_loop-cuda-4096-2]	0.4898ms	0.3992ms	2.5052 KOps/s	2.5078 KOps/s	$\color{#d91a1a}-0.10\%$
test_collector_preallocated_write[set_at_loop-cuda-4096-8]	1.6136ms	1.4933ms	669.6766 Ops/s	673.1900 Ops/s	$\color{#d91a1a}-0.52\%$
test_collector_preallocated_write[set_at_loop-cuda-4096-32]	6.0458ms	5.7937ms	172.6016 Ops/s	176.7606 Ops/s	$\color{#d91a1a}-2.35\%$
test_collector_preallocated_write[copy_at-cpu-32-2]	0.2234ms	0.1726ms	5.7931 KOps/s	5.7172 KOps/s	$\color{#35bf28}+1.33\%$
test_collector_preallocated_write[copy_at-cpu-32-8]	0.7499ms	0.6770ms	1.4771 KOps/s	1.4751 KOps/s	$\color{#35bf28}+0.13\%$
test_collector_preallocated_write[copy_at-cpu-32-32]	2.7610ms	2.6873ms	372.1202 Ops/s	372.7704 Ops/s	$\color{#d91a1a}-0.17\%$
test_collector_preallocated_write[copy_at-cpu-4096-2]	1.9551ms	0.3204ms	3.1212 KOps/s	3.0763 KOps/s	$\color{#35bf28}+1.46\%$
test_collector_preallocated_write[copy_at-cpu-4096-8]	1.9266ms	1.5215ms	657.2449 Ops/s	668.6886 Ops/s	$\color{#d91a1a}-1.71\%$
test_collector_preallocated_write[copy_at-cpu-4096-32]	10.0061ms	9.7474ms	102.5911 Ops/s	102.4820 Ops/s	$\color{#35bf28}+0.11\%$
test_collector_preallocated_write[copy_at-cuda-32-2]	0.3628ms	0.3231ms	3.0947 KOps/s	3.0923 KOps/s	$\color{#35bf28}+0.08\%$
test_collector_preallocated_write[copy_at-cuda-32-8]	1.5267ms	1.2072ms	828.3514 Ops/s	827.3358 Ops/s	$\color{#35bf28}+0.12\%$
test_collector_preallocated_write[copy_at-cuda-32-32]	4.8358ms	4.7186ms	211.9275 Ops/s	211.2908 Ops/s	$\color{#35bf28}+0.30\%$
test_collector_preallocated_write[copy_at-cuda-4096-2]	0.3898ms	0.3245ms	3.0814 KOps/s	3.0692 KOps/s	$\color{#35bf28}+0.40\%$
test_collector_preallocated_write[copy_at-cuda-4096-8]	1.2871ms	1.2124ms	824.7801 Ops/s	823.6429 Ops/s	$\color{#35bf28}+0.14\%$
test_collector_preallocated_write[copy_at-cuda-4096-32]	4.8318ms	4.7500ms	210.5252 Ops/s	208.8753 Ops/s	$\color{#35bf28}+0.79\%$
test_collector_preallocated_write[manual_leaf_copy-cpu-32-2]	0.1585ms	0.1190ms	8.4030 KOps/s	8.3527 KOps/s	$\color{#35bf28}+0.60\%$
test_collector_preallocated_write[manual_leaf_copy-cpu-32-8]	0.5171ms	0.4686ms	2.1338 KOps/s	2.1424 KOps/s	$\color{#d91a1a}-0.40\%$
test_collector_preallocated_write[manual_leaf_copy-cpu-32-32]	1.9511ms	1.8570ms	538.5158 Ops/s	540.0538 Ops/s	$\color{#d91a1a}-0.28\%$
test_collector_preallocated_write[manual_leaf_copy-cpu-4096-2]	0.2988ms	0.2518ms	3.9711 KOps/s	3.7867 KOps/s	$\color{#35bf28}+4.87\%$
test_collector_preallocated_write[manual_leaf_copy-cpu-4096-8]	1.4679ms	1.2474ms	801.6928 Ops/s	819.1025 Ops/s	$\color{#d91a1a}-2.13\%$
test_collector_preallocated_write[manual_leaf_copy-cpu-4096-32]	9.2080ms	8.7844ms	113.8388 Ops/s	115.0272 Ops/s	$\color{#d91a1a}-1.03\%$
test_collector_preallocated_write[manual_leaf_copy-cuda-32-2]	0.3292ms	0.2709ms	3.6920 KOps/s	3.5729 KOps/s	$\color{#35bf28}+3.33\%$
test_collector_preallocated_write[manual_leaf_copy-cuda-32-8]	1.1240ms	1.0000ms	1.0000 KOps/s	972.5482 Ops/s	$\color{#35bf28}+2.83\%$
test_collector_preallocated_write[manual_leaf_copy-cuda-32-32]	4.0927ms	3.9034ms	256.1889 Ops/s	253.9847 Ops/s	$\color{#35bf28}+0.87\%$
test_collector_preallocated_write[manual_leaf_copy-cuda-4096-2]	0.3278ms	0.2687ms	3.7210 KOps/s	3.6329 KOps/s	$\color{#35bf28}+2.43\%$
test_collector_preallocated_write[manual_leaf_copy-cuda-4096-8]	1.3051ms	0.9995ms	1.0005 KOps/s	979.7078 Ops/s	$\color{#35bf28}+2.13\%$
test_collector_preallocated_write[manual_leaf_copy-cuda-4096-32]	4.1155ms	3.9113ms	255.6673 Ops/s	252.0038 Ops/s	$\color{#35bf28}+1.45\%$
test_wide_nested_preallocated_write[stack_out-cpu-128-128]	16.4405ms	16.2542ms	61.5225 Ops/s	61.4278 Ops/s	$\color{#35bf28}+0.15\%$
test_wide_nested_preallocated_write[stack_out-cpu-128-256]	33.0832ms	32.9079ms	30.3878 Ops/s	30.5056 Ops/s	$\color{#d91a1a}-0.39\%$
test_wide_nested_preallocated_write[stack_out-cpu-4096-128]	89.6247ms	87.4636ms	11.4333 Ops/s	11.6075 Ops/s	$\color{#d91a1a}-1.50\%$
test_wide_nested_preallocated_write[stack_out-cpu-4096-256]	0.2034s	0.2019s	4.9540 Ops/s	4.9922 Ops/s	$\color{#d91a1a}-0.76\%$
test_wide_nested_preallocated_write[stack_out-cuda-128-128]	13.9445ms	13.8755ms	72.0695 Ops/s	71.9199 Ops/s	$\color{#35bf28}+0.21\%$
test_wide_nested_preallocated_write[stack_out-cuda-128-256]	27.2147ms	27.0504ms	36.9681 Ops/s	36.9669 Ops/s	$+0.00\%$
test_wide_nested_preallocated_write[stack_out-cuda-4096-128]	14.0694ms	13.9648ms	71.6088 Ops/s	71.3354 Ops/s	$\color{#35bf28}+0.38\%$
test_wide_nested_preallocated_write[stack_out-cuda-4096-256]	27.6150ms	27.5059ms	36.3558 Ops/s	36.5123 Ops/s	$\color{#d91a1a}-0.43\%$
