[Feature] Allow update to increment the number of tds in a lazy stack

[vmoens]

Stack from ghstack (oldest at bottom):

• [BugFix] Consolidate lazy stacks of non-tensors #1222
• [Feature] capture_non_tensor_stack #1221
• -> [Feature] Allow update to increment the number of tds in a lazy stack #1220

[github-actions]

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

Total Benchmarks: 217. Improved: $\large\color{#35bf28}29$. Worsened: $\large\color{#d91a1a}9$.

Expand to view detailed results

Name	Max	Mean	Ops	Ops on Repo HEAD	Change
test_plain_set_nested	47.5390μs	20.9710μs	47.6848 KOps/s	45.9395 KOps/s	$\color{#35bf28}+3.80\%$
test_plain_set_stack_nested	55.4240μs	21.1826μs	47.2086 KOps/s	45.3991 KOps/s	$\color{#35bf28}+3.99\%$
test_plain_set_nested_inplace	53.9410μs	22.8528μs	43.7584 KOps/s	42.3928 KOps/s	$\color{#35bf28}+3.22\%$
test_plain_set_stack_nested_inplace	78.3360μs	22.9130μs	43.6434 KOps/s	43.1842 KOps/s	$\color{#35bf28}+1.06\%$
test_items	38.5320μs	4.3481μs	229.9851 KOps/s	242.1970 KOps/s	$\textbf{\color{#d91a1a}-5.04\%}$
test_items_nested	0.5306ms	0.4070ms	2.4570 KOps/s	2.4087 KOps/s	$\color{#35bf28}+2.00\%$
test_items_nested_locked	0.6234ms	0.4025ms	2.4845 KOps/s	2.4209 KOps/s	$\color{#35bf28}+2.63\%$
test_items_nested_leaf	0.1380ms	77.5382μs	12.8969 KOps/s	12.6145 KOps/s	$\color{#35bf28}+2.24\%$
test_items_stack_nested	0.5981ms	0.4041ms	2.4744 KOps/s	2.3930 KOps/s	$\color{#35bf28}+3.40\%$
test_items_stack_nested_leaf	0.1387ms	78.7203μs	12.7032 KOps/s	12.0448 KOps/s	$\textbf{\color{#35bf28}+5.47\%}$
test_items_stack_nested_locked	0.7096ms	0.4035ms	2.4785 KOps/s	2.3779 KOps/s	$\color{#35bf28}+4.23\%$
test_keys	23.2740μs	3.5050μs	285.3103 KOps/s	280.7495 KOps/s	$\color{#35bf28}+1.62\%$
test_keys_nested	0.2649ms	0.1624ms	6.1577 KOps/s	6.0311 KOps/s	$\color{#35bf28}+2.10\%$
test_keys_nested_locked	0.7716ms	0.1686ms	5.9300 KOps/s	5.7937 KOps/s	$\color{#35bf28}+2.35\%$
test_keys_nested_leaf	0.2102ms	0.1420ms	7.0425 KOps/s	6.9151 KOps/s	$\color{#35bf28}+1.84\%$
test_keys_stack_nested	0.2927ms	0.1634ms	6.1209 KOps/s	6.1968 KOps/s	$\color{#d91a1a}-1.22\%$
test_keys_stack_nested_leaf	0.2289ms	0.1422ms	7.0311 KOps/s	7.1584 KOps/s	$\color{#d91a1a}-1.78\%$
test_keys_stack_nested_locked	0.3193ms	0.1688ms	5.9244 KOps/s	6.0006 KOps/s	$\color{#d91a1a}-1.27\%$
test_values	36.0880μs	1.1913μs	839.4346 KOps/s	935.6387 KOps/s	$\textbf{\color{#d91a1a}-10.28\%}$
test_values_nested	0.1230ms	62.0990μs	16.1033 KOps/s	15.6529 KOps/s	$\color{#35bf28}+2.88\%$
test_values_nested_locked	0.1102ms	61.6517μs	16.2201 KOps/s	15.7507 KOps/s	$\color{#35bf28}+2.98\%$
test_values_nested_leaf	0.1198ms	70.6305μs	14.1582 KOps/s	13.3461 KOps/s	$\textbf{\color{#35bf28}+6.09\%}$
test_values_stack_nested	0.1158ms	62.4799μs	16.0051 KOps/s	15.7248 KOps/s	$\color{#35bf28}+1.78\%$
test_values_stack_nested_leaf	0.1229ms	70.6027μs	14.1638 KOps/s	14.0583 KOps/s	$\color{#35bf28}+0.75\%$
test_values_stack_nested_locked	0.1213ms	62.7733μs	15.9303 KOps/s	15.8867 KOps/s	$\color{#35bf28}+0.27\%$
test_membership	28.7240μs	0.8604μs	1.1622 MOps/s	1.1526 MOps/s	$\color{#35bf28}+0.84\%$
test_membership_nested	36.5690μs	2.9104μs	343.6001 KOps/s	339.8354 KOps/s	$\color{#35bf28}+1.11\%$
test_membership_nested_leaf	34.6140μs	2.9319μs	341.0785 KOps/s	342.5004 KOps/s	$\color{#d91a1a}-0.42\%$
test_membership_stacked_nested	39.3430μs	2.8552μs	350.2417 KOps/s	345.7953 KOps/s	$\color{#35bf28}+1.29\%$
test_membership_stacked_nested_leaf	21.1190μs	2.9028μs	344.4952 KOps/s	341.7338 KOps/s	$\color{#35bf28}+0.81\%$
test_membership_nested_last	37.7910μs	4.3330μs	230.7884 KOps/s	228.4639 KOps/s	$\color{#35bf28}+1.02\%$
test_membership_nested_leaf_last	39.1530μs	4.3362μs	230.6165 KOps/s	228.1520 KOps/s	$\color{#35bf28}+1.08\%$
test_membership_stacked_nested_last	24.4350μs	5.0949μs	196.2739 KOps/s	73.4898 KOps/s	$\textbf{\color{#35bf28}+167.08\%}$
test_membership_stacked_nested_leaf_last	82.6440μs	5.3680μs	186.2904 KOps/s	71.9950 KOps/s	$\textbf{\color{#35bf28}+158.75\%}$
test_nested_getleaf	54.0610μs	10.3212μs	96.8883 KOps/s	93.1301 KOps/s	$\color{#35bf28}+4.04\%$
test_nested_get	45.7060μs	9.7596μs	102.4629 KOps/s	98.5229 KOps/s	$\color{#35bf28}+4.00\%$
test_stacked_getleaf	58.0480μs	10.2383μs	97.6727 KOps/s	93.8724 KOps/s	$\color{#35bf28}+4.05\%$
test_stacked_get	56.5260μs	9.9470μs	100.5326 KOps/s	97.7707 KOps/s	$\color{#35bf28}+2.82\%$
test_nested_getitemleaf	43.1700μs	10.9187μs	91.5864 KOps/s	87.7311 KOps/s	$\color{#35bf28}+4.39\%$
test_nested_getitem	54.3110μs	10.4476μs	95.7156 KOps/s	92.4056 KOps/s	$\color{#35bf28}+3.58\%$
test_stacked_getitemleaf	34.9450μs	10.9653μs	91.1965 KOps/s	89.3051 KOps/s	$\color{#35bf28}+2.12\%$
test_stacked_getitem	64.4600μs	10.4449μs	95.7409 KOps/s	93.3780 KOps/s	$\color{#35bf28}+2.53\%$
test_lock_nested	0.5190ms	0.4018ms	2.4890 KOps/s	2.4243 KOps/s	$\color{#35bf28}+2.67\%$
test_lock_stack_nested	0.5236ms	0.4137ms	2.4173 KOps/s	2.3767 KOps/s	$\color{#35bf28}+1.71\%$
test_unlock_nested	0.3971ms	0.3272ms	3.0559 KOps/s	2.9712 KOps/s	$\color{#35bf28}+2.85\%$
test_unlock_stack_nested	0.4654ms	0.3352ms	2.9830 KOps/s	2.9285 KOps/s	$\color{#35bf28}+1.86\%$
test_flatten_speed	0.1679ms	99.3333μs	10.0671 KOps/s	9.7479 KOps/s	$\color{#35bf28}+3.28\%$
test_unflatten_speed	0.6973ms	0.5149ms	1.9423 KOps/s	1.8832 KOps/s	$\color{#35bf28}+3.14\%$
test_common_ops	5.0484ms	0.8103ms	1.2341 KOps/s	1.1970 KOps/s	$\color{#35bf28}+3.10\%$
test_creation	55.4220μs	2.4591μs	406.6555 KOps/s	396.9766 KOps/s	$\color{#35bf28}+2.44\%$
test_creation_empty	98.2530μs	12.4480μs	80.3343 KOps/s	77.5917 KOps/s	$\color{#35bf28}+3.53\%$
test_creation_nested_1	45.0140μs	15.3396μs	65.1906 KOps/s	63.2497 KOps/s	$\color{#35bf28}+3.07\%$
test_creation_nested_2	77.0030μs	19.9413μs	50.1472 KOps/s	48.1109 KOps/s	$\color{#35bf28}+4.23\%$
