[Feature] TensorDict.logsumexp

[vmoens]

Stack from ghstack (oldest at bottom):

• [Feature] broadcast pointwise ops for tensor/tensordict mixed inputs #1166
• [Feature] TensorDict.clamp #1165
• [BugFix] Fix unitary ops for tensorclass #1164
• [Feature] TensorDict.softmax #1163
• -> [Feature] TensorDict.logsumexp #1162

[github-actions]

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

Total Benchmarks: 217. Improved: $\large\color{#35bf28}15$. Worsened: $\large\color{#d91a1a}10$.

Expand to view detailed results

Name	Max	Mean	Ops	Ops on Repo HEAD	Change
test_plain_set_nested	50.4340μs	20.4703μs	48.8513 KOps/s	49.3359 KOps/s	$\color{#d91a1a}-0.98\%$
test_plain_set_stack_nested	71.9850μs	20.7550μs	48.1811 KOps/s	48.9888 KOps/s	$\color{#d91a1a}-1.65\%$
test_plain_set_nested_inplace	70.0810μs	22.3978μs	44.6473 KOps/s	45.3499 KOps/s	$\color{#d91a1a}-1.55\%$
test_plain_set_stack_nested_inplace	62.7380μs	22.3389μs	44.7649 KOps/s	45.3778 KOps/s	$\color{#d91a1a}-1.35\%$
test_items	41.9590μs	4.2437μs	235.6446 KOps/s	233.6879 KOps/s	$\color{#35bf28}+0.84\%$
test_items_nested	0.5891ms	0.4059ms	2.4637 KOps/s	2.5035 KOps/s	$\color{#d91a1a}-1.59\%$
test_items_nested_locked	0.5983ms	0.4044ms	2.4726 KOps/s	2.4820 KOps/s	$\color{#d91a1a}-0.38\%$
test_items_nested_leaf	0.1444ms	75.6109μs	13.2256 KOps/s	13.0567 KOps/s	$\color{#35bf28}+1.29\%$
test_items_stack_nested	0.5601ms	0.4084ms	2.4485 KOps/s	2.3384 KOps/s	$\color{#35bf28}+4.71\%$
test_items_stack_nested_leaf	0.1514ms	79.9433μs	12.5089 KOps/s	12.9124 KOps/s	$\color{#d91a1a}-3.12\%$
test_items_stack_nested_locked	0.6408ms	0.4082ms	2.4499 KOps/s	2.4772 KOps/s	$\color{#d91a1a}-1.10\%$
test_keys	35.6260μs	3.6883μs	271.1303 KOps/s	286.7006 KOps/s	$\textbf{\color{#d91a1a}-5.43\%}$
test_keys_nested	0.2694ms	0.1654ms	6.0442 KOps/s	6.1159 KOps/s	$\color{#d91a1a}-1.17\%$
test_keys_nested_locked	0.8289ms	0.1712ms	5.8422 KOps/s	5.8866 KOps/s	$\color{#d91a1a}-0.75\%$
test_keys_nested_leaf	1.9405ms	0.1434ms	6.9749 KOps/s	7.0659 KOps/s	$\color{#d91a1a}-1.29\%$
test_keys_stack_nested	0.2713ms	0.1621ms	6.1681 KOps/s	6.1285 KOps/s	$\color{#35bf28}+0.65\%$
test_keys_stack_nested_leaf	0.2706ms	0.1411ms	7.0857 KOps/s	7.0084 KOps/s	$\color{#35bf28}+1.10\%$
test_keys_stack_nested_locked	0.3015ms	0.1680ms	5.9532 KOps/s	5.8985 KOps/s	$\color{#35bf28}+0.93\%$
test_values	8.3776μs	1.0678μs	936.5387 KOps/s	970.4972 KOps/s	$\color{#d91a1a}-3.50\%$
test_values_nested	0.1156ms	63.2229μs	15.8170 KOps/s	16.1697 KOps/s	$\color{#d91a1a}-2.18\%$
test_values_nested_locked	0.1278ms	63.2148μs	15.8191 KOps/s	15.7226 KOps/s	$\color{#35bf28}+0.61\%$
test_values_nested_leaf	0.1296ms	71.3069μs	14.0239 KOps/s	13.9983 KOps/s	$\color{#35bf28}+0.18\%$
test_values_stack_nested	0.1093ms	63.0715μs	15.8550 KOps/s	16.1473 KOps/s	$\color{#d91a1a}-1.81\%$
test_values_stack_nested_leaf	0.1370ms	71.6848μs	13.9500 KOps/s	13.8904 KOps/s	$\color{#35bf28}+0.43\%$
test_values_stack_nested_locked	0.1279ms	63.0124μs	15.8699 KOps/s	16.0729 KOps/s	$\color{#d91a1a}-1.26\%$
test_membership	90.2490μs	0.9374μs	1.0667 MOps/s	1.0763 MOps/s	$\color{#d91a1a}-0.88\%$
test_membership_nested	24.0950μs	2.9700μs	336.7004 KOps/s	343.3502 KOps/s	$\color{#d91a1a}-1.94\%$
test_membership_nested_leaf	34.1940μs	3.0246μs	330.6174 KOps/s	335.2337 KOps/s	$\color{#d91a1a}-1.38\%$
test_membership_stacked_nested	27.2910μs	2.9329μs	340.9614 KOps/s	345.4431 KOps/s	$\color{#d91a1a}-1.30\%$
test_membership_stacked_nested_leaf	37.8210μs	2.9459μs	339.4570 KOps/s	347.1989 KOps/s	$\color{#d91a1a}-2.23\%$
test_membership_nested_last	63.4290μs	4.4898μs	222.7277 KOps/s	232.4412 KOps/s	$\color{#d91a1a}-4.18\%$
test_membership_nested_leaf_last	34.1140μs	4.5082μs	221.8201 KOps/s	233.4658 KOps/s	$\color{#d91a1a}-4.99\%$
test_membership_stacked_nested_last	34.6050μs	7.0489μs	141.8656 KOps/s	231.0250 KOps/s	$\textbf{\color{#d91a1a}-38.59\%}$
test_membership_stacked_nested_leaf_last	34.9260μs	7.0732μs	141.3782 KOps/s	230.8089 KOps/s	$\textbf{\color{#d91a1a}-38.75\%}$
test_nested_getleaf	0.1080ms	11.0052μs	90.8665 KOps/s	91.9170 KOps/s	$\color{#d91a1a}-1.14\%$
test_nested_get	58.2590μs	10.4240μs	95.9326 KOps/s	96.2426 KOps/s	$\color{#d91a1a}-0.32\%$
test_stacked_getleaf	39.7850μs	10.9492μs	91.3309 KOps/s	92.1377 KOps/s	$\color{#d91a1a}-0.88\%$
test_stacked_get	35.8270μs	10.4723μs	95.4904 KOps/s	96.5395 KOps/s	$\color{#d91a1a}-1.09\%$
test_nested_getitemleaf	45.3750μs	11.5202μs	86.8037 KOps/s	90.1878 KOps/s	$\color{#d91a1a}-3.75\%$
test_nested_getitem	43.4510μs	10.8116μs	92.4931 KOps/s	96.2714 KOps/s	$\color{#d91a1a}-3.92\%$
test_stacked_getitemleaf	67.8570μs	11.3644μs	87.9940 KOps/s	90.6273 KOps/s	$\color{#d91a1a}-2.91\%$
test_stacked_getitem	52.8700μs	10.8774μs	91.9340 KOps/s	95.2799 KOps/s	$\color{#d91a1a}-3.51\%$
test_lock_nested	1.9022ms	0.4659ms	2.1464 KOps/s	2.1805 KOps/s	$\color{#d91a1a}-1.56\%$
test_lock_stack_nested	0.9231ms	0.4322ms	2.3138 KOps/s	2.3233 KOps/s	$\color{#d91a1a}-0.41\%$
test_unlock_nested	0.9308ms	0.3839ms	2.6052 KOps/s	2.6004 KOps/s	$\color{#35bf28}+0.18\%$
test_unlock_stack_nested	0.4515ms	0.3492ms	2.8636 KOps/s	2.8578 KOps/s	$\color{#35bf28}+0.20\%$
test_flatten_speed	0.1775ms	0.1005ms	9.9529 KOps/s	10.1095 KOps/s	$\color{#d91a1a}-1.55\%$
test_unflatten_speed	0.6379ms	0.5319ms	1.8801 KOps/s	1.9262 KOps/s	$\color{#d91a1a}-2.39\%$
test_common_ops	4.0306ms	0.8343ms	1.1985 KOps/s	1.2836 KOps/s	$\textbf{\color{#d91a1a}-6.63\%}$
test_creation	39.5050μs	2.5197μs	396.8779 KOps/s	409.9017 KOps/s	$\color{#d91a1a}-3.18\%$
test_creation_empty	36.2070μs	11.4103μs	87.6399 KOps/s	97.3942 KOps/s	$\textbf{\color{#d91a1a}-10.02\%}$
test_creation_nested_1	1.3290ms	14.1622μs	70.6104 KOps/s	75.0557 KOps/s	$\textbf{\color{#d91a1a}-5.92\%}$
test_creation_nested_2	44.3940μs	18.8613μs	53.0187 KOps/s	55.3232 KOps/s	$\color{#d91a1a}-4.17\%$