test_wide_nested_preallocated_write[set_at_loop-cpu-128-128]	52.4975ms	52.2078ms	19.1542 Ops/s	19.4772 Ops/s	$\color{#d91a1a}-1.66\%$
test_wide_nested_preallocated_write[set_at_loop-cpu-128-256]	0.1054s	0.1052s	9.5075 Ops/s	9.5748 Ops/s	$\color{#d91a1a}-0.70\%$
test_wide_nested_preallocated_write[set_at_loop-cpu-4096-128]	0.1492s	0.1474s	6.7837 Ops/s	6.8655 Ops/s	$\color{#d91a1a}-1.19\%$
test_wide_nested_preallocated_write[set_at_loop-cpu-4096-256]	0.3241s	0.3218s	3.1075 Ops/s	3.0292 Ops/s	$\color{#35bf28}+2.59\%$
test_wide_nested_preallocated_write[set_at_loop-cuda-128-128]	74.7750ms	74.1698ms	13.4826 Ops/s	13.7077 Ops/s	$\color{#d91a1a}-1.64\%$
test_wide_nested_preallocated_write[set_at_loop-cuda-128-256]	0.1494s	0.1491s	6.7060 Ops/s	6.5184 Ops/s	$\color{#35bf28}+2.88\%$
test_wide_nested_preallocated_write[set_at_loop-cuda-4096-128]	75.1708ms	74.5853ms	13.4075 Ops/s	12.7836 Ops/s	$\color{#35bf28}+4.88\%$
test_wide_nested_preallocated_write[set_at_loop-cuda-4096-256]	0.1502s	0.1499s	6.6726 Ops/s	6.3189 Ops/s	$\textbf{\color{#35bf28}+5.60\%}$
test_wide_nested_preallocated_write[copy_at-cpu-128-128]	30.6475ms	30.4842ms	32.8038 Ops/s	32.8433 Ops/s	$\color{#d91a1a}-0.12\%$
test_wide_nested_preallocated_write[copy_at-cpu-128-256]	61.8491ms	61.7000ms	16.2075 Ops/s	16.2305 Ops/s	$\color{#d91a1a}-0.14\%$
test_wide_nested_preallocated_write[copy_at-cpu-4096-128]	0.1236s	0.1230s	8.1316 Ops/s	8.2140 Ops/s	$\color{#d91a1a}-1.00\%$
test_wide_nested_preallocated_write[copy_at-cpu-4096-256]	0.2737s	0.2724s	3.6717 Ops/s	3.5722 Ops/s	$\color{#35bf28}+2.79\%$
test_wide_nested_preallocated_write[copy_at-cuda-128-128]	43.3170ms	43.0576ms	23.2247 Ops/s	23.2093 Ops/s	$\color{#35bf28}+0.07\%$
test_wide_nested_preallocated_write[copy_at-cuda-128-256]	86.4868ms	86.2763ms	11.5907 Ops/s	11.6032 Ops/s	$\color{#d91a1a}-0.11\%$
test_wide_nested_preallocated_write[copy_at-cuda-4096-128]	43.5825ms	43.4998ms	22.9886 Ops/s	23.0671 Ops/s	$\color{#d91a1a}-0.34\%$
test_wide_nested_preallocated_write[copy_at-cuda-4096-256]	86.9914ms	86.8141ms	11.5189 Ops/s	11.5128 Ops/s	$\color{#35bf28}+0.05\%$
test_wide_nested_preallocated_write[manual_leaf_copy-cpu-128-128]	20.4604ms	20.2454ms	49.3939 Ops/s	49.3883 Ops/s	$\color{#35bf28}+0.01\%$
test_wide_nested_preallocated_write[manual_leaf_copy-cpu-128-256]	41.5180ms	41.1739ms	24.2872 Ops/s	24.2407 Ops/s	$\color{#35bf28}+0.19\%$
test_wide_nested_preallocated_write[manual_leaf_copy-cpu-4096-128]	0.1150s	0.1094s	9.1417 Ops/s	9.2181 Ops/s	$\color{#d91a1a}-0.83\%$
test_wide_nested_preallocated_write[manual_leaf_copy-cpu-4096-256]	0.2491s	0.2464s	4.0581 Ops/s	3.9998 Ops/s	$\color{#35bf28}+1.46\%$
test_wide_nested_preallocated_write[manual_leaf_copy-cuda-128-128]	35.8203ms	35.4300ms	28.2247 Ops/s	27.9255 Ops/s	$\color{#35bf28}+1.07\%$
test_wide_nested_preallocated_write[manual_leaf_copy-cuda-128-256]	71.0765ms	70.6757ms	14.1491 Ops/s	14.0102 Ops/s	$\color{#35bf28}+0.99\%$
test_wide_nested_preallocated_write[manual_leaf_copy-cuda-4096-128]	35.7573ms	35.3157ms	28.3160 Ops/s	28.0373 Ops/s	$\color{#35bf28}+0.99\%$
test_wide_nested_preallocated_write[manual_leaf_copy-cuda-4096-256]	71.3791ms	70.7906ms	14.1262 Ops/s	13.8778 Ops/s	$\color{#35bf28}+1.79\%$
test_plain_set_nested	46.0910μs	15.5179μs	64.4417 KOps/s	66.0746 KOps/s	$\color{#d91a1a}-2.47\%$
test_plain_set_stack_nested	51.4420μs	15.2025μs	65.7785 KOps/s	65.4502 KOps/s	$\color{#35bf28}+0.50\%$
test_plain_set_nested_inplace	78.5920μs	16.9828μs	58.8830 KOps/s	58.6265 KOps/s	$\color{#35bf28}+0.44\%$
test_plain_set_stack_nested_inplace	55.2020μs	16.4758μs	60.6952 KOps/s	59.9546 KOps/s	$\color{#35bf28}+1.24\%$
test_items	22.9510μs	6.1081μs	163.7158 KOps/s	166.1714 KOps/s	$\color{#d91a1a}-1.48\%$
test_items_nested	0.5310ms	0.4589ms	2.1793 KOps/s	2.1294 KOps/s	$\color{#35bf28}+2.34\%$
test_items_nested_locked	0.5347ms	0.4588ms	2.1797 KOps/s	2.1373 KOps/s	$\color{#35bf28}+1.99\%$
test_items_nested_leaf	0.1331ms	97.3493μs	10.2723 KOps/s	10.3371 KOps/s	$\color{#d91a1a}-0.63\%$
test_items_stack_nested	0.5323ms	0.4601ms	2.1736 KOps/s	2.1256 KOps/s	$\color{#35bf28}+2.26\%$
test_items_stack_nested_leaf	0.1344ms	98.8282μs	10.1186 KOps/s	10.3108 KOps/s	$\color{#d91a1a}-1.86\%$
test_items_stack_nested_locked	0.5545ms	0.4598ms	2.1746 KOps/s	2.1152 KOps/s	$\color{#35bf28}+2.81\%$
test_keys	31.1010μs	4.2069μs	237.7059 KOps/s	237.3858 KOps/s	$\color{#35bf28}+0.13\%$
test_keys_nested	0.1792ms	0.1304ms	7.6670 KOps/s	7.6829 KOps/s	$\color{#d91a1a}-0.21\%$