test_clone	0.1569ms	13.2436μs	75.5083 KOps/s	72.7569 KOps/s	$\color{#35bf28}+3.78\%$
test_getitem[int]	0.8829ms	12.7599μs	78.3704 KOps/s	78.1875 KOps/s	$\color{#35bf28}+0.23\%$
test_getitem[slice_int]	0.1256ms	23.8502μs	41.9284 KOps/s	40.6759 KOps/s	$\color{#35bf28}+3.08\%$
test_getitem[range]	0.1943ms	50.0205μs	19.9918 KOps/s	19.5331 KOps/s	$\color{#35bf28}+2.35\%$
test_getitem[tuple]	0.1283ms	20.0414μs	49.8968 KOps/s	48.7592 KOps/s	$\color{#35bf28}+2.33\%$
test_getitem[list]	0.1816ms	45.9600μs	21.7581 KOps/s	21.5686 KOps/s	$\color{#35bf28}+0.88\%$
test_setitem_dim[int]	47.6990μs	25.1569μs	39.7505 KOps/s	38.6959 KOps/s	$\color{#35bf28}+2.73\%$
test_setitem_dim[slice_int]	0.1319ms	50.3094μs	19.8770 KOps/s	19.5930 KOps/s	$\color{#35bf28}+1.45\%$
test_setitem_dim[range]	0.1234ms	74.6241μs	13.4005 KOps/s	12.8237 KOps/s	$\color{#35bf28}+4.50\%$
test_setitem_dim[tuple]	85.5500μs	39.1472μs	25.5446 KOps/s	24.5072 KOps/s	$\color{#35bf28}+4.23\%$
test_setitem	0.2403ms	20.4924μs	48.7986 KOps/s	45.7525 KOps/s	$\textbf{\color{#35bf28}+6.66\%}$
test_set	0.2020ms	20.1844μs	49.5433 KOps/s	47.5697 KOps/s	$\color{#35bf28}+4.15\%$
test_set_shared	4.3360ms	0.1800ms	5.5557 KOps/s	5.5661 KOps/s	$\color{#d91a1a}-0.19\%$
test_update	0.2446ms	23.4970μs	42.5587 KOps/s	40.5283 KOps/s	$\textbf{\color{#35bf28}+5.01\%}$
test_update_nested	0.2140ms	34.5909μs	28.9093 KOps/s	28.2041 KOps/s	$\color{#35bf28}+2.50\%$
test_update__nested	0.4553ms	33.5825μs	29.7774 KOps/s	28.4767 KOps/s	$\color{#35bf28}+4.57\%$
test_set_nested	0.1017ms	22.0899μs	45.2695 KOps/s	42.8356 KOps/s	$\textbf{\color{#35bf28}+5.68\%}$
test_set_nested_new	0.1027ms	26.2374μs	38.1135 KOps/s	35.8851 KOps/s	$\textbf{\color{#35bf28}+6.21\%}$
test_select	0.1059ms	43.6118μs	22.9296 KOps/s	22.4953 KOps/s	$\color{#35bf28}+1.93\%$
test_select_nested	0.1283ms	62.5546μs	15.9860 KOps/s	15.6726 KOps/s	$\color{#35bf28}+2.00\%$
test_exclude_nested	0.1629ms	80.9303μs	12.3563 KOps/s	12.2166 KOps/s	$\color{#35bf28}+1.14\%$
test_empty[True]	0.7299ms	0.4012ms	2.4926 KOps/s	2.4370 KOps/s	$\color{#35bf28}+2.28\%$
test_empty[False]	47.9773μs	1.4002μs	714.1586 KOps/s	722.8709 KOps/s	$\color{#d91a1a}-1.21\%$
test_unbind_speed	0.4290ms	0.2649ms	3.7747 KOps/s	3.6623 KOps/s	$\color{#35bf28}+3.07\%$
test_unbind_speed_stack0	0.4482ms	0.2633ms	3.7986 KOps/s	3.8335 KOps/s	$\color{#d91a1a}-0.91\%$
test_unbind_speed_stack1	0.1099s	0.7282ms	1.3733 KOps/s	1.3807 KOps/s	$\color{#d91a1a}-0.54\%$
test_split	0.1158s	1.7546ms	569.9277 Ops/s	502.4054 Ops/s	$\textbf{\color{#35bf28}+13.44\%}$
test_chunk	0.1092s	1.7355ms	576.2039 Ops/s	619.1901 Ops/s	$\textbf{\color{#d91a1a}-6.94\%}$
test_consolidate_njt[False-None]	9.7715ms	8.0807ms	123.7520 Ops/s	118.5728 Ops/s	$\color{#35bf28}+4.37\%$
test_creation[device0]	3.6637ms	94.1987μs	10.6159 KOps/s	10.4656 KOps/s	$\color{#35bf28}+1.44\%$
test_creation_from_tensor	0.2714ms	94.1103μs	10.6258 KOps/s	10.4892 KOps/s	$\color{#35bf28}+1.30\%$
test_add_one[memmap_tensor0]	95.1380μs	4.8022μs	208.2367 KOps/s	205.3798 KOps/s	$\color{#35bf28}+1.39\%$
test_contiguous[memmap_tensor0]	16.4400μs	0.5088μs	1.9654 MOps/s	1.9490 MOps/s	$\color{#35bf28}+0.84\%$
test_stack[memmap_tensor0]	34.9250μs	3.3186μs	301.3363 KOps/s	292.0765 KOps/s	$\color{#35bf28}+3.17\%$
test_memmaptd_index	1.3155ms	0.2246ms	4.4523 KOps/s	4.3134 KOps/s	$\color{#35bf28}+3.22\%$
test_memmaptd_index_astensor	0.7014ms	0.3128ms	3.1966 KOps/s	3.1244 KOps/s	$\color{#35bf28}+2.31\%$
test_memmaptd_index_op	1.1614ms	0.5918ms	1.6898 KOps/s	1.6440 KOps/s	$\color{#35bf28}+2.79\%$
test_serialize_model	0.2323s	0.1329s	7.5273 Ops/s	8.2217 Ops/s	$\textbf{\color{#d91a1a}-8.45\%}$
test_serialize_model_pickle	0.4462s	0.4010s	2.4939 Ops/s	2.4498 Ops/s	$\color{#35bf28}+1.80\%$
test_serialize_weights	0.1296s	0.1162s	8.6023 Ops/s	8.4445 Ops/s	$\color{#35bf28}+1.87\%$
test_serialize_weights_returnearly	0.1846s	0.1669s	5.9918 Ops/s	6.0679 Ops/s	$\color{#d91a1a}-1.25\%$
test_serialize_weights_pickle	1.2513s	0.7327s	1.3649 Ops/s	2.4935 Ops/s	$\textbf{\color{#d91a1a}-45.26\%}$
test_serialize_weights_filesystem	0.1520s	0.1452s	6.8880 Ops/s	6.1313 Ops/s	$\textbf{\color{#35bf28}+12.34\%}$
test_serialize_model_filesystem	0.1602s	0.1474s	6.7828 Ops/s	6.5327 Ops/s	$\color{#35bf28}+3.83\%$
test_reshape_pytree	73.2560μs	26.1678μs	38.2149 KOps/s	36.4052 KOps/s	$\color{#35bf28}+4.97\%$
test_reshape_td	0.1065ms	34.0690μs	29.3522 KOps/s	29.5760 KOps/s	$\color{#d91a1a}-0.76\%$
test_view_pytree	68.7180μs	26.0008μs	38.4604 KOps/s	37.2476 KOps/s	$\color{#35bf28}+3.26\%$
test_view_td	0.1201ms	40.5325μs	24.6716 KOps/s	24.0674 KOps/s	$\color{#35bf28}+2.51\%$
test_unbind_pytree	0.1110ms	28.9967μs	34.4867 KOps/s	33.1244 KOps/s	$\color{#35bf28}+4.11\%$
test_unbind_td	0.3345ms	39.4881μs	25.3241 KOps/s	24.8575 KOps/s	$\color{#35bf28}+1.88\%$
test_split_pytree	71.1820μs	28.4289μs	35.1755 KOps/s	33.5137 KOps/s	$\color{#35bf28}+4.96\%$
test_split_td	0.5320ms	44.5377μs	22.4529 KOps/s	21.5890 KOps/s	$\color{#35bf28}+4.00\%$
test_add_pytree	99.0450μs	36.2997μs	27.5484 KOps/s	28.2620 KOps/s	$\color{#d91a1a}-2.52\%$
test_add_td	0.2707ms	65.4222μs	15.2853 KOps/s	17.1198 KOps/s	$\textbf{\color{#d91a1a}-10.72\%}$
test_compile_add_one_nested[tensordict-compile]	0.1748ms	67.4365μs	14.8288 KOps/s	14.6825 KOps/s	$\color{#35bf28}+1.00\%$
test_compile_add_one_nested[tensordict-eager]	0.3760ms	0.1737ms	5.7568 KOps/s	5.7193 KOps/s	$\color{#35bf28}+0.65\%$
test_compile_add_one_nested[pytree-compile]	0.1070ms	45.4244μs	22.0146 KOps/s	21.6428 KOps/s	$\color{#35bf28}+1.72\%$
test_compile_add_one_nested[pytree-eager]	0.2381ms	0.1180ms	8.4759 KOps/s	8.3105 KOps/s	$\color{#35bf28}+1.99\%$
test_compile_copy_nested[tensordict-compile]	0.1271ms	28.6005μs	34.9644 KOps/s	34.9504 KOps/s	$\color{#35bf28}+0.04\%$
test_compile_copy_nested[tensordict-eager]	0.1188ms	59.0987μs	16.9208 KOps/s	16.8001 KOps/s	$\color{#35bf28}+0.72\%$
