[BugFix] Use separate streams for cudagraph warmup

[vmoens]

Stack from ghstack (oldest at bottom):

• -> [BugFix] Use separate streams for cudagraph warmup #1010

[github-actions]

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

Total Benchmarks: 222. Improved: $\large\color{#35bf28}46$. Worsened: $\large\color{#d91a1a}39$.

Expand to view detailed results

Name	Max	Mean	Ops	Ops on Repo HEAD	Change
test_plain_set_nested	44.3130μs	21.4363μs	46.6499 KOps/s	50.5395 KOps/s	$\textbf{\color{#d91a1a}-7.70\%}$
test_plain_set_stack_nested	52.0980μs	22.1748μs	45.0963 KOps/s	50.3384 KOps/s	$\textbf{\color{#d91a1a}-10.41\%}$
test_plain_set_nested_inplace	73.7380μs	23.7829μs	42.0470 KOps/s	47.1037 KOps/s	$\textbf{\color{#d91a1a}-10.74\%}$
test_plain_set_stack_nested_inplace	57.2770μs	23.7994μs	42.0179 KOps/s	46.7057 KOps/s	$\textbf{\color{#d91a1a}-10.04\%}$
test_items	22.8320μs	4.2248μs	236.7003 KOps/s	247.9907 KOps/s	$\color{#d91a1a}-4.55\%$
test_items_nested	0.4747ms	0.3636ms	2.7499 KOps/s	2.7517 KOps/s	$\color{#d91a1a}-0.06\%$
test_items_nested_locked	0.9009ms	0.3675ms	2.7212 KOps/s	2.7362 KOps/s	$\color{#d91a1a}-0.55\%$
test_items_nested_leaf	0.1256ms	69.7525μs	14.3364 KOps/s	14.7878 KOps/s	$\color{#d91a1a}-3.05\%$
test_items_stack_nested	0.4640ms	0.3707ms	2.6979 KOps/s	2.7421 KOps/s	$\color{#d91a1a}-1.61\%$
test_items_stack_nested_leaf	0.1568ms	71.6523μs	13.9563 KOps/s	13.5695 KOps/s	$\color{#35bf28}+2.85\%$
test_items_stack_nested_locked	0.9797ms	0.4781ms	2.0918 KOps/s	2.6973 KOps/s	$\textbf{\color{#d91a1a}-22.45\%}$
test_keys	42.3600μs	4.0253μs	248.4290 KOps/s	278.8630 KOps/s	$\textbf{\color{#d91a1a}-10.91\%}$
test_keys_nested	0.2991ms	0.1120ms	8.9306 KOps/s	9.9469 KOps/s	$\textbf{\color{#d91a1a}-10.22\%}$
test_keys_nested_locked	1.6327ms	0.1177ms	8.4995 KOps/s	9.4559 KOps/s	$\textbf{\color{#d91a1a}-10.11\%}$
test_keys_nested_leaf	0.2551ms	94.7757μs	10.5512 KOps/s	12.0032 KOps/s	$\textbf{\color{#d91a1a}-12.10\%}$
test_keys_stack_nested	0.1822ms	0.1010ms	9.9017 KOps/s	9.8418 KOps/s	$\color{#35bf28}+0.61\%$
test_keys_stack_nested_leaf	0.1688ms	83.4255μs	11.9867 KOps/s	11.8356 KOps/s	$\color{#35bf28}+1.28\%$
test_keys_stack_nested_locked	0.1962ms	0.1052ms	9.5042 KOps/s	9.4387 KOps/s	$\color{#35bf28}+0.69\%$
test_values	11.3572μs	1.0482μs	954.0309 KOps/s	943.4178 KOps/s	$\color{#35bf28}+1.12\%$
test_values_nested	0.1432ms	73.9116μs	13.5297 KOps/s	13.0624 KOps/s	$\color{#35bf28}+3.58\%$
test_values_nested_locked	0.1443ms	72.9312μs	13.7116 KOps/s	13.2868 KOps/s	$\color{#35bf28}+3.20\%$