test_keys_nested_locked	0.7909ms	0.1385ms	7.2193 KOps/s	7.2034 KOps/s	$\color{#35bf28}+0.22\%$
test_keys_nested_leaf	0.1671ms	0.1214ms	8.2350 KOps/s	8.2962 KOps/s	$\color{#d91a1a}-0.74\%$
test_keys_stack_nested	0.2012ms	0.1308ms	7.6452 KOps/s	7.6394 KOps/s	$\color{#35bf28}+0.08\%$
test_keys_stack_nested_leaf	0.1747ms	0.1201ms	8.3254 KOps/s	8.2832 KOps/s	$\color{#35bf28}+0.51\%$
test_keys_stack_nested_locked	0.1936ms	0.1397ms	7.1594 KOps/s	7.2419 KOps/s	$\color{#d91a1a}-1.14\%$
test_values	7.8162μs	1.0191μs	981.2268 KOps/s	983.6768 KOps/s	$\color{#d91a1a}-0.25\%$
test_values_nested	92.9730μs	52.7333μs	18.9633 KOps/s	18.9119 KOps/s	$\color{#35bf28}+0.27\%$
test_values_nested_locked	96.7830μs	55.8559μs	17.9032 KOps/s	17.7636 KOps/s	$\color{#35bf28}+0.79\%$
test_values_nested_leaf	92.1330μs	60.2012μs	16.6110 KOps/s	16.5775 KOps/s	$\color{#35bf28}+0.20\%$
test_values_stack_nested	88.3030μs	52.0797μs	19.2014 KOps/s	19.0947 KOps/s	$\color{#35bf28}+0.56\%$
test_values_stack_nested_leaf	0.1029ms	60.0096μs	16.6640 KOps/s	16.7567 KOps/s	$\color{#d91a1a}-0.55\%$
test_values_stack_nested_locked	0.1172ms	55.6993μs	17.9536 KOps/s	17.9238 KOps/s	$\color{#35bf28}+0.17\%$
test_membership	10.3402μs	0.8520μs	1.1737 MOps/s	1.1854 MOps/s	$\color{#d91a1a}-0.99\%$
test_membership_nested	38.4510μs	3.0201μs	331.1180 KOps/s	332.3156 KOps/s	$\color{#d91a1a}-0.36\%$
test_membership_nested_leaf	37.7210μs	2.9943μs	333.9629 KOps/s	336.5140 KOps/s	$\color{#d91a1a}-0.76\%$
test_membership_stacked_nested	40.7810μs	2.8944μs	345.4907 KOps/s	346.1014 KOps/s	$\color{#d91a1a}-0.18\%$
test_membership_stacked_nested_leaf	36.0110μs	2.8786μs	347.3923 KOps/s	350.9826 KOps/s	$\color{#d91a1a}-1.02\%$
test_membership_nested_last	40.2210μs	4.5036μs	222.0466 KOps/s	222.4423 KOps/s	$\color{#d91a1a}-0.18\%$
test_membership_nested_leaf_last	37.4710μs	4.4586μs	224.2869 KOps/s	223.9727 KOps/s	$\color{#35bf28}+0.14\%$
test_membership_stacked_nested_last	50.3510μs	4.3433μs	230.2403 KOps/s	230.1690 KOps/s	$\color{#35bf28}+0.03\%$
test_membership_stacked_nested_leaf_last	37.5910μs	4.2704μs	234.1690 KOps/s	230.4600 KOps/s	$\color{#35bf28}+1.61\%$
test_nested_getleaf	61.8720μs	21.5163μs	46.4765 KOps/s	46.4394 KOps/s	$\color{#35bf28}+0.08\%$
test_nested_get	48.0410μs	20.3334μs	49.1800 KOps/s	48.6067 KOps/s	$\color{#35bf28}+1.18\%$
test_stacked_getleaf	59.0620μs	21.2233μs	47.1181 KOps/s	47.0232 KOps/s	$\color{#35bf28}+0.20\%$
test_stacked_get	49.3510μs	20.3764μs	49.0763 KOps/s	49.3911 KOps/s	$\color{#d91a1a}-0.64\%$
test_nested_getitemleaf	47.5720μs	22.1130μs	45.2223 KOps/s	45.1606 KOps/s	$\color{#35bf28}+0.14\%$
test_nested_getitem	47.0810μs	20.7425μs	48.2103 KOps/s	47.6603 KOps/s	$\color{#35bf28}+1.15\%$
test_stacked_getitemleaf	53.2110μs	21.7338μs	46.0112 KOps/s	45.5319 KOps/s	$\color{#35bf28}+1.05\%$
test_stacked_getitem	57.2720μs	20.5654μs	48.6253 KOps/s	48.2082 KOps/s	$\color{#35bf28}+0.87\%$
test_lock_nested	7.8684ms	0.5094ms	1.9632 KOps/s	1.9938 KOps/s	$\color{#d91a1a}-1.53\%$
test_lock_stack_nested	0.6099ms	0.5114ms	1.9553 KOps/s	1.9851 KOps/s	$\color{#d91a1a}-1.50\%$
test_unlock_nested	0.4686ms	0.3984ms	2.5103 KOps/s	2.5440 KOps/s	$\color{#d91a1a}-1.32\%$
test_unlock_stack_nested	0.5345ms	0.4041ms	2.4747 KOps/s	2.5174 KOps/s	$\color{#d91a1a}-1.70\%$
test_flatten_speed	0.1741ms	0.1217ms	8.2183 KOps/s	8.2722 KOps/s	$\color{#d91a1a}-0.65\%$
test_unflatten_speed	0.6725ms	0.5702ms	1.7537 KOps/s	1.7341 KOps/s	$\color{#35bf28}+1.13\%$
test_common_ops	0.1001s	1.5898ms	629.0063 Ops/s	320.7477 Ops/s	$\textbf{\color{#35bf28}+96.11\%}$
test_creation	0.1150ms	3.2002μs	312.4841 KOps/s	320.6534 KOps/s	$\color{#d91a1a}-2.55\%$
test_creation_empty	35.4310μs	7.1389μs	140.0785 KOps/s	143.2826 KOps/s	$\color{#d91a1a}-2.24\%$
test_creation_nested_1	86.2520μs	11.6497μs	85.8393 KOps/s	86.4103 KOps/s	$\color{#d91a1a}-0.66\%$
test_creation_nested_2	41.0010μs	13.8329μs	72.2915 KOps/s	72.8710 KOps/s	$\color{#d91a1a}-0.80\%$
test_creation_many_keys[10]	48.1310μs	21.4918μs	46.5294 KOps/s	47.0297 KOps/s	$\color{#d91a1a}-1.06\%$
test_creation_many_keys[50]	0.1378ms	92.7120μs	10.7861 KOps/s	10.9823 KOps/s	$\color{#d91a1a}-1.79\%$
test_creation_many_keys[100]	0.2355ms	0.1798ms	5.5604 KOps/s	5.5522 KOps/s	$\color{#35bf28}+0.15\%$
test_creation_nested_many_keys[10]	0.1137ms	45.8301μs	21.8197 KOps/s	21.9741 KOps/s	$\color{#d91a1a}-0.70\%$
test_creation_nested_many_keys[50]	0.2543ms	0.1860ms	5.3757 KOps/s	5.3855 KOps/s	$\color{#d91a1a}-0.18\%$
test_clone	46.8310μs	13.3631μs	74.8328 KOps/s	75.4675 KOps/s	$\color{#d91a1a}-0.84\%$