test_compile_copy_nested[pytree-compile]	0.1536ms	80.1640μs	12.4744 KOps/s	11.9492 KOps/s	$\color{#35bf28}+4.40\%$
test_compile_copy_nested[pytree-eager]	0.1625ms	67.3566μs	14.8464 KOps/s	14.5430 KOps/s	$\color{#35bf28}+2.09\%$
test_compile_add_one_flat[tensordict-compile]	0.2108ms	0.1072ms	9.3258 KOps/s	9.3108 KOps/s	$\color{#35bf28}+0.16\%$
test_compile_add_one_flat[tensordict-eager]	0.3915ms	0.2164ms	4.6220 KOps/s	4.6013 KOps/s	$\color{#35bf28}+0.45\%$
test_compile_add_one_flat[tensorclass-compile]	0.1197ms	46.5823μs	21.4674 KOps/s	21.1905 KOps/s	$\color{#35bf28}+1.31\%$
test_compile_add_one_flat[tensorclass-eager]	0.4002ms	69.8153μs	14.3235 KOps/s	14.6419 KOps/s	$\color{#d91a1a}-2.17\%$
test_compile_add_one_flat[pytree-compile]	0.1813ms	0.1004ms	9.9580 KOps/s	9.8929 KOps/s	$\color{#35bf28}+0.66\%$
test_compile_add_one_flat[pytree-eager]	0.3360ms	0.1979ms	5.0530 KOps/s	4.8599 KOps/s	$\color{#35bf28}+3.97\%$
test_compile_add_self_flat[tensordict-eager]	0.3984ms	0.2297ms	4.3526 KOps/s	4.2649 KOps/s	$\color{#35bf28}+2.06\%$
test_compile_add_self_flat[tensordict-compile]	0.2426ms	0.1071ms	9.3365 KOps/s	9.2269 KOps/s	$\color{#35bf28}+1.19\%$
test_compile_add_self_flat[tensorclass-eager]	0.4043ms	65.1571μs	15.3475 KOps/s	15.3377 KOps/s	$\color{#35bf28}+0.06\%$
test_compile_add_self_flat[tensorclass-compile]	0.1331ms	47.7189μs	20.9560 KOps/s	20.5017 KOps/s	$\color{#35bf28}+2.22\%$
test_compile_add_self_flat[pytree-eager]	0.8970ms	0.1573ms	6.3575 KOps/s	6.2406 KOps/s	$\color{#35bf28}+1.87\%$
test_compile_add_self_flat[pytree-compile]	0.2355ms	0.1008ms	9.9199 KOps/s	9.6602 KOps/s	$\color{#35bf28}+2.69\%$
test_compile_copy_flat[tensordict-compile]	79.0080μs	20.6165μs	48.5048 KOps/s	45.2078 KOps/s	$\textbf{\color{#35bf28}+7.29\%}$
test_compile_copy_flat[tensordict-eager]	0.3916ms	69.3567μs	14.4182 KOps/s	14.7912 KOps/s	$\color{#d91a1a}-2.52\%$
test_compile_copy_flat[pytree-compile]	0.2715ms	81.2516μs	12.3075 KOps/s	11.9244 KOps/s	$\color{#35bf28}+3.21\%$
test_compile_copy_flat[pytree-eager]	0.1933ms	66.7020μs	14.9921 KOps/s	14.6149 KOps/s	$\color{#35bf28}+2.58\%$
test_compile_assign_and_add[tensordict-compile]	0.2908ms	0.2130ms	4.6956 KOps/s	4.6192 KOps/s	$\color{#35bf28}+1.65\%$
test_compile_assign_and_add[tensordict-eager]	1.7384ms	1.3971ms	715.7786 Ops/s	704.2911 Ops/s	$\color{#35bf28}+1.63\%$
test_compile_assign_and_add[pytree-compile]	0.3424ms	0.2097ms	4.7698 KOps/s	4.7368 KOps/s	$\color{#35bf28}+0.70\%$
test_compile_assign_and_add[pytree-eager]	1.0437ms	0.8180ms	1.2226 KOps/s	1.1854 KOps/s	$\color{#35bf28}+3.13\%$
test_compile_assign_and_add_stack[compile]	0.9003ms	0.4572ms	2.1871 KOps/s	2.1817 KOps/s	$\color{#35bf28}+0.25\%$
test_compile_assign_and_add_stack[eager]	4.1117ms	2.7406ms	364.8883 Ops/s	355.6559 Ops/s	$\color{#35bf28}+2.60\%$
test_compile_indexing[tensor-tensordict-compile]	0.1004ms	38.8906μs	25.7132 KOps/s	25.3066 KOps/s	$\color{#35bf28}+1.61\%$
test_compile_indexing[tensor-tensordict-eager]	0.5475ms	33.0855μs	30.2247 KOps/s	28.5617 KOps/s	$\textbf{\color{#35bf28}+5.82\%}$
test_compile_indexing[tensor-tensorclass-compile]	80.1190μs	31.4691μs	31.7772 KOps/s	32.0703 KOps/s	$\color{#d91a1a}-0.91\%$
test_compile_indexing[tensor-tensorclass-eager]	96.9800μs	22.6679μs	44.1152 KOps/s	42.5161 KOps/s	$\color{#35bf28}+3.76\%$
test_compile_indexing[tensor-pytree-compile]	71.7440μs	31.0825μs	32.1724 KOps/s	30.7612 KOps/s	$\color{#35bf28}+4.59\%$
test_compile_indexing[tensor-pytree-eager]	69.9000μs	22.1929μs	45.0595 KOps/s	42.7921 KOps/s	$\textbf{\color{#35bf28}+5.30\%}$
test_compile_indexing[slice-tensordict-compile]	0.1638ms	53.2265μs	18.7876 KOps/s	18.5107 KOps/s	$\color{#35bf28}+1.50\%$
test_compile_indexing[slice-tensordict-eager]	0.5117ms	19.5022μs	51.2764 KOps/s	46.6792 KOps/s	$\textbf{\color{#35bf28}+9.85\%}$
test_compile_indexing[slice-tensorclass-compile]	0.1273ms	45.8451μs	21.8126 KOps/s	20.9472 KOps/s	$\color{#35bf28}+4.13\%$
test_compile_indexing[slice-tensorclass-eager]	73.8280μs	18.4346μs	54.2457 KOps/s	52.4539 KOps/s	$\color{#35bf28}+3.42\%$
test_compile_indexing[slice-pytree-compile]	0.1040ms	46.5541μs	21.4804 KOps/s	20.6908 KOps/s	$\color{#35bf28}+3.82\%$
test_compile_indexing[slice-pytree-eager]	57.7480μs	18.4156μs	54.3019 KOps/s	52.8610 KOps/s	$\color{#35bf28}+2.73\%$
test_compile_indexing[int-tensordict-compile]	0.2289ms	57.6594μs	17.3432 KOps/s	18.4474 KOps/s	$\textbf{\color{#d91a1a}-5.99\%}$
test_compile_indexing[int-tensordict-eager]	0.8632ms	19.4316μs	51.4627 KOps/s	48.0468 KOps/s	$\textbf{\color{#35bf28}+7.11\%}$
test_compile_indexing[int-tensorclass-compile]	0.1629ms	46.6994μs	21.4135 KOps/s	20.9104 KOps/s	$\color{#35bf28}+2.41\%$
test_compile_indexing[int-tensorclass-eager]	69.7700μs	18.2384μs	54.8295 KOps/s	54.1703 KOps/s	$\color{#35bf28}+1.22\%$
test_compile_indexing[int-pytree-compile]	99.5460μs	46.1395μs	21.6734 KOps/s	21.2319 KOps/s	$\color{#35bf28}+2.08\%$
test_compile_indexing[int-pytree-eager]	56.3050μs	18.1649μs	55.0512 KOps/s	53.1571 KOps/s	$\color{#35bf28}+3.56\%$
test_mod_add[eager]	0.1074ms	34.5279μs	28.9621 KOps/s	26.8650 KOps/s	$\textbf{\color{#35bf28}+7.81\%}$
test_mod_add[compile]	0.1696ms	63.1480μs	15.8358 KOps/s	15.1409 KOps/s	$\color{#35bf28}+4.59\%$
test_mod_add[compile-overhead]	0.1178ms	63.1047μs	15.8467 KOps/s	14.9558 KOps/s	$\textbf{\color{#35bf28}+5.96\%}$
test_mod_wrap[eager]	0.5230ms	0.2191ms	4.5643 KOps/s	4.4254 KOps/s	$\color{#35bf28}+3.14\%$
test_mod_wrap[compile]	2.3553ms	0.2224ms	4.4961 KOps/s	4.3201 KOps/s	$\color{#35bf28}+4.07\%$
test_mod_wrap[compile-overhead]	0.3418ms	0.2203ms	4.5387 KOps/s	4.3844 KOps/s	$\color{#35bf28}+3.52\%$
test_mod_wrap_and_backward[eager]	12.5299ms	11.1865ms	89.3931 Ops/s	71.8616 Ops/s	$\textbf{\color{#35bf28}+24.40\%}$