test_values_nested_leaf	0.1306ms	61.5322μs	16.2517 KOps/s	15.9078 KOps/s	$\color{#35bf28}+2.16\%$
test_values_stack_nested	0.1541ms	74.4470μs	13.4324 KOps/s	13.5041 KOps/s	$\color{#d91a1a}-0.53\%$
test_values_stack_nested_leaf	0.1181ms	61.4980μs	16.2607 KOps/s	16.3509 KOps/s	$\color{#d91a1a}-0.55\%$
test_values_stack_nested_locked	0.1423ms	74.6268μs	13.4000 KOps/s	13.6149 KOps/s	$\color{#d91a1a}-1.58\%$
test_membership	3.9831μs	0.7206μs	1.3877 MOps/s	1.3752 MOps/s	$\color{#35bf28}+0.91\%$
test_membership_nested	21.9310μs	2.7890μs	358.5549 KOps/s	358.4124 KOps/s	$\color{#35bf28}+0.04\%$
test_membership_nested_leaf	22.3730μs	2.7979μs	357.4127 KOps/s	358.1231 KOps/s	$\color{#d91a1a}-0.20\%$
test_membership_stacked_nested	24.1050μs	2.7959μs	357.6612 KOps/s	360.7745 KOps/s	$\color{#d91a1a}-0.86\%$
test_membership_stacked_nested_leaf	30.9780μs	2.7687μs	361.1738 KOps/s	356.6135 KOps/s	$\color{#35bf28}+1.28\%$
test_membership_nested_last	22.3520μs	4.0250μs	248.4462 KOps/s	249.8532 KOps/s	$\color{#d91a1a}-0.56\%$
test_membership_nested_leaf_last	49.8740μs	4.0094μs	249.4144 KOps/s	247.7781 KOps/s	$\color{#35bf28}+0.66\%$
test_membership_stacked_nested_last	23.5850μs	4.5467μs	219.9405 KOps/s	169.6517 KOps/s	$\textbf{\color{#35bf28}+29.64\%}$
test_membership_stacked_nested_leaf_last	48.8740μs	4.5779μs	218.4397 KOps/s	167.9384 KOps/s	$\textbf{\color{#35bf28}+30.07\%}$
test_nested_getleaf	35.1060μs	10.8432μs	92.2234 KOps/s	94.5291 KOps/s	$\color{#d91a1a}-2.44\%$
test_nested_get	38.9930μs	10.2902μs	97.1799 KOps/s	83.9280 KOps/s	$\textbf{\color{#35bf28}+15.79\%}$
test_stacked_getleaf	68.6590μs	10.6957μs	93.4959 KOps/s	73.3471 KOps/s	$\textbf{\color{#35bf28}+27.47\%}$
test_stacked_get	43.0600μs	10.1818μs	98.2145 KOps/s	82.4284 KOps/s	$\textbf{\color{#35bf28}+19.15\%}$
test_nested_getitemleaf	63.2590μs	11.0311μs	90.6529 KOps/s	82.5964 KOps/s	$\textbf{\color{#35bf28}+9.75\%}$
test_nested_getitem	30.9990μs	10.4068μs	96.0908 KOps/s	88.6168 KOps/s	$\textbf{\color{#35bf28}+8.43\%}$
test_stacked_getitemleaf	63.2960μs	10.9339μs	91.4585 KOps/s	80.3606 KOps/s	$\textbf{\color{#35bf28}+13.81\%}$
test_stacked_getitem	44.8140μs	10.4258μs	95.9155 KOps/s	87.5394 KOps/s	$\textbf{\color{#35bf28}+9.57\%}$
test_lock_nested	83.9395ms	0.5747ms	1.7399 KOps/s	1.9433 KOps/s	$\textbf{\color{#d91a1a}-10.46\%}$
test_lock_stack_nested	0.7025ms	0.4539ms	2.2032 KOps/s	2.1884 KOps/s	$\color{#35bf28}+0.68\%$