test_clone	0.3547ms	13.9531μs	71.6686 KOps/s	72.0804 KOps/s	$\color{#d91a1a}-0.57\%$
test_getitem[int]	0.7651ms	12.9733μs	77.0815 KOps/s	75.9413 KOps/s	$\color{#35bf28}+1.50\%$
test_getitem[slice_int]	0.1641ms	24.5352μs	40.7577 KOps/s	39.6311 KOps/s	$\color{#35bf28}+2.84\%$
test_getitem[range]	0.2064ms	48.6305μs	20.5632 KOps/s	20.3704 KOps/s	$\color{#35bf28}+0.95\%$
test_getitem[tuple]	0.1350ms	20.0197μs	49.9507 KOps/s	48.5602 KOps/s	$\color{#35bf28}+2.86\%$
test_getitem[list]	0.9478ms	44.9454μs	22.2492 KOps/s	22.3017 KOps/s	$\color{#d91a1a}-0.24\%$
test_setitem_dim[int]	51.3160μs	26.9956μs	37.0431 KOps/s	37.0793 KOps/s	$\color{#d91a1a}-0.10\%$
test_setitem_dim[slice_int]	0.1061ms	55.2237μs	18.1082 KOps/s	19.3797 KOps/s	$\textbf{\color{#d91a1a}-6.56\%}$
test_setitem_dim[range]	0.1311ms	75.4050μs	13.2617 KOps/s	13.6884 KOps/s	$\color{#d91a1a}-3.12\%$
test_setitem_dim[tuple]	75.1410μs	41.8945μs	23.8695 KOps/s	24.7841 KOps/s	$\color{#d91a1a}-3.69\%$
test_setitem	0.3708ms	21.1655μs	47.2467 KOps/s	49.6354 KOps/s	$\color{#d91a1a}-4.81\%$
test_set	0.4471ms	20.5694μs	48.6158 KOps/s	50.1295 KOps/s	$\color{#d91a1a}-3.02\%$
test_set_shared	1.4198ms	0.1805ms	5.5396 KOps/s	5.8032 KOps/s	$\color{#d91a1a}-4.54\%$
test_update	0.3904ms	23.6808μs	42.2282 KOps/s	45.5893 KOps/s	$\textbf{\color{#d91a1a}-7.37\%}$
test_update_nested	0.4018ms	33.6478μs	29.7196 KOps/s	31.1285 KOps/s	$\color{#d91a1a}-4.53\%$
test_update__nested	0.5997ms	33.8521μs	29.5403 KOps/s	29.8669 KOps/s	$\color{#d91a1a}-1.09\%$
test_set_nested	0.3881ms	22.8092μs	43.8420 KOps/s	45.7666 KOps/s	$\color{#d91a1a}-4.21\%$
test_set_nested_new	0.3940ms	27.4287μs	36.4581 KOps/s	37.6232 KOps/s	$\color{#d91a1a}-3.10\%$
test_select	0.3885ms	44.6601μs	22.3914 KOps/s	23.5763 KOps/s	$\textbf{\color{#d91a1a}-5.03\%}$
test_select_nested	0.1249ms	63.2436μs	15.8119 KOps/s	15.8704 KOps/s	$\color{#d91a1a}-0.37\%$
test_exclude_nested	0.1503ms	82.1797μs	12.1685 KOps/s	12.2927 KOps/s	$\color{#d91a1a}-1.01\%$
test_empty[True]	0.6855ms	0.4133ms	2.4193 KOps/s	2.4537 KOps/s	$\color{#d91a1a}-1.40\%$
test_empty[False]	11.5517μs	1.4180μs	705.2149 KOps/s	724.9780 KOps/s	$\color{#d91a1a}-2.73\%$
test_unbind_speed	0.4113ms	0.2707ms	3.6942 KOps/s	3.5704 KOps/s	$\color{#35bf28}+3.47\%$
test_unbind_speed_stack0	0.4734ms	0.2655ms	3.7662 KOps/s	3.7134 KOps/s	$\color{#35bf28}+1.42\%$
test_unbind_speed_stack1	0.1199s	0.8203ms	1.2190 KOps/s	1.3253 KOps/s	$\textbf{\color{#d91a1a}-8.02\%}$
test_split	0.1183s	1.7918ms	558.1097 Ops/s	565.2210 Ops/s	$\color{#d91a1a}-1.26\%$
test_chunk	0.1195s	1.8082ms	553.0427 Ops/s	560.0776 Ops/s	$\color{#d91a1a}-1.26\%$
test_consolidate_njt[False-None]	9.1403ms	8.4921ms	117.7564 Ops/s	118.2509 Ops/s	$\color{#d91a1a}-0.42\%$
test_creation[device0]	0.2431ms	93.3784μs	10.7091 KOps/s	10.7414 KOps/s	$\color{#d91a1a}-0.30\%$
test_creation_from_tensor	3.4649ms	95.8525μs	10.4327 KOps/s	10.2171 KOps/s	$\color{#35bf28}+2.11\%$
test_add_one[memmap_tensor0]	0.4983ms	4.9439μs	202.2687 KOps/s	205.9986 KOps/s	$\color{#d91a1a}-1.81\%$
test_contiguous[memmap_tensor0]	13.8550μs	0.5394μs	1.8539 MOps/s	1.9377 MOps/s	$\color{#d91a1a}-4.33\%$
test_stack[memmap_tensor0]	0.1726ms	3.4805μs	287.3183 KOps/s	300.2551 KOps/s	$\color{#d91a1a}-4.31\%$
test_memmaptd_index	1.1262ms	0.2416ms	4.1383 KOps/s	3.9818 KOps/s	$\color{#35bf28}+3.93\%$
test_memmaptd_index_astensor	0.7206ms	0.3332ms	3.0008 KOps/s	2.9628 KOps/s	$\color{#35bf28}+1.28\%$
test_memmaptd_index_op	1.1251ms	0.6152ms	1.6254 KOps/s	1.7065 KOps/s	$\color{#d91a1a}-4.76\%$
test_serialize_model	0.1359s	0.1224s	8.1678 Ops/s	7.1137 Ops/s	$\textbf{\color{#35bf28}+14.82\%}$
test_serialize_model_pickle	0.4585s	0.4052s	2.4680 Ops/s	2.5573 Ops/s	$\color{#d91a1a}-3.49\%$
test_serialize_weights	0.1220s	0.1168s	8.5647 Ops/s	8.4934 Ops/s	$\color{#35bf28}+0.84\%$
test_serialize_weights_returnearly	0.1621s	0.1583s	6.3153 Ops/s	6.3165 Ops/s	$\color{#d91a1a}-0.02\%$
test_serialize_weights_pickle	0.4725s	0.4084s	2.4483 Ops/s	2.5275 Ops/s	$\color{#d91a1a}-3.14\%$
test_serialize_weights_filesystem	0.1537s	0.1494s	6.6956 Ops/s	6.2451 Ops/s	$\textbf{\color{#35bf28}+7.21\%}$
test_serialize_model_filesystem	0.1651s	0.1564s	6.3942 Ops/s	6.5506 Ops/s	$\color{#d91a1a}-2.39\%$
test_reshape_pytree	65.9640μs	25.9567μs	38.5257 KOps/s	37.1023 KOps/s	$\color{#35bf28}+3.84\%$
test_reshape_td	73.8480μs	33.0534μs	30.2541 KOps/s	29.5653 KOps/s	$\color{#35bf28}+2.33\%$
test_view_pytree	0.1119ms	26.4459μs	37.8131 KOps/s	36.5494 KOps/s	$\color{#35bf28}+3.46\%$
test_view_td	0.1058ms	39.7578μs	25.1523 KOps/s	23.7717 KOps/s	$\textbf{\color{#35bf28}+5.81\%}$
test_unbind_pytree	65.8740μs	29.2884μs	34.1432 KOps/s	33.0713 KOps/s	$\color{#35bf28}+3.24\%$
test_unbind_td	0.3726ms	39.7241μs	25.1736 KOps/s	24.4951 KOps/s	$\color{#35bf28}+2.77\%$
test_split_pytree	72.5670μs	29.3769μs	34.0403 KOps/s	33.7233 KOps/s	$\color{#35bf28}+0.94\%$
test_split_td	0.5821ms	44.9875μs	22.2284 KOps/s	22.2477 KOps/s	$\color{#d91a1a}-0.09\%$
test_add_pytree	0.1097ms	34.8597μs	28.6864 KOps/s	27.9083 KOps/s	$\color{#35bf28}+2.79\%$
test_add_td	0.1224ms	57.5083μs	17.3888 KOps/s	17.7053 KOps/s	$\color{#d91a1a}-1.79\%$
test_compile_add_one_nested[tensordict-compile]	0.1398ms	62.2987μs	16.0517 KOps/s	15.7510 KOps/s	$\color{#35bf28}+1.91\%$
test_compile_add_one_nested[tensordict-eager]	0.4260ms	0.1677ms	5.9630 KOps/s	5.7960 KOps/s	$\color{#35bf28}+2.88\%$
test_compile_add_one_nested[pytree-compile]	0.1171ms	45.4652μs	21.9948 KOps/s	21.3580 KOps/s	$\color{#35bf28}+2.98\%$
test_compile_add_one_nested[pytree-eager]	0.2316ms	0.1188ms	8.4188 KOps/s	8.2771 KOps/s	$\color{#35bf28}+1.71\%$
test_compile_copy_nested[tensordict-compile]	97.7210μs	26.0365μs	38.4076 KOps/s	37.1172 KOps/s	$\color{#35bf28}+3.48\%$