test_getitem[int]	1.6494ms	15.3477μs	65.1565 KOps/s	67.2522 KOps/s	$\color{#d91a1a}-3.12\%$
test_getitem[slice_int]	0.1432ms	24.4742μs	40.8594 KOps/s	41.2835 KOps/s	$\color{#d91a1a}-1.03\%$
test_getitem[range]	0.1745ms	65.0240μs	15.3789 KOps/s	15.5628 KOps/s	$\color{#d91a1a}-1.18\%$
test_getitem[tuple]	0.1486ms	24.0906μs	41.5100 KOps/s	41.6222 KOps/s	$\color{#d91a1a}-0.27\%$
test_getitem[list]	0.1864ms	59.8803μs	16.7000 KOps/s	16.5993 KOps/s	$\color{#35bf28}+0.61\%$
test_setitem_dim[int]	49.9110μs	25.2358μs	39.6262 KOps/s	39.0959 KOps/s	$\color{#35bf28}+1.36\%$
test_setitem_dim[slice_int]	62.4220μs	42.0065μs	23.8058 KOps/s	23.5079 KOps/s	$\color{#35bf28}+1.27\%$
test_setitem_dim[range]	0.1301ms	94.6372μs	10.5667 KOps/s	10.2587 KOps/s	$\color{#35bf28}+3.00\%$
test_setitem_dim[tuple]	58.3520μs	38.8731μs	25.7247 KOps/s	25.9489 KOps/s	$\color{#d91a1a}-0.86\%$
test_setitem	46.0510μs	17.6847μs	56.5461 KOps/s	57.0140 KOps/s	$\color{#d91a1a}-0.82\%$
test_set	40.7210μs	16.9665μs	58.9395 KOps/s	59.2345 KOps/s	$\color{#d91a1a}-0.50\%$
test_set_shared	0.5146ms	0.2086ms	4.7932 KOps/s	4.6806 KOps/s	$\color{#35bf28}+2.41\%$
test_update	0.2121ms	21.9926μs	45.4698 KOps/s	45.5020 KOps/s	$\color{#d91a1a}-0.07\%$
test_update_nested	72.7720μs	33.4913μs	29.8585 KOps/s	30.6019 KOps/s	$\color{#d91a1a}-2.43\%$
test_update__nested	0.4866ms	33.1240μs	30.1896 KOps/s	30.4507 KOps/s	$\color{#d91a1a}-0.86\%$
test_set_nested	68.9620μs	18.7535μs	53.3234 KOps/s	53.0430 KOps/s	$\color{#35bf28}+0.53\%$
test_set_nested_new	61.8120μs	23.6824μs	42.2254 KOps/s	42.5042 KOps/s	$\color{#d91a1a}-0.66\%$
test_select	83.3020μs	39.7609μs	25.1504 KOps/s	24.9410 KOps/s	$\color{#35bf28}+0.84\%$
test_select_nested	0.1159ms	75.2704μs	13.2854 KOps/s	13.4716 KOps/s	$\color{#d91a1a}-1.38\%$
test_exclude_nested	0.1424ms	91.6146μs	10.9153 KOps/s	10.9806 KOps/s	$\color{#d91a1a}-0.59\%$
test_empty[True]	0.5101ms	0.3973ms	2.5167 KOps/s	2.5081 KOps/s	$\color{#35bf28}+0.34\%$
test_empty[False]	18.5955μs	1.4002μs	714.1663 KOps/s	729.5376 KOps/s	$\color{#d91a1a}-2.11\%$
test_to	0.1200ms	69.0003μs	14.4927 KOps/s	14.2143 KOps/s	$\color{#35bf28}+1.96\%$
test_to_nonblocking	0.1167ms	68.5612μs	14.5855 KOps/s	14.5542 KOps/s	$\color{#35bf28}+0.22\%$
test_unbind_speed	0.4014ms	0.3411ms	2.9315 KOps/s	2.9742 KOps/s	$\color{#d91a1a}-1.44\%$
test_unbind_speed_stack0	0.4505ms	0.3360ms	2.9759 KOps/s	3.0248 KOps/s	$\color{#d91a1a}-1.62\%$
test_unbind_speed_stack1	0.1167s	0.8594ms	1.1636 KOps/s	1.3283 KOps/s	$\textbf{\color{#d91a1a}-12.40\%}$
test_split	0.1153s	1.2821ms	779.9890 Ops/s	728.4951 Ops/s	$\textbf{\color{#35bf28}+7.07\%}$
test_chunk	1.1865ms	1.0875ms	919.5752 Ops/s	921.3132 Ops/s	$\color{#d91a1a}-0.19\%$
test_to_cpu_blocking	20.8986ms	19.4652ms	51.3737 Ops/s	49.9333 Ops/s	$\color{#35bf28}+2.88\%$
test_to_cpu_global_sync	11.8254ms	11.5825ms	86.3371 Ops/s	85.7101 Ops/s	$\color{#35bf28}+0.73\%$
test_to_cpu_event_sync	12.8501ms	12.4944ms	80.0357 Ops/s	70.0706 Ops/s	$\textbf{\color{#35bf28}+14.22\%}$
test_to_cpu_default	12.9022ms	12.4620ms	80.2438 Ops/s	78.7913 Ops/s	$\color{#35bf28}+1.84\%$
test_consolidate[False-None]	4.2809ms	4.2136ms	237.3282 Ops/s	236.5130 Ops/s	$\color{#35bf28}+0.34\%$
test_consolidate[default-None]	3.5383ms	2.1521ms	464.6622 Ops/s	459.7922 Ops/s	$\color{#35bf28}+1.06\%$
test_consolidate[reduce-overhead-None]	3.5219ms	2.0847ms	479.6886 Ops/s	472.6212 Ops/s	$\color{#35bf28}+1.50\%$
test_consolidate_njt[False-None]	8.7346ms	8.6641ms	115.4183 Ops/s	116.1760 Ops/s	$\color{#d91a1a}-0.65\%$
test_to[False-False-None]	2.2094ms	2.1474ms	465.6858 Ops/s	457.6589 Ops/s	$\color{#35bf28}+1.75\%$
test_to[True-False-None]	2.0981ms	1.9307ms	517.9389 Ops/s	522.0149 Ops/s	$\color{#d91a1a}-0.78\%$
test_to[within-False-None]	6.5449ms	6.2574ms	159.8109 Ops/s	160.2561 Ops/s	$\color{#d91a1a}-0.28\%$
test_to[True-default-None]	11.0523ms	10.8751ms	91.9528 Ops/s	91.7683 Ops/s	$\color{#35bf28}+0.20\%$
test_to_njt[False-False-None]	8.7555ms	8.6555ms	115.5341 Ops/s	115.2169 Ops/s	$\color{#35bf28}+0.28\%$
test_to_njt[True-False-None]	7.1603ms	7.0248ms	142.3524 Ops/s	143.7280 Ops/s	$\color{#d91a1a}-0.96\%$
test_to_njt[within-False-None]	15.8616ms	15.7121ms	63.6453 Ops/s	63.7931 Ops/s	$\color{#d91a1a}-0.23\%$
test_creation[device0]	0.3881ms	0.1170ms	8.5475 KOps/s	8.6034 KOps/s	$\color{#d91a1a}-0.65\%$
test_creation_from_tensor	0.4458ms	0.1136ms	8.8033 KOps/s	8.8284 KOps/s	$\color{#d91a1a}-0.28\%$
test_add_one[memmap_tensor0]	0.3873ms	6.7657μs	147.8052 KOps/s	149.6722 KOps/s	$\color{#d91a1a}-1.25\%$