test_mod_wrap_and_backward[compile]	12.7341ms	11.3546ms	88.0698 Ops/s	82.0002 Ops/s	$\textbf{\color{#35bf28}+7.40\%}$
test_mod_wrap_and_backward[compile-overhead]	17.2612ms	11.6624ms	85.7454 Ops/s	84.8896 Ops/s	$\color{#35bf28}+1.01\%$
test_seq_add[eager]	0.2683ms	0.1154ms	8.6680 KOps/s	8.1916 KOps/s	$\textbf{\color{#35bf28}+5.81\%}$
test_seq_add[compile]	0.1660ms	76.3470μs	13.0981 KOps/s	12.3930 KOps/s	$\textbf{\color{#35bf28}+5.69\%}$
test_seq_add[compile-overhead]	0.1410ms	74.4293μs	13.4356 KOps/s	12.9312 KOps/s	$\color{#35bf28}+3.90\%$
test_seq_wrap[eager]	0.7375ms	0.4441ms	2.2519 KOps/s	2.1537 KOps/s	$\color{#35bf28}+4.56\%$
test_seq_wrap[compile]	0.4603ms	0.2411ms	4.1480 KOps/s	4.0425 KOps/s	$\color{#35bf28}+2.61\%$
test_seq_wrap[compile-overhead]	0.3891ms	0.2385ms	4.1928 KOps/s	4.0719 KOps/s	$\color{#35bf28}+2.97\%$
test_func_call_runtime[False-eager]	0.8944ms	0.5361ms	1.8654 KOps/s	1.8001 KOps/s	$\color{#35bf28}+3.63\%$
test_func_call_runtime[False-compile]	0.8297ms	0.4469ms	2.2374 KOps/s	2.2062 KOps/s	$\color{#35bf28}+1.42\%$
test_func_call_runtime[False-compile-overhead]	0.8603ms	0.4460ms	2.2419 KOps/s	2.2220 KOps/s	$\color{#35bf28}+0.89\%$
test_func_call_runtime[True-eager]	1.0845ms	0.7535ms	1.3271 KOps/s	1.2798 KOps/s	$\color{#35bf28}+3.70\%$
test_func_call_runtime[True-compile]	0.5686ms	0.4646ms	2.1525 KOps/s	2.1334 KOps/s	$\color{#35bf28}+0.89\%$
test_func_call_runtime[True-compile-overhead]	0.5789ms	0.4648ms	2.1515 KOps/s	2.1106 KOps/s	$\color{#35bf28}+1.94\%$
test_func_call_cm_runtime[False-eager]	0.7382ms	0.5367ms	1.8631 KOps/s	1.7927 KOps/s	$\color{#35bf28}+3.93\%$
test_func_call_cm_runtime[False-compile]	0.5585ms	0.4450ms	2.2472 KOps/s	2.2311 KOps/s	$\color{#35bf28}+0.72\%$
test_func_call_cm_runtime[False-compile-overhead]	0.7316ms	0.4419ms	2.2631 KOps/s	2.2099 KOps/s	$\color{#35bf28}+2.41\%$
test_func_call_cm_runtime[True-eager]	1.1854ms	0.9015ms	1.1093 KOps/s	1.0844 KOps/s	$\color{#35bf28}+2.30\%$
test_func_call_cm_runtime[True-compile]	1.2982ms	0.7973ms	1.2542 KOps/s	1.1982 KOps/s	$\color{#35bf28}+4.67\%$
test_func_call_cm_runtime[True-compile-overhead]	1.0346ms	0.7970ms	1.2547 KOps/s	1.1888 KOps/s	$\textbf{\color{#35bf28}+5.55\%}$
test_vmap_func_call_cm_runtime[eager]	2.6929ms	1.9210ms	520.5732 Ops/s	509.9440 Ops/s	$\color{#35bf28}+2.08\%$
test_vmap_func_call_cm_runtime[compile]	0.9170ms	0.5410ms	1.8483 KOps/s	1.7964 KOps/s	$\color{#35bf28}+2.89\%$
test_vmap_func_call_cm_runtime[compile-overhead]	0.9545ms	0.5383ms	1.8579 KOps/s	1.8275 KOps/s	$\color{#35bf28}+1.66\%$
test_distributed	0.2628ms	0.1259ms	7.9451 KOps/s	7.4715 KOps/s	$\textbf{\color{#35bf28}+6.34\%}$
test_tdmodule	0.1331ms	26.6625μs	37.5059 KOps/s	36.4821 KOps/s	$\color{#35bf28}+2.81\%$
test_tdmodule_dispatch	0.1102ms	48.5516μs	20.5966 KOps/s	20.3885 KOps/s	$\color{#35bf28}+1.02\%$
test_tdseq	58.9900μs	28.6339μs	34.9237 KOps/s	33.4719 KOps/s	$\color{#35bf28}+4.34\%$
test_tdseq_dispatch	84.0870μs	53.7526μs	18.6037 KOps/s	17.8154 KOps/s	$\color{#35bf28}+4.42\%$
test_instantiation_functorch	1.7374ms	1.5120ms	661.3945 Ops/s	622.6454 Ops/s	$\textbf{\color{#35bf28}+6.22\%}$
test_exec_functorch	0.3150ms	0.1747ms	5.7249 KOps/s	5.4014 KOps/s	$\textbf{\color{#35bf28}+5.99\%}$
test_exec_functional_call	0.3193ms	0.1680ms	5.9515 KOps/s	5.7028 KOps/s	$\color{#35bf28}+4.36\%$
test_exec_td_decorator	0.5747ms	0.2308ms	4.3331 KOps/s	4.2899 KOps/s	$\color{#35bf28}+1.01\%$
test_vmap_mlp_speed_decorator[True-True]	1.0644ms	0.6556ms	1.5252 KOps/s	1.4995 KOps/s	$\color{#35bf28}+1.72\%$
test_vmap_mlp_speed_decorator[True-False]	1.0254ms	0.6598ms	1.5156 KOps/s	1.5013 KOps/s	$\color{#35bf28}+0.95\%$
test_vmap_mlp_speed_decorator[False-True]	0.8258ms	0.5290ms	1.8902 KOps/s	1.8364 KOps/s	$\color{#35bf28}+2.93\%$
test_vmap_mlp_speed_decorator[False-False]	0.8259ms	0.5287ms	1.8914 KOps/s	1.8541 KOps/s	$\color{#35bf28}+2.01\%$
test_to_module_speed[True]	1.9325ms	1.3341ms	749.5759 Ops/s	749.7244 Ops/s	$\color{#d91a1a}-0.02\%$
test_to_module_speed[False]	1.8051ms	1.3080ms	764.5515 Ops/s	754.5233 Ops/s	$\color{#35bf28}+1.33\%$
test_tc_init	89.5670μs	49.9520μs	20.0192 KOps/s	20.0235 KOps/s	$\color{#d91a1a}-0.02\%$
test_tc_init_nested	0.1652ms	99.2404μs	10.0765 KOps/s	10.2836 KOps/s	$\color{#d91a1a}-2.01\%$
test_tc_first_layer_tensor	39.1230μs	1.5749μs	634.9554 KOps/s	642.6088 KOps/s	$\color{#d91a1a}-1.19\%$
test_tc_first_layer_nontensor	39.1930μs	4.7966μs	208.4790 KOps/s	209.2099 KOps/s	$\color{#d91a1a}-0.35\%$
test_tc_second_layer_tensor	38.3620μs	2.8971μs	345.1728 KOps/s	352.0118 KOps/s	$\color{#d91a1a}-1.94\%$
test_tc_second_layer_nontensor	41.2370μs	6.1773μs	161.8818 KOps/s	166.8481 KOps/s	$\color{#d91a1a}-2.98\%$
test_unbind	0.2543s	14.6081ms	68.4553 Ops/s	57.2371 Ops/s	$\textbf{\color{#35bf28}+19.60\%}$
test_full_like	11.9717ms	8.9160ms	112.1580 Ops/s	118.2353 Ops/s	$\textbf{\color{#d91a1a}-5.14\%}$
test_zeros_like	5.9825ms	4.7072ms	212.4417 Ops/s	299.1298 Ops/s	$\textbf{\color{#d91a1a}-28.98\%}$
test_ones_like	5.0005ms	3.6802ms	271.7208 Ops/s	274.4979 Ops/s	$\color{#d91a1a}-1.01\%$
test_clone	8.0553ms	5.2922ms	188.9564 Ops/s	160.1769 Ops/s	$\textbf{\color{#35bf28}+17.97\%}$
test_squeeze	79.9590μs	12.7776μs	78.2619 KOps/s	75.3991 KOps/s	$\color{#35bf28}+3.80\%$
test_unsqueeze	0.1942ms	91.5747μs	10.9200 KOps/s	10.4787 KOps/s	$\color{#35bf28}+4.21\%$
test_split	0.3877ms	0.1957ms	5.1089 KOps/s	4.9852 KOps/s	$\color{#35bf28}+2.48\%$
test_permute	0.2874ms	0.1980ms	5.0505 KOps/s	4.8651 KOps/s	$\color{#35bf28}+3.81\%$
test_stack	30.9389ms	24.5147ms	40.7918 Ops/s	35.9036 Ops/s	$\textbf{\color{#35bf28}+13.61\%}$
test_cat	28.7327ms	24.1429ms	41.4200 Ops/s	35.9172 Ops/s	$\textbf{\color{#35bf28}+15.32\%}$

[github-actions]

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

Total Benchmarks: 229. Improved: $\large\color{#35bf28}13$. Worsened: $\large\color{#d91a1a}20$.