test_compile_copy_nested[tensordict-eager]	0.1095ms	58.8363μs	16.9963 KOps/s	16.9170 KOps/s	$\color{#35bf28}+0.47\%$
test_compile_copy_nested[pytree-compile]	0.1958ms	80.5254μs	12.4184 KOps/s	12.5442 KOps/s	$\color{#d91a1a}-1.00\%$
test_compile_copy_nested[pytree-eager]	0.1322ms	68.6821μs	14.5598 KOps/s	14.4788 KOps/s	$\color{#35bf28}+0.56\%$
test_compile_add_one_flat[tensordict-compile]	0.2124ms	0.1041ms	9.6063 KOps/s	9.3012 KOps/s	$\color{#35bf28}+3.28\%$
test_compile_add_one_flat[tensordict-eager]	0.3912ms	0.2111ms	4.7378 KOps/s	4.6495 KOps/s	$\color{#35bf28}+1.90\%$
test_compile_add_one_flat[tensorclass-compile]	0.1230ms	43.8671μs	22.7961 KOps/s	21.4819 KOps/s	$\textbf{\color{#35bf28}+6.12\%}$
test_compile_add_one_flat[tensorclass-eager]	0.5664ms	63.4410μs	15.7627 KOps/s	14.3758 KOps/s	$\textbf{\color{#35bf28}+9.65\%}$
test_compile_add_one_flat[pytree-compile]	0.1860ms	0.1033ms	9.6792 KOps/s	9.4717 KOps/s	$\color{#35bf28}+2.19\%$
test_compile_add_one_flat[pytree-eager]	0.3336ms	0.1983ms	5.0422 KOps/s	4.8421 KOps/s	$\color{#35bf28}+4.13\%$
test_compile_add_self_flat[tensordict-eager]	0.3630ms	0.2282ms	4.3830 KOps/s	4.3037 KOps/s	$\color{#35bf28}+1.84\%$
test_compile_add_self_flat[tensordict-compile]	0.2295ms	0.1058ms	9.4502 KOps/s	9.2731 KOps/s	$\color{#35bf28}+1.91\%$
test_compile_add_self_flat[tensorclass-eager]	0.1458ms	56.7768μs	17.6128 KOps/s	17.1374 KOps/s	$\color{#35bf28}+2.77\%$
test_compile_add_self_flat[tensorclass-compile]	0.1404ms	46.0671μs	21.7075 KOps/s	21.4310 KOps/s	$\color{#35bf28}+1.29\%$
test_compile_add_self_flat[pytree-eager]	0.6607ms	0.1606ms	6.2266 KOps/s	6.1915 KOps/s	$\color{#35bf28}+0.57\%$
test_compile_add_self_flat[pytree-compile]	0.2020ms	0.1029ms	9.7196 KOps/s	9.1673 KOps/s	$\textbf{\color{#35bf28}+6.02\%}$
test_compile_copy_flat[tensordict-compile]	77.0840μs	21.1570μs	47.2656 KOps/s	46.6406 KOps/s	$\color{#35bf28}+1.34\%$
test_compile_copy_flat[tensordict-eager]	0.1326ms	65.8634μs	15.1829 KOps/s	14.9418 KOps/s	$\color{#35bf28}+1.61\%$
test_compile_copy_flat[pytree-compile]	0.1509ms	81.6367μs	12.2494 KOps/s	11.9518 KOps/s	$\color{#35bf28}+2.49\%$
test_compile_copy_flat[pytree-eager]	0.1349ms	68.7606μs	14.5432 KOps/s	14.3154 KOps/s	$\color{#35bf28}+1.59\%$
test_compile_assign_and_add[tensordict-compile]	0.4154ms	0.2044ms	4.8933 KOps/s	4.5875 KOps/s	$\textbf{\color{#35bf28}+6.67\%}$
test_compile_assign_and_add[tensordict-eager]	2.0685ms	1.2980ms	770.3903 Ops/s	759.5388 Ops/s	$\color{#35bf28}+1.43\%$
test_compile_assign_and_add[pytree-compile]	0.4026ms	0.2031ms	4.9227 KOps/s	4.8005 KOps/s	$\color{#35bf28}+2.54\%$
test_compile_assign_and_add[pytree-eager]	0.9660ms	0.7694ms	1.2997 KOps/s	1.2603 KOps/s	$\color{#35bf28}+3.12\%$
test_compile_assign_and_add_stack[compile]	0.8046ms	0.4525ms	2.2097 KOps/s	2.1247 KOps/s	$\color{#35bf28}+4.00\%$
test_compile_assign_and_add_stack[eager]	2.9676ms	2.6565ms	376.4364 Ops/s	372.7363 Ops/s	$\color{#35bf28}+0.99\%$
test_compile_indexing[tensor-tensordict-compile]	0.1320ms	36.6255μs	27.3033 KOps/s	27.1747 KOps/s	$\color{#35bf28}+0.47\%$
test_compile_indexing[tensor-tensordict-eager]	0.6567ms	33.4535μs	29.8922 KOps/s	28.3626 KOps/s	$\textbf{\color{#35bf28}+5.39\%}$
test_compile_indexing[tensor-tensorclass-compile]	91.5510μs	29.0925μs	34.3732 KOps/s	32.8393 KOps/s	$\color{#35bf28}+4.67\%$
test_compile_indexing[tensor-tensorclass-eager]	0.1088ms	23.0196μs	43.4412 KOps/s	41.5721 KOps/s	$\color{#35bf28}+4.50\%$
test_compile_indexing[tensor-pytree-compile]	87.1330μs	28.6970μs	34.8469 KOps/s	32.0809 KOps/s	$\textbf{\color{#35bf28}+8.62\%}$
test_compile_indexing[tensor-pytree-eager]	85.9280μs	22.8917μs	43.6840 KOps/s	42.0259 KOps/s	$\color{#35bf28}+3.95\%$
test_compile_indexing[slice-tensordict-compile]	0.1336ms	50.8754μs	19.6559 KOps/s	19.1742 KOps/s	$\color{#35bf28}+2.51\%$
test_compile_indexing[slice-tensordict-eager]	0.6796ms	20.1305μs	49.6759 KOps/s	48.1774 KOps/s	$\color{#35bf28}+3.11\%$
test_compile_indexing[slice-tensorclass-compile]	96.9820μs	43.7743μs	22.8445 KOps/s	22.4228 KOps/s	$\color{#35bf28}+1.88\%$
test_compile_indexing[slice-tensorclass-eager]	79.2680μs	18.7443μs	53.3496 KOps/s	52.4578 KOps/s	$\color{#35bf28}+1.70\%$
test_compile_indexing[slice-pytree-compile]	0.1229ms	44.6937μs	22.3745 KOps/s	21.8632 KOps/s	$\color{#35bf28}+2.34\%$
test_compile_indexing[slice-pytree-eager]	2.2601ms	18.8128μs	53.1553 KOps/s	52.4800 KOps/s	$\color{#35bf28}+1.29\%$
test_compile_indexing[int-tensordict-compile]	0.1229ms	52.3812μs	19.0908 KOps/s	18.6831 KOps/s	$\color{#35bf28}+2.18\%$
test_compile_indexing[int-tensordict-eager]	1.0326ms	20.0080μs	49.9799 KOps/s	48.4772 KOps/s	$\color{#35bf28}+3.10\%$
test_compile_indexing[int-tensorclass-compile]	0.1286ms	44.8290μs	22.3070 KOps/s	21.8069 KOps/s	$\color{#35bf28}+2.29\%$
test_compile_indexing[int-tensorclass-eager]	85.4400μs	18.7280μs	53.3960 KOps/s	52.8610 KOps/s	$\color{#35bf28}+1.01\%$
test_compile_indexing[int-pytree-compile]	0.1294ms	44.4885μs	22.4777 KOps/s	22.0389 KOps/s	$\color{#35bf28}+1.99\%$
test_compile_indexing[int-pytree-eager]	54.1120μs	18.9228μs	52.8462 KOps/s	52.7665 KOps/s	$\color{#35bf28}+0.15\%$
test_mod_add[eager]	0.1076ms	33.9195μs	29.4816 KOps/s	29.4878 KOps/s	$\color{#d91a1a}-0.02\%$
test_mod_add[compile]	0.1071ms	46.9330μs	21.3070 KOps/s	20.7292 KOps/s	$\color{#35bf28}+2.79\%$
test_mod_add[compile-overhead]	0.1232ms	46.4324μs	21.5367 KOps/s	20.5773 KOps/s	$\color{#35bf28}+4.66\%$
test_mod_wrap[eager]	0.3664ms	0.2271ms	4.4040 KOps/s	4.3256 KOps/s	$\color{#35bf28}+1.81\%$
test_mod_wrap[compile]	0.3778ms	0.2099ms	4.7646 KOps/s	4.5985 KOps/s	$\color{#35bf28}+3.61\%$
test_mod_wrap[compile-overhead]	0.3312ms	0.2052ms	4.8729 KOps/s	4.7202 KOps/s	$\color{#35bf28}+3.23\%$
test_mod_wrap_and_backward[eager]	12.7403ms	11.2672ms	88.7533 Ops/s	79.7529 Ops/s	$\textbf{\color{#35bf28}+11.29\%}$