test_contiguous[memmap_tensor0]	15.8410μs	0.7189μs	1.3910 MOps/s	2.1671 MOps/s	$\textbf{\color{#d91a1a}-35.81\%}$
test_stack[memmap_tensor0]	93.7330μs	4.2745μs	233.9448 KOps/s	237.9049 KOps/s	$\color{#d91a1a}-1.66\%$
test_memmaptd_index	1.2635ms	0.2792ms	3.5818 KOps/s	3.6654 KOps/s	$\color{#d91a1a}-2.28\%$
test_memmaptd_index_astensor	0.5428ms	0.3826ms	2.6139 KOps/s	2.6392 KOps/s	$\color{#d91a1a}-0.96\%$
test_memmaptd_index_op	1.0688ms	0.6365ms	1.5712 KOps/s	1.5748 KOps/s	$\color{#d91a1a}-0.23\%$
test_serialize_model	0.1395s	0.1346s	7.4274 Ops/s	7.3732 Ops/s	$\color{#35bf28}+0.74\%$
test_serialize_model_pickle	1.3494s	1.2095s	0.8268 Ops/s	0.8234 Ops/s	$\color{#35bf28}+0.41\%$
test_serialize_weights	0.1416s	0.1364s	7.3335 Ops/s	7.4526 Ops/s	$\color{#d91a1a}-1.60\%$
test_serialize_weights_returnearly	78.3444ms	64.2018ms	15.5759 Ops/s	10.9453 Ops/s	$\textbf{\color{#35bf28}+42.31\%}$
test_serialize_weights_pickle	1.4202s	1.2146s	0.8233 Ops/s	0.8171 Ops/s	$\color{#35bf28}+0.76\%$
test_reshape_pytree	0.2225ms	32.9116μs	30.3844 KOps/s	30.4300 KOps/s	$\color{#d91a1a}-0.15\%$
test_reshape_td	86.4030μs	44.2826μs	22.5822 KOps/s	22.6388 KOps/s	$\color{#d91a1a}-0.25\%$
test_view_pytree	0.2188ms	31.5996μs	31.6460 KOps/s	30.7800 KOps/s	$\color{#35bf28}+2.81\%$
test_view_td	0.1048ms	55.2127μs	18.1118 KOps/s	18.3564 KOps/s	$\color{#d91a1a}-1.33\%$
test_unbind_pytree	0.2434ms	37.2213μs	26.8664 KOps/s	26.0482 KOps/s	$\color{#35bf28}+3.14\%$
test_unbind_td	0.1686ms	51.0430μs	19.5913 KOps/s	19.4131 KOps/s	$\color{#35bf28}+0.92\%$
test_split_pytree	0.2427ms	43.0931μs	23.2056 KOps/s	22.7972 KOps/s	$\color{#35bf28}+1.79\%$
test_split_td	0.1057ms	66.1964μs	15.1066 KOps/s	14.1899 KOps/s	$\textbf{\color{#35bf28}+6.46\%}$
test_add_pytree	0.2151ms	42.1348μs	23.7334 KOps/s	23.3446 KOps/s	$\color{#35bf28}+1.67\%$
test_add_td	0.1188ms	56.9123μs	17.5709 KOps/s	18.0581 KOps/s	$\color{#d91a1a}-2.70\%$
test_compile_add_one_nested[tensordict-compile]	0.3500ms	0.2362ms	4.2344 KOps/s	4.4253 KOps/s	$\color{#d91a1a}-4.31\%$
test_compile_add_one_nested[tensordict-eager]	0.4497ms	0.1991ms	5.0226 KOps/s	4.9849 KOps/s	$\color{#35bf28}+0.76\%$
test_compile_add_one_nested[pytree-compile]	0.3167ms	0.1375ms	7.2716 KOps/s	7.0579 KOps/s	$\color{#35bf28}+3.03\%$
test_compile_add_one_nested[pytree-eager]	0.6149ms	0.1746ms	5.7285 KOps/s	5.6751 KOps/s	$\color{#35bf28}+0.94\%$
test_compile_copy_nested[tensordict-compile]	0.4646ms	16.7449μs	59.7196 KOps/s	66.5960 KOps/s	$\textbf{\color{#d91a1a}-10.33\%}$
test_compile_copy_nested[tensordict-eager]	0.3902ms	53.4132μs	18.7220 KOps/s	18.4030 KOps/s	$\color{#35bf28}+1.73\%$
test_compile_copy_nested[pytree-compile]	0.1039ms	15.6666μs	63.8299 KOps/s	63.9836 KOps/s	$\color{#d91a1a}-0.24\%$
test_compile_copy_nested[pytree-eager]	0.2775ms	67.6434μs	14.7834 KOps/s	14.4772 KOps/s	$\color{#35bf28}+2.12\%$
test_compile_add_one_flat[tensordict-compile]	0.3584ms	0.2545ms	3.9295 KOps/s	3.6982 KOps/s	$\textbf{\color{#35bf28}+6.25\%}$
test_compile_add_one_flat[tensordict-eager]	0.3297ms	0.2853ms	3.5048 KOps/s	3.4638 KOps/s	$\color{#35bf28}+1.18\%$
test_compile_add_one_flat[tensorclass-compile]	0.2863ms	0.1873ms	5.3381 KOps/s	5.1459 KOps/s	$\color{#35bf28}+3.73\%$
test_compile_add_one_flat[tensorclass-eager]	0.1195ms	75.0332μs	13.3274 KOps/s	13.3669 KOps/s	$\color{#d91a1a}-0.30\%$
test_compile_add_one_flat[pytree-compile]	0.3278ms	0.1968ms	5.0801 KOps/s	4.9464 KOps/s	$\color{#35bf28}+2.70\%$
test_compile_add_one_flat[pytree-eager]	0.7930ms	0.5095ms	1.9626 KOps/s	1.9672 KOps/s	$\color{#d91a1a}-0.24\%$
test_compile_add_self_flat[tensordict-eager]	0.3702ms	0.3323ms	3.0095 KOps/s	2.8387 KOps/s	$\textbf{\color{#35bf28}+6.02\%}$
test_compile_add_self_flat[tensordict-compile]	0.3083ms	0.2551ms	3.9205 KOps/s	3.5169 KOps/s	$\textbf{\color{#35bf28}+11.48\%}$
test_compile_add_self_flat[tensorclass-eager]	0.1341ms	91.3875μs	10.9424 KOps/s	10.4444 KOps/s	$\color{#35bf28}+4.77\%$
test_compile_add_self_flat[tensorclass-compile]	0.4522ms	0.1911ms	5.2335 KOps/s	4.8458 KOps/s	$\textbf{\color{#35bf28}+8.00\%}$
test_compile_add_self_flat[pytree-eager]	0.6605ms	0.4147ms	2.4116 KOps/s	2.3306 KOps/s	$\color{#35bf28}+3.48\%$
test_compile_add_self_flat[pytree-compile]	0.2703ms	0.1979ms	5.0534 KOps/s	4.7710 KOps/s	$\textbf{\color{#35bf28}+5.92\%}$
test_compile_copy_flat[tensordict-compile]	62.2720μs	20.0354μs	49.9117 KOps/s	49.7532 KOps/s	$\color{#35bf28}+0.32\%$
test_compile_copy_flat[tensordict-eager]	77.1120μs	42.0234μs	23.7963 KOps/s	24.5038 KOps/s	$\color{#d91a1a}-2.89\%$