Expand to view detailed results

Name	Max	Mean	Ops	Ops on Repo HEAD	Change
test_plain_set_nested	36.3100μs	12.8860μs	77.6037 KOps/s	82.8174 KOps/s	$\textbf{\color{#d91a1a}-6.30\%}$
test_plain_set_stack_nested	35.4410μs	12.9705μs	77.0982 KOps/s	81.5332 KOps/s	$\textbf{\color{#d91a1a}-5.44\%}$
test_plain_set_nested_inplace	55.6710μs	13.9693μs	71.5857 KOps/s	75.6612 KOps/s	$\textbf{\color{#d91a1a}-5.39\%}$
test_plain_set_stack_nested_inplace	44.0010μs	13.9160μs	71.8600 KOps/s	75.9952 KOps/s	$\textbf{\color{#d91a1a}-5.44\%}$
test_items	37.1400μs	2.8914μs	345.8547 KOps/s	342.4807 KOps/s	$\color{#35bf28}+0.99\%$
test_items_nested	0.4186ms	0.3602ms	2.7759 KOps/s	2.7097 KOps/s	$\color{#35bf28}+2.44\%$
test_items_nested_locked	0.4128ms	0.3638ms	2.7486 KOps/s	2.7075 KOps/s	$\color{#35bf28}+1.52\%$
test_items_nested_leaf	93.1110μs	60.5760μs	16.5082 KOps/s	16.6295 KOps/s	$\color{#d91a1a}-0.73\%$
test_items_stack_nested	0.4159ms	0.3595ms	2.7814 KOps/s	2.7280 KOps/s	$\color{#35bf28}+1.96\%$
test_items_stack_nested_leaf	83.9510μs	60.1707μs	16.6194 KOps/s	16.0744 KOps/s	$\color{#35bf28}+3.39\%$
test_items_stack_nested_locked	0.4334ms	0.3619ms	2.7634 KOps/s	2.7335 KOps/s	$\color{#35bf28}+1.10\%$
test_keys	56.4010μs	3.4141μs	292.9057 KOps/s	284.8367 KOps/s	$\color{#35bf28}+2.83\%$
test_keys_nested	0.1157ms	88.2680μs	11.3291 KOps/s	11.3727 KOps/s	$\color{#d91a1a}-0.38\%$
test_keys_nested_locked	0.7798ms	93.6367μs	10.6796 KOps/s	10.6806 KOps/s	$-0.01\%$
test_keys_nested_leaf	0.1064ms	79.9539μs	12.5072 KOps/s	12.7092 KOps/s	$\color{#d91a1a}-1.59\%$
test_keys_stack_nested	0.1203ms	88.3432μs	11.3195 KOps/s	11.2660 KOps/s	$\color{#35bf28}+0.47\%$
test_keys_stack_nested_leaf	0.1060ms	79.2688μs	12.6153 KOps/s	12.4553 KOps/s	$\color{#35bf28}+1.28\%$
test_keys_stack_nested_locked	0.1449ms	93.4460μs	10.7014 KOps/s	10.6480 KOps/s	$\color{#35bf28}+0.50\%$
test_values	8.2852μs	0.8781μs	1.1389 MOps/s	1.1576 MOps/s	$\color{#d91a1a}-1.62\%$
test_values_nested	78.3010μs	37.5380μs	26.6397 KOps/s	26.9128 KOps/s	$\color{#d91a1a}-1.02\%$
test_values_nested_locked	93.5010μs	39.6848μs	25.1986 KOps/s	25.5861 KOps/s	$\color{#d91a1a}-1.51\%$
test_values_nested_leaf	95.1410μs	42.3377μs	23.6196 KOps/s	23.6476 KOps/s	$\color{#d91a1a}-0.12\%$
test_values_stack_nested	71.0510μs	37.6990μs	26.5259 KOps/s	26.4600 KOps/s	$\color{#35bf28}+0.25\%$
test_values_stack_nested_leaf	80.3410μs	42.8117μs	23.3581 KOps/s	23.3128 KOps/s	$\color{#35bf28}+0.19\%$
test_values_stack_nested_locked	67.1010μs	39.6480μs	25.2220 KOps/s	24.9768 KOps/s	$\color{#35bf28}+0.98\%$
test_membership	2.2171μs	0.5004μs	1.9984 MOps/s	1.9912 MOps/s	$\color{#35bf28}+0.36\%$
test_membership_nested	39.0500μs	2.0663μs	483.9506 KOps/s	480.6274 KOps/s	$\color{#35bf28}+0.69\%$
test_membership_nested_leaf	22.8250μs	2.0039μs	499.0237 KOps/s	498.2671 KOps/s	$\color{#35bf28}+0.15\%$
test_membership_stacked_nested	39.5300μs	2.0766μs	481.5566 KOps/s	479.2675 KOps/s	$\color{#35bf28}+0.48\%$
test_membership_stacked_nested_leaf	38.6200μs	2.0704μs	482.9931 KOps/s	478.8443 KOps/s	$\color{#35bf28}+0.87\%$
test_membership_nested_last	25.3300μs	3.0503μs	327.8364 KOps/s	325.2656 KOps/s	$\color{#35bf28}+0.79\%$
test_membership_nested_leaf_last	63.7010μs	3.0320μs	329.8182 KOps/s	324.8731 KOps/s	$\color{#35bf28}+1.52\%$
test_membership_stacked_nested_last	24.3200μs	3.0424μs	328.6844 KOps/s	266.6523 KOps/s	$\textbf{\color{#35bf28}+23.26\%}$
test_membership_stacked_nested_leaf_last	28.6110μs	3.0581μs	326.9969 KOps/s	266.1170 KOps/s	$\textbf{\color{#35bf28}+22.88\%}$
test_nested_getleaf	40.9600μs	6.2988μs	158.7601 KOps/s	159.7387 KOps/s	$\color{#d91a1a}-0.61\%$
test_nested_get	95.9210μs	5.9118μs	169.1521 KOps/s	167.5349 KOps/s	$\color{#35bf28}+0.97\%$
test_stacked_getleaf	87.2210μs	6.2267μs	160.5985 KOps/s	162.0254 KOps/s	$\color{#d91a1a}-0.88\%$
test_stacked_get	97.8510μs	5.7864μs	172.8198 KOps/s	171.8377 KOps/s	$\color{#35bf28}+0.57\%$
test_nested_getitemleaf	38.4700μs	6.4273μs	155.5873 KOps/s	156.1968 KOps/s	$\color{#d91a1a}-0.39\%$
test_nested_getitem	34.5910μs	6.1395μs	162.8792 KOps/s	163.5442 KOps/s	$\color{#d91a1a}-0.41\%$
test_stacked_getitemleaf	38.1210μs	6.4133μs	155.9249 KOps/s	156.2090 KOps/s	$\color{#d91a1a}-0.18\%$
test_stacked_getitem	33.6200μs	6.0141μs	166.2748 KOps/s	165.8428 KOps/s	$\color{#35bf28}+0.26\%$
test_lock_nested	0.4044ms	0.3412ms	2.9309 KOps/s	2.8585 KOps/s	$\color{#35bf28}+2.53\%$
test_lock_stack_nested	0.3932ms	0.3471ms	2.8808 KOps/s	2.8937 KOps/s	$\color{#d91a1a}-0.44\%$
test_unlock_nested	0.3601ms	0.2875ms	3.4786 KOps/s	3.5064 KOps/s	$\color{#d91a1a}-0.79\%$
test_unlock_stack_nested	0.3249ms	0.2849ms	3.5097 KOps/s	3.5195 KOps/s	$\color{#d91a1a}-0.28\%$
test_flatten_speed	0.1146ms	77.2441μs	12.9460 KOps/s	13.1246 KOps/s	$\color{#d91a1a}-1.36\%$
test_unflatten_speed	0.3762ms	0.3193ms	3.1317 KOps/s	3.1324 KOps/s	$\color{#d91a1a}-0.02\%$
test_common_ops	0.7522ms	0.6219ms	1.6080 KOps/s	1.6464 KOps/s	$\color{#d91a1a}-2.34\%$
test_creation	80.0900μs	1.7500μs	571.4268 KOps/s	566.3684 KOps/s	$\color{#35bf28}+0.89\%$
test_creation_empty	46.8400μs	9.0115μs	110.9690 KOps/s	130.3780 KOps/s	$\textbf{\color{#d91a1a}-14.89\%}$
test_creation_nested_1	34.8200μs	10.7884μs	92.6923 KOps/s	106.6787 KOps/s	$\textbf{\color{#d91a1a}-13.11\%}$
test_creation_nested_2	44.6100μs	13.3636μs	74.8300 KOps/s	83.5185 KOps/s	$\textbf{\color{#d91a1a}-10.40\%}$
test_clone	60.9010μs	11.2269μs	89.0715 KOps/s	87.8522 KOps/s	$\color{#35bf28}+1.39\%$
test_getitem[int]	1.2118ms	10.7657μs	92.8877 KOps/s	92.8982 KOps/s	$\color{#d91a1a}-0.01\%$
test_getitem[slice_int]	0.1074ms	20.9363μs	47.7640 KOps/s	47.2233 KOps/s	$\color{#35bf28}+1.14\%$
test_getitem[range]	0.1358ms	37.5655μs	26.6202 KOps/s	26.0949 KOps/s	$\color{#35bf28}+2.01\%$
test_getitem[tuple]	95.9610μs	18.5196μs	53.9968 KOps/s	54.8849 KOps/s	$\color{#d91a1a}-1.62\%$
test_getitem[list]	0.1459ms	32.8356μs	30.4547 KOps/s	29.6736 KOps/s	$\color{#35bf28}+2.63\%$