test_mod_wrap_and_backward[compile]	12.6312ms	11.1992ms	89.2920 Ops/s	72.9906 Ops/s	$\textbf{\color{#35bf28}+22.33\%}$
test_mod_wrap_and_backward[compile-overhead]	12.4840ms	10.9795ms	91.0789 Ops/s	73.3478 Ops/s	$\textbf{\color{#35bf28}+24.17\%}$
test_seq_add[eager]	0.2578ms	0.1168ms	8.5621 KOps/s	8.7431 KOps/s	$\color{#d91a1a}-2.07\%$
test_seq_add[compile]	0.1648ms	62.3807μs	16.0306 KOps/s	15.6364 KOps/s	$\color{#35bf28}+2.52\%$
test_seq_add[compile-overhead]	0.1781ms	60.1429μs	16.6271 KOps/s	15.9882 KOps/s	$\color{#35bf28}+4.00\%$
test_seq_wrap[eager]	0.6625ms	0.4403ms	2.2711 KOps/s	2.1511 KOps/s	$\textbf{\color{#35bf28}+5.58\%}$
test_seq_wrap[compile]	0.4621ms	0.2336ms	4.2804 KOps/s	4.2052 KOps/s	$\color{#35bf28}+1.79\%$
test_seq_wrap[compile-overhead]	0.3551ms	0.2274ms	4.3966 KOps/s	4.1933 KOps/s	$\color{#35bf28}+4.85\%$
test_func_call_runtime[False-eager]	0.8229ms	0.5497ms	1.8191 KOps/s	1.7570 KOps/s	$\color{#35bf28}+3.53\%$
test_func_call_runtime[False-compile]	0.5572ms	0.4311ms	2.3195 KOps/s	2.2871 KOps/s	$\color{#35bf28}+1.42\%$
test_func_call_runtime[False-compile-overhead]	0.5378ms	0.4300ms	2.3255 KOps/s	2.2661 KOps/s	$\color{#35bf28}+2.62\%$
test_func_call_runtime[True-eager]	1.0708ms	0.7636ms	1.3096 KOps/s	1.2791 KOps/s	$\color{#35bf28}+2.38\%$
test_func_call_runtime[True-compile]	0.7853ms	0.4686ms	2.1340 KOps/s	2.0899 KOps/s	$\color{#35bf28}+2.11\%$
test_func_call_runtime[True-compile-overhead]	0.7884ms	0.4762ms	2.0998 KOps/s	2.0615 KOps/s	$\color{#35bf28}+1.86\%$
test_func_call_cm_runtime[False-eager]	1.0559ms	0.5486ms	1.8228 KOps/s	1.7729 KOps/s	$\color{#35bf28}+2.81\%$
test_func_call_cm_runtime[False-compile]	0.5316ms	0.4305ms	2.3231 KOps/s	2.2550 KOps/s	$\color{#35bf28}+3.02\%$
test_func_call_cm_runtime[False-compile-overhead]	0.5468ms	0.4317ms	2.3164 KOps/s	2.2787 KOps/s	$\color{#35bf28}+1.66\%$
test_func_call_cm_runtime[True-eager]	1.1840ms	0.9039ms	1.1063 KOps/s	1.0671 KOps/s	$\color{#35bf28}+3.68\%$
test_func_call_cm_runtime[True-compile]	0.7390ms	0.4979ms	2.0083 KOps/s	2.0006 KOps/s	$\color{#35bf28}+0.38\%$
test_func_call_cm_runtime[True-compile-overhead]	0.8383ms	0.4993ms	2.0026 KOps/s	1.9985 KOps/s	$\color{#35bf28}+0.21\%$
test_vmap_func_call_cm_runtime[eager]	3.0998ms	1.9568ms	511.0473 Ops/s	511.7675 Ops/s	$\color{#d91a1a}-0.14\%$
test_vmap_func_call_cm_runtime[compile]	0.9784ms	0.5252ms	1.9041 KOps/s	1.8894 KOps/s	$\color{#35bf28}+0.78\%$
test_vmap_func_call_cm_runtime[compile-overhead]	0.9154ms	0.5281ms	1.8935 KOps/s	1.9009 KOps/s	$\color{#d91a1a}-0.39\%$
test_distributed	0.3863ms	0.1260ms	7.9358 KOps/s	7.7266 KOps/s	$\color{#35bf28}+2.71\%$
test_tdmodule	57.6080μs	25.7706μs	38.8039 KOps/s	39.0288 KOps/s	$\color{#d91a1a}-0.58\%$
test_tdmodule_dispatch	0.1050ms	47.3603μs	21.1147 KOps/s	21.6366 KOps/s	$\color{#d91a1a}-2.41\%$
test_tdseq	64.2500μs	28.3856μs	35.2291 KOps/s	33.9091 KOps/s	$\color{#35bf28}+3.89\%$
test_tdseq_dispatch	0.1041ms	53.0676μs	18.8439 KOps/s	18.8984 KOps/s	$\color{#d91a1a}-0.29\%$
test_instantiation_functorch	2.2491ms	1.5316ms	652.9192 Ops/s	651.2219 Ops/s	$\color{#35bf28}+0.26\%$
test_exec_functorch	0.4045ms	0.1833ms	5.4552 KOps/s	5.5549 KOps/s	$\color{#d91a1a}-1.80\%$
test_exec_functional_call	0.3188ms	0.1760ms	5.6807 KOps/s	5.7040 KOps/s	$\color{#d91a1a}-0.41\%$
test_exec_td_decorator	0.5072ms	0.2417ms	4.1372 KOps/s	4.2408 KOps/s	$\color{#d91a1a}-2.44\%$
test_vmap_mlp_speed_decorator[True-True]	1.2421ms	0.6710ms	1.4903 KOps/s	1.4843 KOps/s	$\color{#35bf28}+0.41\%$
test_vmap_mlp_speed_decorator[True-False]	1.1854ms	0.6644ms	1.5052 KOps/s	1.4820 KOps/s	$\color{#35bf28}+1.56\%$
test_vmap_mlp_speed_decorator[False-True]	0.7880ms	0.5336ms	1.8739 KOps/s	1.8054 KOps/s	$\color{#35bf28}+3.79\%$
test_vmap_mlp_speed_decorator[False-False]	0.7456ms	0.5320ms	1.8795 KOps/s	1.8321 KOps/s	$\color{#35bf28}+2.59\%$
test_to_module_speed[True]	2.1819ms	1.3388ms	746.9411 Ops/s	734.7716 Ops/s	$\color{#35bf28}+1.66\%$
test_to_module_speed[False]	1.7902ms	1.3178ms	758.8141 Ops/s	749.6107 Ops/s	$\color{#35bf28}+1.23\%$
test_tc_init	80.7410μs	46.2578μs	21.6180 KOps/s	22.4629 KOps/s	$\color{#d91a1a}-3.76\%$
test_tc_init_nested	0.1705ms	91.5840μs	10.9189 KOps/s	10.9079 KOps/s	$\color{#35bf28}+0.10\%$
test_tc_first_layer_tensor	44.6530μs	1.5848μs	630.9985 KOps/s	633.9302 KOps/s	$\color{#d91a1a}-0.46\%$
test_tc_first_layer_nontensor	38.6130μs	4.7425μs	210.8600 KOps/s	209.2564 KOps/s	$\color{#35bf28}+0.77\%$
test_tc_second_layer_tensor	48.5010μs	2.9213μs	342.3165 KOps/s	346.0670 KOps/s	$\color{#d91a1a}-1.08\%$
test_tc_second_layer_nontensor	39.7050μs	6.2125μs	160.9649 KOps/s	164.1140 KOps/s	$\color{#d91a1a}-1.92\%$
test_unbind	0.2471s	14.3891ms	69.4969 Ops/s	71.2481 Ops/s	$\color{#d91a1a}-2.46\%$
test_full_like	18.0294ms	14.8669ms	67.2634 Ops/s	65.8478 Ops/s	$\color{#35bf28}+2.15\%$
test_zeros_like	15.5734ms	7.9803ms	125.3086 Ops/s	122.6833 Ops/s	$\color{#35bf28}+2.14\%$
test_ones_like	12.7969ms	8.0294ms	124.5422 Ops/s	115.6956 Ops/s	$\textbf{\color{#35bf28}+7.65\%}$
test_clone	13.7325ms	9.8813ms	101.2010 Ops/s	95.1072 Ops/s	$\textbf{\color{#35bf28}+6.41\%}$
test_squeeze	62.9180μs	12.7914μs	78.1773 KOps/s	82.1939 KOps/s	$\color{#d91a1a}-4.89\%$
test_unsqueeze	0.3395ms	95.7065μs	10.4486 KOps/s	10.7924 KOps/s	$\color{#d91a1a}-3.19\%$
test_split	0.3366ms	0.1931ms	5.1784 KOps/s	5.0584 KOps/s	$\color{#35bf28}+2.37\%$
test_permute	0.3493ms	0.2146ms	4.6604 KOps/s	4.7719 KOps/s	$\color{#d91a1a}-2.34\%$
test_stack	35.7260ms	28.1859ms	35.4787 Ops/s	36.8882 Ops/s	$\color{#d91a1a}-3.82\%$
test_cat	31.4363ms	27.2144ms	36.7452 Ops/s	36.9260 Ops/s	$\color{#d91a1a}-0.49\%$

[github-actions]

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

Total Benchmarks: 229. Improved: $\large\color{#35bf28}34$. Worsened: $\large\color{#d91a1a}8$.