test_compile_copy_flat[pytree-compile]	0.2223ms	22.2308μs	44.9825 KOps/s	42.1539 KOps/s	$\textbf{\color{#35bf28}+6.71\%}$
test_compile_copy_flat[pytree-eager]	0.3443ms	53.0815μs	18.8389 KOps/s	18.8726 KOps/s	$\color{#d91a1a}-0.18\%$
test_compile_assign_and_add[tensordict-compile]	2.3089ms	0.2424ms	4.1251 KOps/s	3.8934 KOps/s	$\textbf{\color{#35bf28}+5.95\%}$
test_compile_assign_and_add[tensordict-eager]	3.4519ms	3.2662ms	306.1642 Ops/s	304.6255 Ops/s	$\color{#35bf28}+0.51\%$
test_compile_assign_and_add[pytree-compile]	2.2302ms	0.1846ms	5.4177 KOps/s	5.3915 KOps/s	$\color{#35bf28}+0.49\%$
test_compile_assign_and_add[pytree-eager]	2.9412ms	2.6708ms	374.4161 Ops/s	364.2339 Ops/s	$\color{#35bf28}+2.80\%$
test_compile_indexing[tensor-tensordict-compile]	0.2931ms	0.1363ms	7.3362 KOps/s	7.2408 KOps/s	$\color{#35bf28}+1.32\%$
test_compile_indexing[tensor-tensordict-eager]	0.3808ms	73.9078μs	13.5304 KOps/s	12.8428 KOps/s	$\textbf{\color{#35bf28}+5.35\%}$
test_compile_indexing[tensor-tensorclass-compile]	0.2779ms	0.1207ms	8.2835 KOps/s	8.3318 KOps/s	$\color{#d91a1a}-0.58\%$
test_compile_indexing[tensor-tensorclass-eager]	0.2982ms	44.3682μs	22.5387 KOps/s	22.3018 KOps/s	$\color{#35bf28}+1.06\%$
test_compile_indexing[tensor-pytree-compile]	0.1626ms	0.1215ms	8.2285 KOps/s	8.3693 KOps/s	$\color{#d91a1a}-1.68\%$
test_compile_indexing[tensor-pytree-eager]	0.2480ms	44.6141μs	22.4144 KOps/s	22.2363 KOps/s	$\color{#35bf28}+0.80\%$
test_compile_indexing[slice-tensordict-compile]	0.2706ms	85.5748μs	11.6857 KOps/s	11.7745 KOps/s	$\color{#d91a1a}-0.75\%$
test_compile_indexing[slice-tensordict-eager]	0.2150ms	30.6070μs	32.6723 KOps/s	35.8036 KOps/s	$\textbf{\color{#d91a1a}-8.75\%}$
test_compile_indexing[slice-tensorclass-compile]	0.1973ms	72.5778μs	13.7783 KOps/s	14.2227 KOps/s	$\color{#d91a1a}-3.12\%$
test_compile_indexing[slice-tensorclass-eager]	0.2545ms	22.3979μs	44.6470 KOps/s	44.0055 KOps/s	$\color{#35bf28}+1.46\%$
test_compile_indexing[slice-pytree-compile]	0.1138ms	72.6482μs	13.7650 KOps/s	13.9402 KOps/s	$\color{#d91a1a}-1.26\%$
test_compile_indexing[slice-pytree-eager]	0.2891ms	22.3604μs	44.7219 KOps/s	44.2516 KOps/s	$\color{#35bf28}+1.06\%$
test_compile_indexing[int-tensordict-compile]	0.1253ms	85.9670μs	11.6324 KOps/s	11.6650 KOps/s	$\color{#d91a1a}-0.28\%$
test_compile_indexing[int-tensordict-eager]	0.2586ms	28.0018μs	35.7120 KOps/s	35.6305 KOps/s	$\color{#35bf28}+0.23\%$
test_compile_indexing[int-tensorclass-compile]	0.1134ms	71.9929μs	13.8903 KOps/s	13.9194 KOps/s	$\color{#d91a1a}-0.21\%$
test_compile_indexing[int-tensorclass-eager]	0.2302ms	22.2246μs	44.9951 KOps/s	44.3639 KOps/s	$\color{#35bf28}+1.42\%$
test_compile_indexing[int-pytree-compile]	0.1119ms	72.3549μs	13.8208 KOps/s	13.9378 KOps/s	$\color{#d91a1a}-0.84\%$
test_compile_indexing[int-pytree-eager]	0.2608ms	22.3761μs	44.6905 KOps/s	44.1997 KOps/s	$\color{#35bf28}+1.11\%$
test_compile_replace[single-eager]	89.4220μs	47.7211μs	20.9551 KOps/s	19.9946 KOps/s	$\color{#35bf28}+4.80\%$
test_compile_replace[single-compile]	0.2528ms	0.1325ms	7.5453 KOps/s	7.5789 KOps/s	$\color{#d91a1a}-0.44\%$
test_compile_replace[multi-eager]	0.6621ms	0.5669ms	1.7640 KOps/s	1.7656 KOps/s	$\color{#d91a1a}-0.09\%$
test_compile_replace[multi-compile]	0.3274ms	0.1400ms	7.1437 KOps/s	7.2771 KOps/s	$\color{#d91a1a}-1.83\%$
test_compile_tc_getattr_20[eager]	0.2133ms	0.1570ms	6.3711 KOps/s	6.2238 KOps/s	$\color{#35bf28}+2.37\%$
test_compile_tc_getattr_20[compile]	0.3420ms	0.1527ms	6.5495 KOps/s	6.6705 KOps/s	$\color{#d91a1a}-1.81\%$
test_compile_clone_shallow[20-eager]	61.3020μs	19.6183μs	50.9727 KOps/s	52.0514 KOps/s	$\color{#d91a1a}-2.07\%$
test_compile_clone_shallow[20-compile]	8.6302ms	18.0091μs	55.5276 KOps/s	59.0628 KOps/s	$\textbf{\color{#d91a1a}-5.99\%}$
test_compile_clone_shallow[40-eager]	71.9320μs	34.3022μs	29.1526 KOps/s	29.6224 KOps/s	$\color{#d91a1a}-1.59\%$
test_compile_clone_shallow[40-compile]	78.4420μs	18.8104μs	53.1621 KOps/s	54.2664 KOps/s	$\color{#d91a1a}-2.03\%$
test_compile_clone_shallow[80-eager]	93.9820μs	63.6723μs	15.7054 KOps/s	16.1116 KOps/s	$\color{#d91a1a}-2.52\%$
test_compile_clone_shallow[80-compile]	91.6030μs	22.3262μs	44.7905 KOps/s	45.1714 KOps/s	$\color{#d91a1a}-0.84\%$
test_compile_update_inplace[eager]	0.2264ms	61.3009μs	16.3130 KOps/s	16.2080 KOps/s	$\color{#35bf28}+0.65\%$
test_compile_update_inplace[compile]	0.4147ms	0.1490ms	6.7133 KOps/s	6.5656 KOps/s	$\color{#35bf28}+2.25\%$