test_setitem_dim[int]	42.7110μs	19.9595μs	50.1014 KOps/s	49.7415 KOps/s	$\color{#35bf28}+0.72\%$
test_setitem_dim[slice_int]	59.9500μs	38.3450μs	26.0790 KOps/s	24.5534 KOps/s	$\textbf{\color{#35bf28}+6.21\%}$
test_setitem_dim[range]	75.8910μs	53.3625μs	18.7397 KOps/s	19.1711 KOps/s	$\color{#d91a1a}-2.25\%$
test_setitem_dim[tuple]	54.2110μs	33.1615μs	30.1555 KOps/s	30.6064 KOps/s	$\color{#d91a1a}-1.47\%$
test_setitem	66.7710μs	15.9779μs	62.5863 KOps/s	63.2529 KOps/s	$\color{#d91a1a}-1.05\%$
test_set	74.3610μs	15.3980μs	64.9437 KOps/s	66.1499 KOps/s	$\color{#d91a1a}-1.82\%$
test_set_shared	0.5254ms	0.1582ms	6.3207 KOps/s	6.2948 KOps/s	$\color{#35bf28}+0.41\%$
test_update	0.4252ms	18.7792μs	53.2503 KOps/s	56.2339 KOps/s	$\textbf{\color{#d91a1a}-5.31\%}$
test_update_nested	68.1710μs	25.0931μs	39.8515 KOps/s	43.0003 KOps/s	$\textbf{\color{#d91a1a}-7.32\%}$
test_update__nested	0.5558ms	25.2867μs	39.5465 KOps/s	39.3165 KOps/s	$\color{#35bf28}+0.59\%$
test_set_nested	59.4110μs	16.9692μs	58.9304 KOps/s	61.5391 KOps/s	$\color{#d91a1a}-4.24\%$
test_set_nested_new	62.2910μs	19.6262μs	50.9524 KOps/s	52.7833 KOps/s	$\color{#d91a1a}-3.47\%$
test_select	75.8010μs	30.7325μs	32.5389 KOps/s	33.6467 KOps/s	$\color{#d91a1a}-3.29\%$
test_select_nested	74.5110μs	43.4738μs	23.0023 KOps/s	22.5997 KOps/s	$\color{#35bf28}+1.78\%$
test_exclude_nested	0.1253ms	62.3253μs	16.0448 KOps/s	15.6497 KOps/s	$\color{#35bf28}+2.53\%$
test_empty[True]	0.4028ms	0.2917ms	3.4284 KOps/s	3.3903 KOps/s	$\color{#35bf28}+1.13\%$
test_empty[False]	4.9731μs	0.8214μs	1.2175 MOps/s	1.2201 MOps/s	$\color{#d91a1a}-0.21\%$
test_to	88.1410μs	58.5964μs	17.0659 KOps/s	17.5749 KOps/s	$\color{#d91a1a}-2.90\%$
test_to_nonblocking	0.1010ms	47.7186μs	20.9562 KOps/s	19.9552 KOps/s	$\textbf{\color{#35bf28}+5.02\%}$
test_unbind_speed	0.2763ms	0.2469ms	4.0500 KOps/s	4.0537 KOps/s	$\color{#d91a1a}-0.09\%$
test_unbind_speed_stack0	0.2922ms	0.2417ms	4.1381 KOps/s	4.1352 KOps/s	$\color{#35bf28}+0.07\%$
test_unbind_speed_stack1	92.8770ms	0.7429ms	1.3461 KOps/s	1.3623 KOps/s	$\color{#d91a1a}-1.19\%$
test_split	95.2124ms	1.5946ms	627.1206 Ops/s	681.2448 Ops/s	$\textbf{\color{#d91a1a}-7.94\%}$
test_chunk	95.0993ms	1.5916ms	628.3064 Ops/s	568.7204 Ops/s	$\textbf{\color{#35bf28}+10.48\%}$
test_consolidate[False-None]	2.7863ms	2.6084ms	383.3745 Ops/s	375.7593 Ops/s	$\color{#35bf28}+2.03\%$
test_consolidate[default-None]	1.8285ms	1.7139ms	583.4689 Ops/s	597.6450 Ops/s	$\color{#d91a1a}-2.37\%$
test_consolidate[reduce-overhead-None]	1.9265ms	1.7327ms	577.1238 Ops/s	575.1144 Ops/s	$\color{#35bf28}+0.35\%$
test_consolidate_njt[False-None]	6.6733ms	6.3865ms	156.5802 Ops/s	155.4058 Ops/s	$\color{#35bf28}+0.76\%$
test_to[False-False-None]	1.8676ms	1.7427ms	573.8085 Ops/s	576.3117 Ops/s	$\color{#d91a1a}-0.43\%$
test_to[True-False-None]	1.5858ms	1.3522ms	739.5446 Ops/s	752.9897 Ops/s	$\color{#d91a1a}-1.79\%$
test_to[within-False-None]	4.2275ms	4.1110ms	243.2503 Ops/s	244.1300 Ops/s	$\color{#d91a1a}-0.36\%$
test_to[True-default-None]	5.3661ms	5.1771ms	193.1587 Ops/s	191.1560 Ops/s	$\color{#35bf28}+1.05\%$
test_to_njt[False-False-None]	7.0257ms	6.9120ms	144.6754 Ops/s	143.6874 Ops/s	$\color{#35bf28}+0.69\%$
test_to_njt[True-False-None]	5.6471ms	5.4421ms	183.7525 Ops/s	180.8922 Ops/s	$\color{#35bf28}+1.58\%$
test_to_njt[within-False-None]	12.0926ms	11.8740ms	84.2174 Ops/s	82.0161 Ops/s	$\color{#35bf28}+2.68\%$
test_creation[device0]	0.6339ms	79.6813μs	12.5500 KOps/s	12.5311 KOps/s	$\color{#35bf28}+0.15\%$
test_creation_from_tensor	0.6370ms	83.5785μs	11.9648 KOps/s	12.0499 KOps/s	$\color{#d91a1a}-0.71\%$
test_add_one[memmap_tensor0]	0.6326ms	7.0792μs	141.2588 KOps/s	143.1263 KOps/s	$\color{#d91a1a}-1.30\%$
test_contiguous[memmap_tensor0]	1.8505μs	0.4123μs	2.4252 MOps/s	2.4326 MOps/s	$\color{#d91a1a}-0.30\%$
test_stack[memmap_tensor0]	23.9610μs	4.4899μs	222.7223 KOps/s	218.6109 KOps/s	$\color{#35bf28}+1.88\%$
test_memmaptd_index	1.6425ms	0.2417ms	4.1377 KOps/s	4.0618 KOps/s	$\color{#35bf28}+1.87\%$
test_memmaptd_index_astensor	0.4155ms	0.3038ms	3.2918 KOps/s	3.2875 KOps/s	$\color{#35bf28}+0.13\%$
test_memmaptd_index_op	0.7120ms	0.5879ms	1.7009 KOps/s	1.7358 KOps/s	$\color{#d91a1a}-2.01\%$
test_serialize_model	0.1313s	0.1300s	7.6939 Ops/s	7.6926 Ops/s	$\color{#35bf28}+0.02\%$
test_serialize_model_pickle	1.3471s	1.1854s	0.8436 Ops/s	0.8218 Ops/s	$\color{#35bf28}+2.66\%$
test_serialize_weights	0.1308s	0.1292s	7.7386 Ops/s	7.7161 Ops/s	$\color{#35bf28}+0.29\%$
test_serialize_weights_returnearly	0.3142s	53.3916ms	18.7295 Ops/s	23.3814 Ops/s	$\textbf{\color{#d91a1a}-19.90\%}$
test_serialize_weights_pickle	1.3721s	1.2166s	0.8220 Ops/s	0.8189 Ops/s	$\color{#35bf28}+0.38\%$
test_reshape_pytree	47.0400μs	21.5477μs	46.4088 KOps/s	44.7443 KOps/s	$\color{#35bf28}+3.72\%$
test_reshape_td	0.4239ms	25.7052μs	38.9026 KOps/s	37.3024 KOps/s	$\color{#35bf28}+4.29\%$
test_view_pytree	51.1200μs	21.3602μs	46.8160 KOps/s	46.4231 KOps/s	$\color{#35bf28}+0.85\%$
test_view_td	0.4272ms	29.7893μs	33.5691 KOps/s	31.1141 KOps/s	$\textbf{\color{#35bf28}+7.89\%}$
test_unbind_pytree	64.4100μs	28.2474μs	35.4015 KOps/s	34.3948 KOps/s	$\color{#35bf28}+2.93\%$
test_unbind_td	0.5904ms	36.8639μs	27.1268 KOps/s	25.3481 KOps/s	$\textbf{\color{#35bf28}+7.02\%}$
test_split_pytree	0.4367ms	28.9816μs	34.5047 KOps/s	33.9092 KOps/s	$\color{#35bf28}+1.76\%$
test_split_td	0.7511ms	37.6045μs	26.5926 KOps/s	25.9349 KOps/s	$\color{#35bf28}+2.54\%$
test_add_pytree	0.4337ms	34.4531μs	29.0250 KOps/s	28.0642 KOps/s	$\color{#35bf28}+3.42\%$
test_add_td	84.5210μs	48.5206μs	20.6098 KOps/s	21.1898 KOps/s	$\color{#d91a1a}-2.74\%$
test_compile_add_one_nested[tensordict-compile]	0.1793ms	0.1222ms	8.1862 KOps/s	7.8179 KOps/s	$\color{#35bf28}+4.71\%$
test_compile_add_one_nested[tensordict-eager]	0.2230ms	0.1309ms	7.6379 KOps/s	7.4276 KOps/s	$\color{#35bf28}+2.83\%$
test_compile_add_one_nested[pytree-compile]	0.1469ms	98.6030μs	10.1417 KOps/s	10.3071 KOps/s	$\color{#d91a1a}-1.60\%$