Expand to view detailed results

Name	Max	Mean	Ops	Ops on Repo HEAD	Change
test_plain_set_nested	34.7010μs	11.1826μs	89.4243 KOps/s	74.5600 KOps/s	$\textbf{\color{#35bf28}+19.94\%}$
test_plain_set_stack_nested	34.4510μs	11.3817μs	87.8606 KOps/s	74.4123 KOps/s	$\textbf{\color{#35bf28}+18.07\%}$
test_plain_set_nested_inplace	39.9820μs	12.3112μs	81.2270 KOps/s	69.2793 KOps/s	$\textbf{\color{#35bf28}+17.25\%}$
test_plain_set_stack_nested_inplace	38.3010μs	12.3162μs	81.1939 KOps/s	69.7329 KOps/s	$\textbf{\color{#35bf28}+16.44\%}$
test_items	31.3110μs	2.8790μs	347.3403 KOps/s	342.7070 KOps/s	$\color{#35bf28}+1.35\%$
test_items_nested	0.4050ms	0.3517ms	2.8437 KOps/s	2.8021 KOps/s	$\color{#35bf28}+1.48\%$
test_items_nested_locked	0.4303ms	0.3551ms	2.8160 KOps/s	2.8027 KOps/s	$\color{#35bf28}+0.47\%$
test_items_nested_leaf	87.7530μs	58.0541μs	17.2253 KOps/s	17.1160 KOps/s	$\color{#35bf28}+0.64\%$
test_items_stack_nested	0.4092ms	0.3544ms	2.8220 KOps/s	2.8087 KOps/s	$\color{#35bf28}+0.47\%$
test_items_stack_nested_leaf	93.0630μs	57.8895μs	17.2743 KOps/s	17.0672 KOps/s	$\color{#35bf28}+1.21\%$
test_items_stack_nested_locked	0.5241ms	0.3539ms	2.8256 KOps/s	2.8141 KOps/s	$\color{#35bf28}+0.41\%$
test_keys	39.8220μs	3.4304μs	291.5071 KOps/s	291.1401 KOps/s	$\color{#35bf28}+0.13\%$
test_keys_nested	0.1145ms	80.9869μs	12.3477 KOps/s	12.4072 KOps/s	$\color{#d91a1a}-0.48\%$
test_keys_nested_locked	0.7588ms	86.5035μs	11.5602 KOps/s	11.4967 KOps/s	$\color{#35bf28}+0.55\%$
test_keys_nested_leaf	2.6453ms	72.5452μs	13.7845 KOps/s	14.0064 KOps/s	$\color{#d91a1a}-1.58\%$
test_keys_stack_nested	0.1498ms	81.2697μs	12.3047 KOps/s	12.2993 KOps/s	$\color{#35bf28}+0.04\%$
test_keys_stack_nested_leaf	0.1074ms	72.3857μs	13.8149 KOps/s	13.9152 KOps/s	$\color{#d91a1a}-0.72\%$
test_keys_stack_nested_locked	0.1557ms	87.0401μs	11.4890 KOps/s	11.3864 KOps/s	$\color{#35bf28}+0.90\%$
test_values	5.7802μs	0.8529μs	1.1725 MOps/s	1.1827 MOps/s	$\color{#d91a1a}-0.86\%$
test_values_nested	69.8030μs	34.3595μs	29.1041 KOps/s	29.2888 KOps/s	$\color{#d91a1a}-0.63\%$
test_values_nested_locked	73.9820μs	36.2205μs	27.6086 KOps/s	27.9115 KOps/s	$\color{#d91a1a}-1.08\%$
test_values_nested_leaf	91.2640μs	38.5109μs	25.9667 KOps/s	25.6846 KOps/s	$\color{#35bf28}+1.10\%$
test_values_stack_nested	67.8230μs	34.0764μs	29.3459 KOps/s	28.9468 KOps/s	$\color{#35bf28}+1.38\%$
test_values_stack_nested_leaf	75.6520μs	38.6954μs	25.8429 KOps/s	25.3450 KOps/s	$\color{#35bf28}+1.96\%$
test_values_stack_nested_locked	68.3020μs	35.7107μs	28.0028 KOps/s	27.8549 KOps/s	$\color{#35bf28}+0.53\%$
test_membership	1.9050μs	0.5365μs	1.8641 MOps/s	1.8813 MOps/s	$\color{#d91a1a}-0.91\%$
test_membership_nested	15.8005μs	1.9657μs	508.7137 KOps/s	492.0330 KOps/s	$\color{#35bf28}+3.39\%$
test_membership_nested_leaf	13.8355μs	1.9329μs	517.3530 KOps/s	509.4365 KOps/s	$\color{#35bf28}+1.55\%$
test_membership_stacked_nested	26.5710μs	2.0250μs	493.8366 KOps/s	485.9877 KOps/s	$\color{#35bf28}+1.62\%$
test_membership_stacked_nested_leaf	25.5710μs	2.0177μs	495.6146 KOps/s	487.4319 KOps/s	$\color{#35bf28}+1.68\%$
test_membership_nested_last	37.8510μs	3.0524μs	327.6153 KOps/s	338.4295 KOps/s	$\color{#d91a1a}-3.20\%$
test_membership_nested_leaf_last	25.8010μs	3.0225μs	330.8509 KOps/s	330.2875 KOps/s	$\color{#35bf28}+0.17\%$
test_membership_stacked_nested_last	35.4720μs	3.0187μs	331.2659 KOps/s	327.3293 KOps/s	$\color{#35bf28}+1.20\%$
test_membership_stacked_nested_leaf_last	26.8810μs	3.0532μs	327.5245 KOps/s	325.8161 KOps/s	$\color{#35bf28}+0.52\%$
test_nested_getleaf	31.5910μs	6.1891μs	161.5746 KOps/s	163.6444 KOps/s	$\color{#d91a1a}-1.26\%$
test_nested_get	31.4520μs	5.7901μs	172.7077 KOps/s	172.5397 KOps/s	$\color{#35bf28}+0.10\%$
test_stacked_getleaf	29.8710μs	6.1744μs	161.9603 KOps/s	161.7827 KOps/s	$\color{#35bf28}+0.11\%$
test_stacked_get	30.2010μs	5.8194μs	171.8389 KOps/s	172.6538 KOps/s	$\color{#d91a1a}-0.47\%$
test_nested_getitemleaf	30.6810μs	6.2491μs	160.0230 KOps/s	160.8836 KOps/s	$\color{#d91a1a}-0.53\%$
test_nested_getitem	29.8210μs	5.9463μs	168.1707 KOps/s	169.9736 KOps/s	$\color{#d91a1a}-1.06\%$
test_stacked_getitemleaf	37.0010μs	6.2359μs	160.3611 KOps/s	160.5497 KOps/s	$\color{#d91a1a}-0.12\%$
test_stacked_getitem	31.4820μs	5.9519μs	168.0125 KOps/s	170.3483 KOps/s	$\color{#d91a1a}-1.37\%$
test_lock_nested	2.3440ms	0.3686ms	2.7129 KOps/s	2.6286 KOps/s	$\color{#35bf28}+3.21\%$
test_lock_stack_nested	0.3902ms	0.3385ms	2.9543 KOps/s	2.9384 KOps/s	$\color{#35bf28}+0.54\%$
test_unlock_nested	0.6362ms	0.3056ms	3.2728 KOps/s	3.2349 KOps/s	$\color{#35bf28}+1.17\%$
test_unlock_stack_nested	0.3395ms	0.2789ms	3.5859 KOps/s	3.5890 KOps/s	$\color{#d91a1a}-0.08\%$
test_flatten_speed	0.1170ms	73.8946μs	13.5328 KOps/s	13.5342 KOps/s	$\color{#d91a1a}-0.01\%$
test_unflatten_speed	0.3898ms	0.3143ms	3.1813 KOps/s	3.1380 KOps/s	$\color{#35bf28}+1.38\%$
test_common_ops	1.5298ms	0.5662ms	1.7660 KOps/s	1.5782 KOps/s	$\textbf{\color{#35bf28}+11.90\%}$
test_creation	0.1003ms	1.7115μs	584.2768 KOps/s	583.7447 KOps/s	$\color{#35bf28}+0.09\%$
test_creation_empty	33.2310μs	6.2092μs	161.0509 KOps/s	95.9365 KOps/s	$\textbf{\color{#35bf28}+67.87\%}$
test_creation_nested_1	82.5430μs	7.7759μs	128.6030 KOps/s	83.6483 KOps/s	$\textbf{\color{#35bf28}+53.74\%}$
test_creation_nested_2	32.4920μs	10.4917μs	95.3131 KOps/s	67.7429 KOps/s	$\textbf{\color{#35bf28}+40.70\%}$
test_clone	33.7310μs	10.7975μs	92.6141 KOps/s	93.9472 KOps/s	$\color{#d91a1a}-1.42\%$
test_getitem[int]	1.9934ms	10.4663μs	95.5446 KOps/s	95.2679 KOps/s	$\color{#35bf28}+0.29\%$
test_getitem[slice_int]	0.1069ms	20.6486μs	48.4294 KOps/s	49.1041 KOps/s	$\color{#d91a1a}-1.37\%$
test_getitem[range]	0.1323ms	35.5880μs	28.0994 KOps/s	27.9438 KOps/s	$\color{#35bf28}+0.56\%$
test_getitem[tuple]	0.1159ms	17.8381μs	56.0599 KOps/s	56.3356 KOps/s	$\color{#d91a1a}-0.49\%$
test_getitem[list]	0.1306ms	32.1403μs	31.1136 KOps/s	30.5611 KOps/s	$\color{#35bf28}+1.81\%$
test_setitem_dim[int]	36.6320μs	18.2385μs	54.8289 KOps/s	53.1392 KOps/s	$\color{#35bf28}+3.18\%$