test_mod_add[eager]	0.2831ms	0.2267ms	4.4119 KOps/s	4.4025 KOps/s	$\color{#35bf28}+0.21\%$
test_mod_add[compile]	0.2059ms	0.1324ms	7.5527 KOps/s	7.3631 KOps/s	$\color{#35bf28}+2.58\%$
test_mod_add[compile-overhead]	0.2887ms	0.1804ms	5.5419 KOps/s	5.5605 KOps/s	$\color{#d91a1a}-0.33\%$
test_mod_wrap[eager]	0.5489ms	0.4693ms	2.1309 KOps/s	2.1038 KOps/s	$\color{#35bf28}+1.29\%$
test_mod_wrap[compile]	0.5390ms	0.3956ms	2.5280 KOps/s	2.4909 KOps/s	$\color{#35bf28}+1.49\%$
test_mod_wrap[compile-overhead]	9.4321ms	5.0392ms	198.4439 Ops/s	218.8547 Ops/s	$\textbf{\color{#d91a1a}-9.33\%}$
test_mod_wrap_and_backward[eager]	1.9330ms	1.6841ms	593.7919 Ops/s	584.9627 Ops/s	$\color{#35bf28}+1.51\%$
test_mod_wrap_and_backward[compile]	2.5750ms	1.6427ms	608.7594 Ops/s	649.0604 Ops/s	$\textbf{\color{#d91a1a}-6.21\%}$
test_mod_wrap_and_backward[compile-overhead]	1.8684ms	1.0637ms	940.0840 Ops/s	1.0412 KOps/s	$\textbf{\color{#d91a1a}-9.72\%}$
test_seq_add[eager]	0.5786ms	0.5127ms	1.9505 KOps/s	1.9622 KOps/s	$\color{#d91a1a}-0.60\%$
test_seq_add[compile]	0.2468ms	0.1411ms	7.0852 KOps/s	6.8429 KOps/s	$\color{#35bf28}+3.54\%$
test_seq_add[compile-overhead]	0.2408ms	0.1855ms	5.3914 KOps/s	5.2461 KOps/s	$\color{#35bf28}+2.77\%$
test_seq_wrap[eager]	1.5303ms	1.4159ms	706.2648 Ops/s	706.6196 Ops/s	$\color{#d91a1a}-0.05\%$
test_seq_wrap[compile]	0.4972ms	0.4102ms	2.4378 KOps/s	2.3835 KOps/s	$\color{#35bf28}+2.28\%$
test_seq_wrap[compile-overhead]	0.4336ms	0.3102ms	3.2237 KOps/s	3.2232 KOps/s	$\color{#35bf28}+0.01\%$
test_func_call_runtime[False-eager]	0.9434ms	0.8407ms	1.1895 KOps/s	1.2420 KOps/s	$\color{#d91a1a}-4.23\%$
test_func_call_runtime[False-compile]	1.1058ms	0.9851ms	1.0152 KOps/s	1.0108 KOps/s	$\color{#35bf28}+0.44\%$
test_func_call_runtime[False-compile-overhead]	0.5864ms	0.5216ms	1.9172 KOps/s	1.9177 KOps/s	$\color{#d91a1a}-0.03\%$
test_func_call_runtime[True-eager]	1.1608ms	1.0663ms	937.8168 Ops/s	939.8920 Ops/s	$\color{#d91a1a}-0.22\%$
test_func_call_runtime[True-compile]	1.1174ms	1.0054ms	994.6499 Ops/s	988.8921 Ops/s	$\color{#35bf28}+0.58\%$
test_func_call_runtime[True-compile-overhead]	0.6039ms	0.5350ms	1.8692 KOps/s	1.8607 KOps/s	$\color{#35bf28}+0.46\%$
test_func_call_cm_runtime[False-eager]	0.8960ms	0.8271ms	1.2090 KOps/s	1.2621 KOps/s	$\color{#d91a1a}-4.20\%$
test_func_call_cm_runtime[False-compile]	1.1045ms	0.9849ms	1.0153 KOps/s	1.0252 KOps/s	$\color{#d91a1a}-0.96\%$
test_func_call_cm_runtime[False-compile-overhead]	0.6946ms	0.4939ms	2.0249 KOps/s	2.0000 KOps/s	$\color{#35bf28}+1.24\%$
test_func_call_cm_runtime[True-eager]	1.3142ms	1.2069ms	828.5594 Ops/s	823.5992 Ops/s	$\color{#35bf28}+0.60\%$
test_func_call_cm_runtime[True-compile]	1.1162ms	1.0188ms	981.5651 Ops/s	972.9510 Ops/s	$\color{#35bf28}+0.89\%$
test_func_call_cm_runtime[True-compile-overhead]	0.6282ms	0.5525ms	1.8100 KOps/s	1.7788 KOps/s	$\color{#35bf28}+1.75\%$
test_vmap_func_call_cm_runtime[eager]	2.8755ms	2.4031ms	416.1266 Ops/s	414.4031 Ops/s	$\color{#35bf28}+0.42\%$
test_vmap_func_call_cm_runtime[compile]	3.0838ms	2.6098ms	383.1747 Ops/s	385.7176 Ops/s	$\color{#d91a1a}-0.66\%$
test_vmap_func_call_cm_runtime[compile-overhead]	3.1055ms	2.6125ms	382.7726 Ops/s	385.3603 Ops/s	$\color{#d91a1a}-0.67\%$
test_distributed	3.1012ms	0.1684ms	5.9395 KOps/s	6.4135 KOps/s	$\textbf{\color{#d91a1a}-7.39\%}$
test_tdmodule	0.4736ms	0.1922ms	5.2035 KOps/s	5.0997 KOps/s	$\color{#35bf28}+2.03\%$
test_tdmodule_dispatch	0.4027ms	0.3708ms	2.6968 KOps/s	2.6682 KOps/s	$\color{#35bf28}+1.07\%$
test_tdseq	46.3010μs	26.9022μs	37.1717 KOps/s	37.6120 KOps/s	$\color{#d91a1a}-1.17\%$
test_tdseq_dispatch	0.2801ms	0.2183ms	4.5814 KOps/s	4.6134 KOps/s	$\color{#d91a1a}-0.70\%$
test_instantiation_functorch	2.1597ms	2.0697ms	483.1730 Ops/s	482.0510 Ops/s	$\color{#35bf28}+0.23\%$
test_exec_functorch	0.2318ms	0.1804ms	5.5435 KOps/s	5.4914 KOps/s	$\color{#35bf28}+0.95\%$
test_exec_functional_call	0.2407ms	0.1594ms	6.2746 KOps/s	6.2372 KOps/s	$\color{#35bf28}+0.60\%$
test_exec_td_decorator	0.4955ms	0.2484ms	4.0252 KOps/s	3.9975 KOps/s	$\color{#35bf28}+0.69\%$
test_vmap_mlp_speed_decorator[True-True]	1.2742ms	1.0335ms	967.6094 Ops/s	957.7768 Ops/s	$\color{#35bf28}+1.03\%$
test_vmap_mlp_speed_decorator[True-False]	1.2318ms	1.0362ms	965.1063 Ops/s	961.7027 Ops/s	$\color{#35bf28}+0.35\%$
test_vmap_mlp_speed_decorator[False-True]	0.9451ms	0.7373ms	1.3564 KOps/s	1.3525 KOps/s	$\color{#35bf28}+0.29\%$
test_vmap_mlp_speed_decorator[False-False]	0.9285ms	0.7381ms	1.3548 KOps/s	1.3552 KOps/s	$\color{#d91a1a}-0.03\%$