test_compile_add_one_nested[pytree-eager]	1.2939ms	0.1528ms	6.5462 KOps/s	6.4389 KOps/s	$\color{#35bf28}+1.67\%$
test_compile_copy_nested[tensordict-compile]	56.5010μs	22.6837μs	44.0845 KOps/s	43.9189 KOps/s	$\color{#35bf28}+0.38\%$
test_compile_copy_nested[tensordict-eager]	84.6200μs	29.5724μs	33.8153 KOps/s	34.2029 KOps/s	$\color{#d91a1a}-1.13\%$
test_compile_copy_nested[pytree-compile]	0.4510ms	64.6464μs	15.4688 KOps/s	15.5123 KOps/s	$\color{#d91a1a}-0.28\%$
test_compile_copy_nested[pytree-eager]	84.6410μs	49.0258μs	20.3974 KOps/s	20.5527 KOps/s	$\color{#d91a1a}-0.76\%$
test_compile_add_one_flat[tensordict-compile]	0.2135ms	0.1443ms	6.9291 KOps/s	7.0638 KOps/s	$\color{#d91a1a}-1.91\%$
test_compile_add_one_flat[tensordict-eager]	0.3123ms	0.2198ms	4.5506 KOps/s	4.6900 KOps/s	$\color{#d91a1a}-2.97\%$
test_compile_add_one_flat[tensorclass-compile]	0.1501ms	99.3104μs	10.0694 KOps/s	9.7849 KOps/s	$\color{#35bf28}+2.91\%$
test_compile_add_one_flat[tensorclass-eager]	0.4201ms	55.0243μs	18.1738 KOps/s	17.1592 KOps/s	$\textbf{\color{#35bf28}+5.91\%}$
test_compile_add_one_flat[pytree-compile]	0.2295ms	0.1387ms	7.2088 KOps/s	7.0194 KOps/s	$\color{#35bf28}+2.70\%$
test_compile_add_one_flat[pytree-eager]	0.5911ms	0.4971ms	2.0115 KOps/s	1.9777 KOps/s	$\color{#35bf28}+1.71\%$
test_compile_add_self_flat[tensordict-eager]	0.3958ms	0.2588ms	3.8642 KOps/s	3.7682 KOps/s	$\color{#35bf28}+2.55\%$
test_compile_add_self_flat[tensordict-compile]	0.2160ms	0.1468ms	6.8123 KOps/s	6.7637 KOps/s	$\color{#35bf28}+0.72\%$
test_compile_add_self_flat[tensorclass-eager]	0.2974ms	67.5300μs	14.8082 KOps/s	14.1995 KOps/s	$\color{#35bf28}+4.29\%$
test_compile_add_self_flat[tensorclass-compile]	0.4461ms	0.1012ms	9.8856 KOps/s	9.6526 KOps/s	$\color{#35bf28}+2.41\%$
test_compile_add_self_flat[pytree-eager]	0.5174ms	0.4137ms	2.4174 KOps/s	2.3823 KOps/s	$\color{#35bf28}+1.47\%$
test_compile_add_self_flat[pytree-compile]	0.1825ms	0.1344ms	7.4397 KOps/s	7.4264 KOps/s	$\color{#35bf28}+0.18\%$
test_compile_copy_flat[tensordict-compile]	52.5100μs	17.6642μs	56.6117 KOps/s	56.5996 KOps/s	$\color{#35bf28}+0.02\%$
test_compile_copy_flat[tensordict-eager]	66.9210μs	31.6535μs	31.5921 KOps/s	31.9560 KOps/s	$\color{#d91a1a}-1.14\%$
test_compile_copy_flat[pytree-compile]	0.1038ms	70.0766μs	14.2701 KOps/s	14.3392 KOps/s	$\color{#d91a1a}-0.48\%$
test_compile_copy_flat[pytree-eager]	0.1560ms	51.9109μs	19.2638 KOps/s	18.9442 KOps/s	$\color{#35bf28}+1.69\%$
test_compile_assign_and_add[tensordict-compile]	1.6275ms	0.3950ms	2.5316 KOps/s	2.2220 KOps/s	$\textbf{\color{#35bf28}+13.94\%}$
test_compile_assign_and_add[tensordict-eager]	2.8494ms	2.6573ms	376.3260 Ops/s	363.7004 Ops/s	$\color{#35bf28}+3.47\%$
test_compile_assign_and_add[pytree-compile]	1.5971ms	0.3835ms	2.6072 KOps/s	2.2800 KOps/s	$\textbf{\color{#35bf28}+14.35\%}$
test_compile_assign_and_add[pytree-eager]	2.8668ms	2.7117ms	368.7784 Ops/s	368.7640 Ops/s	$+0.00\%$
test_compile_indexing[tensor-tensordict-compile]	0.5988ms	0.1194ms	8.3721 KOps/s	8.7841 KOps/s	$\color{#d91a1a}-4.69\%$
test_compile_indexing[tensor-tensordict-eager]	0.5804ms	84.9563μs	11.7708 KOps/s	12.3793 KOps/s	$\color{#d91a1a}-4.92\%$
test_compile_indexing[tensor-tensorclass-compile]	0.6200ms	0.1131ms	8.8452 KOps/s	9.4695 KOps/s	$\textbf{\color{#d91a1a}-6.59\%}$
test_compile_indexing[tensor-tensorclass-eager]	0.1136ms	72.6234μs	13.7697 KOps/s	13.8870 KOps/s	$\color{#d91a1a}-0.84\%$
test_compile_indexing[tensor-pytree-compile]	0.1677ms	0.1134ms	8.8185 KOps/s	8.9892 KOps/s	$\color{#d91a1a}-1.90\%$
test_compile_indexing[tensor-pytree-eager]	0.1260ms	73.2876μs	13.6449 KOps/s	13.8387 KOps/s	$\color{#d91a1a}-1.40\%$
test_compile_indexing[slice-tensordict-compile]	0.2390ms	0.1042ms	9.5951 KOps/s	9.9116 KOps/s	$\color{#d91a1a}-3.19\%$
test_compile_indexing[slice-tensordict-eager]	0.1407ms	17.0375μs	58.6940 KOps/s	56.4360 KOps/s	$\color{#35bf28}+4.00\%$
test_compile_indexing[slice-tensorclass-compile]	0.1475ms	98.6112μs	10.1408 KOps/s	10.2511 KOps/s	$\color{#d91a1a}-1.08\%$
test_compile_indexing[slice-tensorclass-eager]	53.5310μs	15.9479μs	62.7043 KOps/s	62.0060 KOps/s	$\color{#35bf28}+1.13\%$
test_compile_indexing[slice-pytree-compile]	0.1562ms	0.1015ms	9.8510 KOps/s	10.3322 KOps/s	$\color{#d91a1a}-4.66\%$
test_compile_indexing[slice-pytree-eager]	51.6610μs	15.8579μs	63.0601 KOps/s	62.5806 KOps/s	$\color{#35bf28}+0.77\%$
test_compile_indexing[int-tensordict-compile]	0.1507ms	0.1056ms	9.4654 KOps/s	9.8382 KOps/s	$\color{#d91a1a}-3.79\%$
test_compile_indexing[int-tensordict-eager]	0.5886ms	17.1237μs	58.3985 KOps/s	57.3588 KOps/s	$\color{#35bf28}+1.81\%$
test_compile_indexing[int-tensorclass-compile]	0.1860ms	97.7643μs	10.2287 KOps/s	10.3260 KOps/s	$\color{#d91a1a}-0.94\%$
test_compile_indexing[int-tensorclass-eager]	50.3400μs	15.7340μs	63.5567 KOps/s	62.4231 KOps/s	$\color{#35bf28}+1.82\%$
test_compile_indexing[int-pytree-compile]	0.1506ms	0.1003ms	9.9743 KOps/s	10.1408 KOps/s	$\color{#d91a1a}-1.64\%$
test_compile_indexing[int-pytree-eager]	0.4091ms	15.8521μs	63.0830 KOps/s	62.5171 KOps/s	$\color{#35bf28}+0.91\%$
test_mod_add[eager]	0.1819ms	40.7273μs	24.5536 KOps/s	26.5244 KOps/s	$\textbf{\color{#d91a1a}-7.43\%}$
test_mod_add[compile]	0.3147ms	80.6883μs	12.3934 KOps/s	12.3370 KOps/s	$\color{#35bf28}+0.46\%$
test_mod_add[compile-overhead]	0.3210ms	0.1666ms	6.0026 KOps/s	5.4186 KOps/s	$\textbf{\color{#35bf28}+10.78\%}$
test_mod_wrap[eager]	0.4039ms	0.2573ms	3.8867 KOps/s	3.7502 KOps/s	$\color{#35bf28}+3.64\%$
test_mod_wrap[compile]	0.5414ms	0.2841ms	3.5195 KOps/s	3.3950 KOps/s	$\color{#35bf28}+3.67\%$
test_mod_wrap[compile-overhead]	7.3763ms	3.8742ms	258.1197 Ops/s	271.2869 Ops/s	$\color{#d91a1a}-4.85\%$
test_mod_wrap_and_backward[eager]	1.7917ms	1.4782ms	676.4861 Ops/s	683.5139 Ops/s	$\color{#d91a1a}-1.03\%$
test_mod_wrap_and_backward[compile]	1.4955ms	1.3867ms	721.1189 Ops/s	721.7766 Ops/s	$\color{#d91a1a}-0.09\%$
test_mod_wrap_and_backward[compile-overhead]	1.4378ms	1.0041ms	995.9027 Ops/s	949.9888 Ops/s	$\color{#35bf28}+4.83\%$