test_setitem_dim[slice_int]	68.3920μs	37.9782μs	26.3309 KOps/s	26.1562 KOps/s	$\color{#35bf28}+0.67\%$
test_setitem_dim[range]	73.8320μs	51.0880μs	19.5741 KOps/s	19.4625 KOps/s	$\color{#35bf28}+0.57\%$
test_setitem_dim[tuple]	53.0020μs	30.9979μs	32.2603 KOps/s	31.4887 KOps/s	$\color{#35bf28}+2.45\%$
test_setitem	50.0420μs	14.2077μs	70.3845 KOps/s	62.3595 KOps/s	$\textbf{\color{#35bf28}+12.87\%}$
test_set	46.0010μs	13.4858μs	74.1521 KOps/s	63.7782 KOps/s	$\textbf{\color{#35bf28}+16.27\%}$
test_set_shared	1.6336ms	0.1504ms	6.6509 KOps/s	6.6563 KOps/s	$\color{#d91a1a}-0.08\%$
test_update	0.3329ms	15.6165μs	64.0347 KOps/s	50.9493 KOps/s	$\textbf{\color{#35bf28}+25.68\%}$
test_update_nested	0.1308ms	20.7009μs	48.3072 KOps/s	39.0852 KOps/s	$\textbf{\color{#35bf28}+23.59\%}$
test_update__nested	0.5866ms	25.0753μs	39.8799 KOps/s	40.6615 KOps/s	$\color{#d91a1a}-1.92\%$
test_set_nested	0.1222ms	15.0166μs	66.5931 KOps/s	59.3573 KOps/s	$\textbf{\color{#35bf28}+12.19\%}$
test_set_nested_new	0.1353ms	17.3399μs	57.6703 KOps/s	52.3297 KOps/s	$\textbf{\color{#35bf28}+10.21\%}$
test_select	59.0120μs	29.5560μs	33.8341 KOps/s	32.9088 KOps/s	$\color{#35bf28}+2.81\%$
test_select_nested	75.9630μs	43.3134μs	23.0875 KOps/s	22.9035 KOps/s	$\color{#35bf28}+0.80\%$
test_exclude_nested	98.8540μs	61.5568μs	16.2452 KOps/s	15.8880 KOps/s	$\color{#35bf28}+2.25\%$
test_empty[True]	0.3559ms	0.2841ms	3.5204 KOps/s	3.4973 KOps/s	$\color{#35bf28}+0.66\%$
test_empty[False]	4.4312μs	0.8354μs	1.1970 MOps/s	1.1827 MOps/s	$\color{#35bf28}+1.21\%$
test_to	86.0830μs	58.4019μs	17.1227 KOps/s	18.0035 KOps/s	$\color{#d91a1a}-4.89\%$
test_to_nonblocking	94.8240μs	46.4267μs	21.5394 KOps/s	21.1403 KOps/s	$\color{#35bf28}+1.89\%$
test_unbind_speed	1.5072ms	0.2317ms	4.3166 KOps/s	4.2537 KOps/s	$\color{#35bf28}+1.48\%$
test_unbind_speed_stack0	0.2771ms	0.2290ms	4.3671 KOps/s	4.2382 KOps/s	$\color{#35bf28}+3.04\%$
test_unbind_speed_stack1	92.6569ms	0.6656ms	1.5023 KOps/s	1.4762 KOps/s	$\color{#35bf28}+1.77\%$
test_split	94.2614ms	1.7219ms	580.7513 Ops/s	581.8324 Ops/s	$\color{#d91a1a}-0.19\%$
test_chunk	95.9325ms	1.5889ms	629.3794 Ops/s	693.5531 Ops/s	$\textbf{\color{#d91a1a}-9.25\%}$
test_consolidate[False-None]	2.8052ms	2.6543ms	376.7537 Ops/s	374.1646 Ops/s	$\color{#35bf28}+0.69\%$
test_consolidate[default-None]	1.7215ms	1.6386ms	610.2632 Ops/s	610.4653 Ops/s	$\color{#d91a1a}-0.03\%$
test_consolidate[reduce-overhead-None]	1.7668ms	1.6926ms	590.7919 Ops/s	597.5385 Ops/s	$\color{#d91a1a}-1.13\%$
test_consolidate_njt[False-None]	6.6304ms	6.4047ms	156.1358 Ops/s	158.2437 Ops/s	$\color{#d91a1a}-1.33\%$
test_to[False-False-None]	1.8453ms	1.7685ms	565.4575 Ops/s	574.5291 Ops/s	$\color{#d91a1a}-1.58\%$
test_to[True-False-None]	1.5164ms	1.2951ms	772.1285 Ops/s	784.4977 Ops/s	$\color{#d91a1a}-1.58\%$
test_to[within-False-None]	4.3522ms	4.0257ms	248.4029 Ops/s	245.7512 Ops/s	$\color{#35bf28}+1.08\%$
test_to[True-default-None]	5.4757ms	5.2565ms	190.2392 Ops/s	192.0730 Ops/s	$\color{#d91a1a}-0.95\%$
test_to_njt[False-False-None]	7.0400ms	6.8207ms	146.6131 Ops/s	145.3592 Ops/s	$\color{#35bf28}+0.86\%$
test_to_njt[True-False-None]	5.6598ms	5.4945ms	182.0007 Ops/s	185.1377 Ops/s	$\color{#d91a1a}-1.69\%$
test_to_njt[within-False-None]	12.2565ms	12.0042ms	83.3040 Ops/s	84.1532 Ops/s	$\color{#d91a1a}-1.01\%$
test_creation[device0]	0.4572ms	79.9583μs	12.5065 KOps/s	12.6041 KOps/s	$\color{#d91a1a}-0.77\%$
test_creation_from_tensor	0.5317ms	83.3568μs	11.9966 KOps/s	12.1425 KOps/s	$\color{#d91a1a}-1.20\%$
test_add_one[memmap_tensor0]	0.3927ms	6.9646μs	143.5825 KOps/s	147.8943 KOps/s	$\color{#d91a1a}-2.92\%$
test_contiguous[memmap_tensor0]	1.8321μs	0.4083μs	2.4494 MOps/s	2.4281 MOps/s	$\color{#35bf28}+0.88\%$
test_stack[memmap_tensor0]	43.1810μs	4.3410μs	230.3610 KOps/s	234.4584 KOps/s	$\color{#d91a1a}-1.75\%$
test_memmaptd_index	1.5756ms	0.2459ms	4.0670 KOps/s	4.0748 KOps/s	$\color{#d91a1a}-0.19\%$
test_memmaptd_index_astensor	0.6015ms	0.3069ms	3.2589 KOps/s	3.2845 KOps/s	$\color{#d91a1a}-0.78\%$
test_memmaptd_index_op	1.0354ms	0.5523ms	1.8107 KOps/s	1.6524 KOps/s	$\textbf{\color{#35bf28}+9.59\%}$
test_serialize_model	0.1319s	0.1307s	7.6523 Ops/s	7.6656 Ops/s	$\color{#d91a1a}-0.17\%$
test_serialize_model_pickle	1.3491s	1.2114s	0.8255 Ops/s	0.8243 Ops/s	$\color{#35bf28}+0.14\%$
test_serialize_weights	0.1304s	0.1298s	7.7018 Ops/s	5.4363 Ops/s	$\textbf{\color{#35bf28}+41.67\%}$
test_serialize_weights_returnearly	0.3271s	55.2289ms	18.1065 Ops/s	23.8738 Ops/s	$\textbf{\color{#d91a1a}-24.16\%}$
test_serialize_weights_pickle	1.3690s	1.2147s	0.8233 Ops/s	0.8374 Ops/s	$\color{#d91a1a}-1.69\%$
test_reshape_pytree	52.7520μs	21.6759μs	46.1342 KOps/s	45.1361 KOps/s	$\color{#35bf28}+2.21\%$
test_reshape_td	56.1820μs	25.8947μs	38.6180 KOps/s	36.5800 KOps/s	$\textbf{\color{#35bf28}+5.57\%}$
test_view_pytree	54.4320μs	21.6531μs	46.1828 KOps/s	46.1698 KOps/s	$\color{#35bf28}+0.03\%$
test_view_td	63.2620μs	28.9946μs	34.4892 KOps/s	32.3249 KOps/s	$\textbf{\color{#35bf28}+6.70\%}$
test_unbind_pytree	59.3220μs	27.6500μs	36.1664 KOps/s	36.3687 KOps/s	$\color{#d91a1a}-0.56\%$
test_unbind_td	0.7705ms	36.2153μs	27.6126 KOps/s	28.4786 KOps/s	$\color{#d91a1a}-3.04\%$
test_split_pytree	63.7020μs	29.8625μs	33.4868 KOps/s	34.0048 KOps/s	$\color{#d91a1a}-1.52\%$
test_split_td	0.9908ms	37.8714μs	26.4051 KOps/s	26.6327 KOps/s	$\color{#d91a1a}-0.85\%$
test_add_pytree	77.5620μs	34.3422μs	29.1187 KOps/s	29.8698 KOps/s	$\color{#d91a1a}-2.51\%$
test_add_td	0.1920ms	44.8118μs	22.3156 KOps/s	20.1993 KOps/s	$\textbf{\color{#35bf28}+10.48\%}$
test_compile_add_one_nested[tensordict-compile]	0.1741ms	0.1191ms	8.3986 KOps/s	8.2696 KOps/s	$\color{#35bf28}+1.56\%$
test_compile_add_one_nested[tensordict-eager]	0.2225ms	0.1289ms	7.7583 KOps/s	7.8420 KOps/s	$\color{#d91a1a}-1.07\%$
test_compile_add_one_nested[pytree-compile]	0.1882ms	96.8632μs	10.3238 KOps/s	10.6706 KOps/s	$\color{#d91a1a}-3.25\%$