test_vmap_transformer_speed_decorator[True-True]	21.7769ms	20.9964ms	47.6272 Ops/s	47.3831 Ops/s	$\color{#35bf28}+0.52\%$
test_vmap_transformer_speed_decorator[True-False]	21.8053ms	21.0439ms	47.5197 Ops/s	47.2960 Ops/s	$\color{#35bf28}+0.47\%$
test_vmap_transformer_speed_decorator[False-True]	21.2401ms	20.6033ms	48.5360 Ops/s	48.2800 Ops/s	$\color{#35bf28}+0.53\%$
test_vmap_transformer_speed_decorator[False-False]	21.1557ms	20.6214ms	48.4934 Ops/s	48.4046 Ops/s	$\color{#35bf28}+0.18\%$
test_to_module_speed[True]	1.7388ms	1.6563ms	603.7388 Ops/s	598.0720 Ops/s	$\color{#35bf28}+0.95\%$
test_to_module_speed[False]	1.7368ms	1.6515ms	605.5243 Ops/s	619.8231 Ops/s	$\color{#d91a1a}-2.31\%$
test_to_module_plain_tensor_speed[None]	2.5498ms	1.7539ms	570.1663 Ops/s	576.9414 Ops/s	$\color{#d91a1a}-1.17\%$
test_to_module_plain_tensor_speed[False]	2.1820ms	1.5840ms	631.3265 Ops/s	633.1253 Ops/s	$\color{#d91a1a}-0.28\%$
test_to_module_plain_tensor_speed[True]	3.5346ms	3.4316ms	291.4126 Ops/s	290.4849 Ops/s	$\color{#35bf28}+0.32\%$
test_tc_init	70.5120μs	42.6407μs	23.4517 KOps/s	23.1064 KOps/s	$\color{#35bf28}+1.49\%$
test_tc_init_tensor_only	40.6710μs	9.2352μs	108.2809 KOps/s	110.4116 KOps/s	$\color{#d91a1a}-1.93\%$
test_tc_init_nested	0.1191ms	84.7664μs	11.7971 KOps/s	11.7837 KOps/s	$\color{#35bf28}+0.11\%$
test_tc_init_many_fields	48.2810μs	15.0881μs	66.2775 KOps/s	67.6619 KOps/s	$\color{#d91a1a}-2.05\%$
test_tc_first_layer_tensor	34.9210μs	1.8138μs	551.3380 KOps/s	549.5883 KOps/s	$\color{#35bf28}+0.32\%$
test_tc_first_layer_tensor_only	3.1001μs	0.4048μs	2.4701 MOps/s	2.5001 MOps/s	$\color{#d91a1a}-1.20\%$
test_tc_first_layer_tensor_set	24.7900μs	3.9114μs	255.6617 KOps/s	252.6906 KOps/s	$\color{#35bf28}+1.18\%$
test_tc_first_layer_tensor_only_set	35.4200μs	3.2580μs	306.9349 KOps/s	303.9696 KOps/s	$\color{#35bf28}+0.98\%$
test_tc_first_layer_nontensor	35.6310μs	6.1834μs	161.7235 KOps/s	161.9682 KOps/s	$\color{#d91a1a}-0.15\%$
test_tc_second_layer_tensor	22.5310μs	4.4024μs	227.1510 KOps/s	228.0365 KOps/s	$\color{#d91a1a}-0.39\%$
test_tc_second_layer_nontensor	31.8510μs	8.6488μs	115.6232 KOps/s	115.0254 KOps/s	$\color{#35bf28}+0.52\%$
test_unbind	0.2769s	14.1164ms	70.8397 Ops/s	65.4076 Ops/s	$\textbf{\color{#35bf28}+8.31\%}$
test_full_like	4.7596ms	4.4295ms	225.7597 Ops/s	226.1880 Ops/s	$\color{#d91a1a}-0.19\%$
test_zeros_like	4.6411ms	4.4094ms	226.7890 Ops/s	227.9413 Ops/s	$\color{#d91a1a}-0.51\%$
test_ones_like	4.8624ms	4.4261ms	225.9323 Ops/s	227.2131 Ops/s	$\color{#d91a1a}-0.56\%$
test_clone	7.3198ms	6.7167ms	148.8837 Ops/s	150.6180 Ops/s	$\color{#d91a1a}-1.15\%$
test_squeeze	69.9820μs	14.4420μs	69.2426 KOps/s	69.4890 KOps/s	$\color{#d91a1a}-0.35\%$
test_unsqueeze	0.3067ms	0.1122ms	8.9087 KOps/s	8.8644 KOps/s	$\color{#35bf28}+0.50\%$
test_split	0.2464ms	0.1806ms	5.5361 KOps/s	5.4742 KOps/s	$\color{#35bf28}+1.13\%$
test_permute	0.2730ms	0.2078ms	4.8117 KOps/s	4.6607 KOps/s	$\color{#35bf28}+3.24\%$
test_stack	20.5645ms	18.9495ms	52.7719 Ops/s	22.9878 Ops/s	$\textbf{\color{#35bf28}+129.56\%}$
test_cat	19.8167ms	18.8601ms	53.0220 Ops/s	22.6090 Ops/s	$\textbf{\color{#35bf28}+134.52\%}$
test_sequential_tensordict	0.9306ms	0.5687ms	1.7583 KOps/s	1.6379 KOps/s	$\textbf{\color{#35bf28}+7.35\%}$
test_sequential_graph_module	0.1859ms	0.1164ms	8.5880 KOps/s	8.3653 KOps/s	$\color{#35bf28}+2.66\%$
test_nested_tensordict	1.2128ms	0.8171ms	1.2238 KOps/s	1.2206 KOps/s	$\color{#35bf28}+0.26\%$
test_nested_graph_module	0.2096ms	0.1262ms	7.9243 KOps/s	7.6718 KOps/s	$\color{#35bf28}+3.29\%$

[vmoens]

Added a second commit [CI] Fix broken logging call in test_distributed exposed by pytest 9.1.

The CPU/macOS jobs were failing on TestDTensor::test_dtensor (and the FSDP tests) with TypeError: not all arguments converted during string formatting. Root cause is unrelated to the pad change:

• test/test_distributed.py had tdlogger.info("start method already set to", mp.get_start_method()) — a logging call whose message lacks a %s placeholder, so msg % args raises.
• This was latent for a long time (the logging framework swallowed the formatting error) until pytest 9.1.0 (landed between Jun 12 and Jun 15) changed its log-capture handler to let the exception propagate, turning it into a hard test failure. Confirmed: passing runs used pytest 9.0.3, failing runs use 9.1.0.

The fix adds the missing %s placeholder. The remaining red jobs are the GPU / Unit-tests (RL) jobs, which are pre-existing failures on main (flaky distributed/RL-GPU), not introduced here.