test_seq_add[eager]	0.1857ms	0.1167ms	8.5715 KOps/s	8.7131 KOps/s	$\color{#d91a1a}-1.63\%$
test_seq_add[compile]	0.1384ms	87.5944μs	11.4163 KOps/s	11.4924 KOps/s	$\color{#d91a1a}-0.66\%$
test_seq_add[compile-overhead]	0.1922ms	0.1295ms	7.7199 KOps/s	7.8179 KOps/s	$\color{#d91a1a}-1.25\%$
test_seq_wrap[eager]	0.5411ms	0.4245ms	2.3559 KOps/s	2.2907 KOps/s	$\color{#35bf28}+2.85\%$
test_seq_wrap[compile]	0.4429ms	0.2995ms	3.3389 KOps/s	3.2681 KOps/s	$\color{#35bf28}+2.16\%$
test_seq_wrap[compile-overhead]	0.3299ms	0.2232ms	4.4807 KOps/s	4.3611 KOps/s	$\color{#35bf28}+2.74\%$
test_func_call_runtime[False-eager]	0.8639ms	0.7360ms	1.3588 KOps/s	1.2709 KOps/s	$\textbf{\color{#35bf28}+6.91\%}$
test_func_call_runtime[False-compile]	1.1260ms	0.7371ms	1.3567 KOps/s	1.3273 KOps/s	$\color{#35bf28}+2.21\%$
test_func_call_runtime[False-compile-overhead]	0.4349ms	0.3620ms	2.7625 KOps/s	2.7211 KOps/s	$\color{#35bf28}+1.52\%$
test_func_call_runtime[True-eager]	1.3451ms	0.9094ms	1.0996 KOps/s	1.0574 KOps/s	$\color{#35bf28}+3.98\%$
test_func_call_runtime[True-compile]	1.1697ms	0.7601ms	1.3157 KOps/s	1.2968 KOps/s	$\color{#35bf28}+1.46\%$
test_func_call_runtime[True-compile-overhead]	0.8081ms	0.3835ms	2.6078 KOps/s	2.5663 KOps/s	$\color{#35bf28}+1.61\%$
test_func_call_cm_runtime[False-eager]	1.1722ms	0.7331ms	1.3641 KOps/s	1.2532 KOps/s	$\textbf{\color{#35bf28}+8.86\%}$
test_func_call_cm_runtime[False-compile]	1.1673ms	0.7342ms	1.3619 KOps/s	1.3222 KOps/s	$\color{#35bf28}+3.00\%$
test_func_call_cm_runtime[False-compile-overhead]	0.4664ms	0.3602ms	2.7759 KOps/s	2.6968 KOps/s	$\color{#35bf28}+2.93\%$
test_func_call_cm_runtime[True-eager]	1.4409ms	0.9942ms	1.0058 KOps/s	966.4844 Ops/s	$\color{#35bf28}+4.07\%$
test_func_call_cm_runtime[True-compile]	1.4285ms	0.9908ms	1.0093 KOps/s	976.2432 Ops/s	$\color{#35bf28}+3.39\%$
test_func_call_cm_runtime[True-compile-overhead]	1.4127ms	0.9908ms	1.0093 KOps/s	970.7679 Ops/s	$\color{#35bf28}+3.97\%$
test_vmap_func_call_cm_runtime[eager]	2.4870ms	2.0859ms	479.4032 Ops/s	463.5644 Ops/s	$\color{#35bf28}+3.42\%$
test_vmap_func_call_cm_runtime[compile]	0.9575ms	0.8034ms	1.2447 KOps/s	1.2365 KOps/s	$\color{#35bf28}+0.66\%$
test_vmap_func_call_cm_runtime[compile-overhead]	0.4890ms	0.4158ms	2.4048 KOps/s	2.3425 KOps/s	$\color{#35bf28}+2.66\%$
test_distributed	5.0659ms	0.2542ms	3.9340 KOps/s	8.7451 KOps/s	$\textbf{\color{#d91a1a}-55.01\%}$
test_tdmodule	0.4315ms	21.4298μs	46.6640 KOps/s	52.8064 KOps/s	$\textbf{\color{#d91a1a}-11.63\%}$
test_tdmodule_dispatch	71.6510μs	36.4022μs	27.4709 KOps/s	29.9220 KOps/s	$\textbf{\color{#d91a1a}-8.19\%}$
test_tdseq	44.5600μs	20.6232μs	48.4890 KOps/s	50.9007 KOps/s	$\color{#d91a1a}-4.74\%$
test_tdseq_dispatch	58.4510μs	38.5900μs	25.9134 KOps/s	26.9111 KOps/s	$\color{#d91a1a}-3.71\%$
test_instantiation_functorch	1.7350ms	1.5358ms	651.1439 Ops/s	648.9585 Ops/s	$\color{#35bf28}+0.34\%$
test_exec_functorch	0.1909ms	0.1434ms	6.9733 KOps/s	6.9880 KOps/s	$\color{#d91a1a}-0.21\%$
test_exec_functional_call	0.2079ms	0.1407ms	7.1056 KOps/s	7.3331 KOps/s	$\color{#d91a1a}-3.10\%$
test_exec_td_decorator	0.3728ms	0.1890ms	5.2902 KOps/s	5.3395 KOps/s	$\color{#d91a1a}-0.92\%$
test_vmap_mlp_speed_decorator[True-True]	0.8416ms	0.7017ms	1.4251 KOps/s	1.4646 KOps/s	$\color{#d91a1a}-2.69\%$
test_vmap_mlp_speed_decorator[True-False]	1.0940ms	0.6887ms	1.4520 KOps/s	1.4583 KOps/s	$\color{#d91a1a}-0.43\%$
test_vmap_mlp_speed_decorator[False-True]	0.9875ms	0.5925ms	1.6877 KOps/s	1.6639 KOps/s	$\color{#35bf28}+1.43\%$
test_vmap_mlp_speed_decorator[False-False]	0.9851ms	0.5946ms	1.6818 KOps/s	1.6701 KOps/s	$\color{#35bf28}+0.70\%$
test_vmap_transformer_speed_decorator[True-True]	20.2123ms	19.2853ms	51.8530 Ops/s	51.5752 Ops/s	$\color{#35bf28}+0.54\%$
test_vmap_transformer_speed_decorator[True-False]	20.2558ms	19.5621ms	51.1194 Ops/s	51.7029 Ops/s	$\color{#d91a1a}-1.13\%$
test_vmap_transformer_speed_decorator[False-True]	20.1859ms	19.8063ms	50.4891 Ops/s	52.3495 Ops/s	$\color{#d91a1a}-3.55\%$
test_vmap_transformer_speed_decorator[False-False]	20.0111ms	19.1527ms	52.2119 Ops/s	51.8182 Ops/s	$\color{#35bf28}+0.76\%$
test_to_module_speed[True]	1.5058ms	0.9540ms	1.0482 KOps/s	1.0490 KOps/s	$\color{#d91a1a}-0.08\%$
test_to_module_speed[False]	1.0343ms	0.9476ms	1.0552 KOps/s	1.0763 KOps/s	$\color{#d91a1a}-1.95\%$
test_tc_init	89.4010μs	36.4786μs	27.4133 KOps/s	30.4610 KOps/s	$\textbf{\color{#d91a1a}-10.01\%}$
test_tc_init_nested	0.1150ms	75.3200μs	13.2767 KOps/s	14.9419 KOps/s	$\textbf{\color{#d91a1a}-11.14\%}$
test_tc_first_layer_tensor	24.1500μs	0.8089μs	1.2363 MOps/s	1.2506 MOps/s	$\color{#d91a1a}-1.15\%$
test_tc_first_layer_nontensor	36.4810μs	2.2002μs	454.5026 KOps/s	446.9891 KOps/s	$\color{#35bf28}+1.68\%$
test_tc_second_layer_tensor	95.9810μs	1.3837μs	722.7001 KOps/s	718.6849 KOps/s	$\color{#35bf28}+0.56\%$
test_tc_second_layer_nontensor	22.4200μs	2.9296μs	341.3420 KOps/s	344.8639 KOps/s	$\color{#d91a1a}-1.02\%$
test_unbind	0.2168s	11.9258ms	83.8515 Ops/s	141.4074 Ops/s	$\textbf{\color{#d91a1a}-40.70\%}$
test_full_like	9.7789ms	9.3376ms	107.0939 Ops/s	105.0931 Ops/s	$\color{#35bf28}+1.90\%$
test_zeros_like	5.0359ms	4.3472ms	230.0351 Ops/s	230.4989 Ops/s	$\color{#d91a1a}-0.20\%$
test_ones_like	9.0880ms	4.4621ms	224.1111 Ops/s	231.1717 Ops/s	$\color{#d91a1a}-3.05\%$
test_clone	7.0872ms	6.6707ms	149.9091 Ops/s	155.6649 Ops/s	$\color{#d91a1a}-3.70\%$
test_squeeze	59.0510μs	9.4195μs	106.1629 KOps/s	105.9108 KOps/s	$\color{#35bf28}+0.24\%$
test_unsqueeze	0.1276ms	74.4458μs	13.4326 KOps/s	14.1725 KOps/s	$\textbf{\color{#d91a1a}-5.22\%}$
test_split	0.5718ms	0.1589ms	6.2921 KOps/s	6.3617 KOps/s	$\color{#d91a1a}-1.09\%$
test_permute	0.2501ms	0.1858ms	5.3826 KOps/s	5.6132 KOps/s	$\color{#d91a1a}-4.11\%$
test_stack	51.4680ms	50.7329ms	19.7111 Ops/s	19.7586 Ops/s	$\color{#d91a1a}-0.24\%$
test_cat	50.9789ms	50.6237ms	19.7536 Ops/s	20.1009 Ops/s	$\color{#d91a1a}-1.73\%$