test_compile_add_one_nested[pytree-eager]	0.2072ms	0.1512ms	6.6128 KOps/s	6.6108 KOps/s	$\color{#35bf28}+0.03\%$
test_compile_copy_nested[tensordict-compile]	54.1610μs	22.2418μs	44.9603 KOps/s	43.6363 KOps/s	$\color{#35bf28}+3.03\%$
test_compile_copy_nested[tensordict-eager]	56.7820μs	28.6312μs	34.9270 KOps/s	34.0963 KOps/s	$\color{#35bf28}+2.44\%$
test_compile_copy_nested[pytree-compile]	0.2207ms	64.4382μs	15.5187 KOps/s	15.4522 KOps/s	$\color{#35bf28}+0.43\%$
test_compile_copy_nested[pytree-eager]	85.0030μs	49.1937μs	20.3278 KOps/s	20.3331 KOps/s	$\color{#d91a1a}-0.03\%$
test_compile_add_one_flat[tensordict-compile]	0.1812ms	0.1409ms	7.0958 KOps/s	7.0778 KOps/s	$\color{#35bf28}+0.25\%$
test_compile_add_one_flat[tensordict-eager]	0.3014ms	0.2121ms	4.7150 KOps/s	4.7333 KOps/s	$\color{#d91a1a}-0.39\%$
test_compile_add_one_flat[tensorclass-compile]	0.2207ms	99.3796μs	10.0624 KOps/s	10.3871 KOps/s	$\color{#d91a1a}-3.13\%$
test_compile_add_one_flat[tensorclass-eager]	0.1577ms	56.8304μs	17.5962 KOps/s	19.3316 KOps/s	$\textbf{\color{#d91a1a}-8.98\%}$
test_compile_add_one_flat[pytree-compile]	0.1782ms	0.1347ms	7.4259 KOps/s	7.3976 KOps/s	$\color{#35bf28}+0.38\%$
test_compile_add_one_flat[pytree-eager]	0.5863ms	0.4939ms	2.0246 KOps/s	2.0623 KOps/s	$\color{#d91a1a}-1.83\%$
test_compile_add_self_flat[tensordict-eager]	0.3698ms	0.2557ms	3.9111 KOps/s	3.8854 KOps/s	$\color{#35bf28}+0.66\%$
test_compile_add_self_flat[tensordict-compile]	0.1925ms	0.1425ms	7.0175 KOps/s	7.0685 KOps/s	$\color{#d91a1a}-0.72\%$
test_compile_add_self_flat[tensorclass-eager]	0.1412ms	65.0329μs	15.3768 KOps/s	15.8480 KOps/s	$\color{#d91a1a}-2.97\%$
test_compile_add_self_flat[tensorclass-compile]	0.1505ms	97.9901μs	10.2051 KOps/s	10.4407 KOps/s	$\color{#d91a1a}-2.26\%$
test_compile_add_self_flat[pytree-eager]	0.4677ms	0.4203ms	2.3793 KOps/s	2.4545 KOps/s	$\color{#d91a1a}-3.06\%$
test_compile_add_self_flat[pytree-compile]	0.1735ms	0.1345ms	7.4325 KOps/s	7.3305 KOps/s	$\color{#35bf28}+1.39\%$
test_compile_copy_flat[tensordict-compile]	60.1030μs	19.1064μs	52.3385 KOps/s	55.1272 KOps/s	$\textbf{\color{#d91a1a}-5.06\%}$
test_compile_copy_flat[tensordict-eager]	74.1630μs	31.1171μs	32.1367 KOps/s	31.8551 KOps/s	$\color{#35bf28}+0.88\%$
test_compile_copy_flat[pytree-compile]	0.1829ms	70.7968μs	14.1249 KOps/s	14.5590 KOps/s	$\color{#d91a1a}-2.98\%$
test_compile_copy_flat[pytree-eager]	86.1030μs	51.8485μs	19.2870 KOps/s	19.7155 KOps/s	$\color{#d91a1a}-2.17\%$
test_compile_assign_and_add[tensordict-compile]	1.6089ms	0.3874ms	2.5816 KOps/s	2.1419 KOps/s	$\textbf{\color{#35bf28}+20.53\%}$
test_compile_assign_and_add[tensordict-eager]	2.8407ms	2.6319ms	379.9570 Ops/s	380.4529 Ops/s	$\color{#d91a1a}-0.13\%$
test_compile_assign_and_add[pytree-compile]	1.5685ms	0.3756ms	2.6622 KOps/s	2.2558 KOps/s	$\textbf{\color{#35bf28}+18.01\%}$
test_compile_assign_and_add[pytree-eager]	3.1570ms	2.6933ms	371.2954 Ops/s	370.2126 Ops/s	$\color{#35bf28}+0.29\%$
test_compile_indexing[tensor-tensordict-compile]	0.6352ms	0.1157ms	8.6452 KOps/s	8.8079 KOps/s	$\color{#d91a1a}-1.85\%$
test_compile_indexing[tensor-tensordict-eager]	0.5720ms	79.7242μs	12.5432 KOps/s	11.9920 KOps/s	$\color{#35bf28}+4.60\%$
test_compile_indexing[tensor-tensorclass-compile]	0.5608ms	0.1074ms	9.3153 KOps/s	9.5119 KOps/s	$\color{#d91a1a}-2.07\%$
test_compile_indexing[tensor-tensorclass-eager]	0.4954ms	68.5071μs	14.5970 KOps/s	14.3774 KOps/s	$\color{#35bf28}+1.53\%$
test_compile_indexing[tensor-pytree-compile]	0.5540ms	0.1088ms	9.1871 KOps/s	9.4391 KOps/s	$\color{#d91a1a}-2.67\%$
test_compile_indexing[tensor-pytree-eager]	0.5008ms	68.8363μs	14.5272 KOps/s	14.5160 KOps/s	$\color{#35bf28}+0.08\%$
test_compile_indexing[slice-tensordict-compile]	0.1678ms	0.1023ms	9.7769 KOps/s	10.1436 KOps/s	$\color{#d91a1a}-3.61\%$
test_compile_indexing[slice-tensordict-eager]	0.4451ms	16.4343μs	60.8483 KOps/s	58.7826 KOps/s	$\color{#35bf28}+3.51\%$
test_compile_indexing[slice-tensorclass-compile]	0.5261ms	98.7430μs	10.1273 KOps/s	10.3157 KOps/s	$\color{#d91a1a}-1.83\%$
test_compile_indexing[slice-tensorclass-eager]	0.4312ms	15.6806μs	63.7732 KOps/s	65.0979 KOps/s	$\color{#d91a1a}-2.03\%$
test_compile_indexing[slice-pytree-compile]	0.5152ms	98.2987μs	10.1731 KOps/s	10.4326 KOps/s	$\color{#d91a1a}-2.49\%$
test_compile_indexing[slice-pytree-eager]	51.6320μs	15.2441μs	65.5993 KOps/s	64.4069 KOps/s	$\color{#35bf28}+1.85\%$
test_compile_indexing[int-tensordict-compile]	0.5144ms	0.1035ms	9.6598 KOps/s	10.0376 KOps/s	$\color{#d91a1a}-3.76\%$
test_compile_indexing[int-tensordict-eager]	0.5931ms	16.2417μs	61.5698 KOps/s	60.1354 KOps/s	$\color{#35bf28}+2.39\%$
test_compile_indexing[int-tensorclass-compile]	0.5194ms	98.3690μs	10.1658 KOps/s	10.5240 KOps/s	$\color{#d91a1a}-3.40\%$
test_compile_indexing[int-tensorclass-eager]	0.4165ms	15.4372μs	64.7784 KOps/s	64.3365 KOps/s	$\color{#35bf28}+0.69\%$
test_compile_indexing[int-pytree-compile]	0.5028ms	96.8096μs	10.3295 KOps/s	10.5176 KOps/s	$\color{#d91a1a}-1.79\%$
test_compile_indexing[int-pytree-eager]	46.6620μs	15.5614μs	64.2614 KOps/s	65.0505 KOps/s	$\color{#d91a1a}-1.21\%$
test_mod_add[eager]	83.1030μs	36.7518μs	27.2095 KOps/s	25.7176 KOps/s	$\textbf{\color{#35bf28}+5.80\%}$
test_mod_add[compile]	0.2975ms	77.3196μs	12.9333 KOps/s	12.4937 KOps/s	$\color{#35bf28}+3.52\%$
test_mod_add[compile-overhead]	0.3303ms	0.1669ms	5.9932 KOps/s	5.7870 KOps/s	$\color{#35bf28}+3.56\%$
test_mod_wrap[eager]	0.3298ms	0.2538ms	3.9400 KOps/s	3.9625 KOps/s	$\color{#d91a1a}-0.57\%$
test_mod_wrap[compile]	0.4726ms	0.2784ms	3.5913 KOps/s	3.5537 KOps/s	$\color{#35bf28}+1.06\%$
test_mod_wrap[compile-overhead]	7.0615ms	3.7643ms	265.6532 Ops/s	250.2545 Ops/s	$\textbf{\color{#35bf28}+6.15\%}$
test_mod_wrap_and_backward[eager]	1.4833ms	1.3637ms	733.3172 Ops/s	686.0149 Ops/s	$\textbf{\color{#35bf28}+6.90\%}$
test_mod_wrap_and_backward[compile]	1.3215ms	1.2512ms	799.2266 Ops/s	730.8608 Ops/s	$\textbf{\color{#35bf28}+9.35\%}$
test_mod_wrap_and_backward[compile-overhead]	1.3897ms	0.9228ms	1.0836 KOps/s	968.5119 Ops/s	$\textbf{\color{#35bf28}+11.88\%}$
test_seq_add[eager]	0.1628ms	0.1131ms	8.8427 KOps/s	8.4423 KOps/s	$\color{#35bf28}+4.74\%$
test_seq_add[compile]	0.1375ms	90.9918μs	10.9900 KOps/s	11.4806 KOps/s	$\color{#d91a1a}-4.27\%$
test_seq_add[compile-overhead]	0.1679ms	0.1317ms	7.5921 KOps/s	7.8908 KOps/s	$\color{#d91a1a}-3.79\%$
test_seq_wrap[eager]	0.4868ms	0.4224ms	2.3675 KOps/s	2.3620 KOps/s	$\color{#35bf28}+0.23\%$
test_seq_wrap[compile]	0.3628ms	0.3094ms	3.2320 KOps/s	3.3443 KOps/s	$\color{#d91a1a}-3.36\%$
test_seq_wrap[compile-overhead]	0.2736ms	0.2216ms	4.5125 KOps/s	4.3243 KOps/s	$\color{#35bf28}+4.35\%$
test_func_call_runtime[False-eager]	0.8045ms	0.7315ms	1.3670 KOps/s	1.2790 KOps/s	$\textbf{\color{#35bf28}+6.88\%}$
test_func_call_runtime[False-compile]	0.7858ms	0.7375ms	1.3559 KOps/s	1.3534 KOps/s	$\color{#35bf28}+0.19\%$
test_func_call_runtime[False-compile-overhead]	0.4176ms	0.3572ms	2.7993 KOps/s	2.7789 KOps/s	$\color{#35bf28}+0.73\%$
test_func_call_runtime[True-eager]	0.9695ms	0.8973ms	1.1145 KOps/s	1.0994 KOps/s	$\color{#35bf28}+1.38\%$
test_func_call_runtime[True-compile]	0.8338ms	0.7580ms	1.3192 KOps/s	1.3195 KOps/s	$\color{#d91a1a}-0.02\%$
test_func_call_runtime[True-compile-overhead]	0.4362ms	0.3803ms	2.6296 KOps/s	2.6600 KOps/s	$\color{#d91a1a}-1.14\%$
test_func_call_cm_runtime[False-eager]	0.8704ms	0.7689ms	1.3005 KOps/s	1.3607 KOps/s	$\color{#d91a1a}-4.43\%$
test_func_call_cm_runtime[False-compile]	0.8278ms	0.7406ms	1.3502 KOps/s	1.3100 KOps/s	$\color{#35bf28}+3.07\%$
test_func_call_cm_runtime[False-compile-overhead]	0.4097ms	0.3630ms	2.7551 KOps/s	2.7550 KOps/s	$+0.00\%$
test_func_call_cm_runtime[True-eager]	1.1095ms	0.9982ms	1.0018 KOps/s	989.3443 Ops/s	$\color{#35bf28}+1.26\%$
test_func_call_cm_runtime[True-compile]	0.8504ms	0.7875ms	1.2698 KOps/s	1.2805 KOps/s	$\color{#d91a1a}-0.83\%$
test_func_call_cm_runtime[True-compile-overhead]	0.5761ms	0.4059ms	2.4639 KOps/s	2.4713 KOps/s	$\color{#d91a1a}-0.30\%$
test_vmap_func_call_cm_runtime[eager]	2.5738ms	2.0913ms	478.1667 Ops/s	473.4929 Ops/s	$\color{#35bf28}+0.99\%$
test_vmap_func_call_cm_runtime[compile]	0.9267ms	0.8380ms	1.1933 KOps/s	1.2412 KOps/s	$\color{#d91a1a}-3.86\%$
test_vmap_func_call_cm_runtime[compile-overhead]	0.4686ms	0.4185ms	2.3897 KOps/s	2.4529 KOps/s	$\color{#d91a1a}-2.58\%$
test_distributed	0.5918ms	0.1253ms	7.9825 KOps/s	8.6264 KOps/s	$\textbf{\color{#d91a1a}-7.46\%}$
test_tdmodule	0.2565ms	18.0765μs	55.3204 KOps/s	46.8007 KOps/s	$\textbf{\color{#35bf28}+18.20\%}$
test_tdmodule_dispatch	69.8620μs	31.7954μs	31.4511 KOps/s	26.6273 KOps/s	$\textbf{\color{#35bf28}+18.12\%}$
test_tdseq	42.2910μs	19.7659μs	50.5921 KOps/s	45.4771 KOps/s	$\textbf{\color{#35bf28}+11.25\%}$
test_tdseq_dispatch	51.3720μs	35.4612μs	28.1998 KOps/s	24.3406 KOps/s	$\textbf{\color{#35bf28}+15.85\%}$
test_instantiation_functorch	1.6656ms	1.5198ms	657.9986 Ops/s	660.2155 Ops/s	$\color{#d91a1a}-0.34\%$
test_exec_functorch	0.1886ms	0.1425ms	7.0192 KOps/s	7.0142 KOps/s	$\color{#35bf28}+0.07\%$
test_exec_functional_call	0.2278ms	0.1377ms	7.2645 KOps/s	7.3401 KOps/s	$\color{#d91a1a}-1.03\%$
test_exec_td_decorator	0.3752ms	0.1828ms	5.4714 KOps/s	5.4387 KOps/s	$\color{#35bf28}+0.60\%$
test_vmap_mlp_speed_decorator[True-True]	0.7900ms	0.6765ms	1.4781 KOps/s	1.4493 KOps/s	$\color{#35bf28}+1.99\%$
test_vmap_mlp_speed_decorator[True-False]	0.8047ms	0.6755ms	1.4805 KOps/s	1.4487 KOps/s	$\color{#35bf28}+2.20\%$
test_vmap_mlp_speed_decorator[False-True]	0.7408ms	0.6067ms	1.6483 KOps/s	1.6771 KOps/s	$\color{#d91a1a}-1.72\%$
test_vmap_mlp_speed_decorator[False-False]	0.7323ms	0.5935ms	1.6849 KOps/s	1.6678 KOps/s	$\color{#35bf28}+1.03\%$
test_vmap_transformer_speed_decorator[True-True]	19.6044ms	19.2990ms	51.8162 Ops/s	51.8509 Ops/s	$\color{#d91a1a}-0.07\%$
test_vmap_transformer_speed_decorator[True-False]	20.0805ms	19.3339ms	51.7225 Ops/s	51.7737 Ops/s	$\color{#d91a1a}-0.10\%$
test_vmap_transformer_speed_decorator[False-True]	19.4061ms	19.1917ms	52.1058 Ops/s	52.3766 Ops/s	$\color{#d91a1a}-0.52\%$
test_vmap_transformer_speed_decorator[False-False]	19.3446ms	19.2441ms	51.9640 Ops/s	52.3606 Ops/s	$\color{#d91a1a}-0.76\%$
test_to_module_speed[True]	1.0556ms	0.9564ms	1.0456 KOps/s	1.0318 KOps/s	$\color{#35bf28}+1.34\%$
test_to_module_speed[False]	1.3241ms	0.9357ms	1.0687 KOps/s	1.0439 KOps/s	$\color{#35bf28}+2.38\%$
test_tc_init	70.3320μs	33.9453μs	29.4592 KOps/s	26.0696 KOps/s	$\textbf{\color{#35bf28}+13.00\%}$
test_tc_init_nested	0.1064ms	67.6032μs	14.7922 KOps/s	12.9944 KOps/s	$\textbf{\color{#35bf28}+13.84\%}$
test_tc_first_layer_tensor	20.8900μs	0.8026μs	1.2459 MOps/s	1.4388 MOps/s	$\textbf{\color{#d91a1a}-13.41\%}$
test_tc_first_layer_nontensor	20.8010μs	2.2244μs	449.5570 KOps/s	450.5018 KOps/s	$\color{#d91a1a}-0.21\%$
test_tc_second_layer_tensor	10.7630μs	1.4094μs	709.5031 KOps/s	690.2949 KOps/s	$\color{#35bf28}+2.78\%$
test_tc_second_layer_nontensor	88.9830μs	2.9526μs	338.6878 KOps/s	334.5468 KOps/s	$\color{#35bf28}+1.24\%$
test_unbind	0.2239s	10.1016ms	98.9942 Ops/s	141.4013 Ops/s	$\textbf{\color{#d91a1a}-29.99\%}$
test_full_like	11.7459ms	9.1392ms	109.4191 Ops/s	107.0710 Ops/s	$\color{#35bf28}+2.19\%$
test_zeros_like	5.2098ms	4.3183ms	231.5724 Ops/s	237.0540 Ops/s	$\color{#d91a1a}-2.31\%$
test_ones_like	4.9650ms	4.2770ms	233.8070 Ops/s	236.9089 Ops/s	$\color{#d91a1a}-1.31\%$
test_clone	6.7667ms	6.3594ms	157.2478 Ops/s	109.9461 Ops/s	$\textbf{\color{#35bf28}+43.02\%}$
test_squeeze	58.6620μs	9.7305μs	102.7697 KOps/s	108.4870 KOps/s	$\textbf{\color{#d91a1a}-5.27\%}$
test_unsqueeze	0.1527ms	70.6127μs	14.1618 KOps/s	14.0393 KOps/s	$\color{#35bf28}+0.87\%$
test_split	0.4242ms	0.1567ms	6.3811 KOps/s	6.2762 KOps/s	$\color{#35bf28}+1.67\%$
test_permute	0.3054ms	0.1775ms	5.6341 KOps/s	5.6821 KOps/s	$\color{#d91a1a}-0.85\%$
test_stack	50.4909ms	50.1878ms	19.9252 Ops/s	19.8953 Ops/s	$\color{#35bf28}+0.15\%$
test_cat	50.4904ms	50.1247ms	19.9503 Ops/s	19.7388 Ops/s	$\color{#35bf28}+1.07\%$