test_unlock_nested	82.3425ms	0.4890ms	2.0449 KOps/s	2.4544 KOps/s	$\textbf{\color{#d91a1a}-16.68\%}$
test_unlock_stack_nested	0.7333ms	0.3719ms	2.6892 KOps/s	2.6481 KOps/s	$\color{#35bf28}+1.55\%$
test_flatten_speed	0.1808ms	87.8620μs	11.3815 KOps/s	9.3829 KOps/s	$\textbf{\color{#35bf28}+21.30\%}$
test_unflatten_speed	0.6489ms	0.4631ms	2.1594 KOps/s	1.8769 KOps/s	$\textbf{\color{#35bf28}+15.05\%}$
test_common_ops	4.8491ms	1.1491ms	870.2625 Ops/s	772.5172 Ops/s	$\textbf{\color{#35bf28}+12.65\%}$
test_creation	30.8880μs	2.0984μs	476.5433 KOps/s	472.8116 KOps/s	$\color{#35bf28}+0.79\%$
test_creation_empty	51.7070μs	20.1048μs	49.7393 KOps/s	45.6777 KOps/s	$\textbf{\color{#35bf28}+8.89\%}$
test_creation_nested_1	64.9810μs	23.2287μs	43.0501 KOps/s	49.7397 KOps/s	$\textbf{\color{#d91a1a}-13.45\%}$
test_creation_nested_2	76.9050μs	27.9362μs	35.7959 KOps/s	34.3380 KOps/s	$\color{#35bf28}+4.25\%$
test_clone	75.9230μs	16.7699μs	59.6308 KOps/s	56.5050 KOps/s	$\textbf{\color{#35bf28}+5.53\%}$
test_getitem[int]	1.0798ms	16.5717μs	60.3437 KOps/s	58.2997 KOps/s	$\color{#35bf28}+3.51\%$
test_getitem[slice_int]	0.1364ms	29.5335μs	33.8599 KOps/s	31.6424 KOps/s	$\textbf{\color{#35bf28}+7.01\%}$
test_getitem[range]	0.1853ms	55.9905μs	17.8602 KOps/s	16.9710 KOps/s	$\textbf{\color{#35bf28}+5.24\%}$
test_getitem[tuple]	0.1328ms	24.3723μs	41.0302 KOps/s	38.7474 KOps/s	$\textbf{\color{#35bf28}+5.89\%}$
test_getitem[list]	0.1779ms	51.2354μs	19.5178 KOps/s	18.2591 KOps/s	$\textbf{\color{#35bf28}+6.89\%}$
test_setitem_dim[int]	99.4160μs	33.9431μs	29.4610 KOps/s	29.8793 KOps/s	$\color{#d91a1a}-1.40\%$
test_setitem_dim[slice_int]	0.1034ms	61.4205μs	16.2812 KOps/s	15.8627 KOps/s	$\color{#35bf28}+2.64\%$
test_setitem_dim[range]	0.1233ms	82.0889μs	12.1819 KOps/s	11.6666 KOps/s	$\color{#35bf28}+4.42\%$
test_setitem_dim[tuple]	0.1194ms	50.2572μs	19.8977 KOps/s	19.6980 KOps/s	$\color{#35bf28}+1.01\%$
test_setitem	91.4420μs	31.4623μs	31.7841 KOps/s	32.9777 KOps/s	$\color{#d91a1a}-3.62\%$
test_set	85.2800μs	30.4454μs	32.8457 KOps/s	33.4991 KOps/s	$\color{#d91a1a}-1.95\%$
test_set_shared	1.9860ms	0.2138ms	4.6766 KOps/s	4.5945 KOps/s	$\color{#35bf28}+1.79\%$
test_update	0.1402ms	38.5018μs	25.9728 KOps/s	27.4220 KOps/s	$\textbf{\color{#d91a1a}-5.28\%}$
test_update_nested	0.1598ms	47.8007μs	20.9202 KOps/s	20.9647 KOps/s	$\color{#d91a1a}-0.21\%$
test_update__nested	88.2550μs	34.8766μs	28.6725 KOps/s	27.9213 KOps/s	$\color{#35bf28}+2.69\%$
test_set_nested	98.4450μs	32.8774μs	30.4160 KOps/s	31.5034 KOps/s	$\color{#d91a1a}-3.45\%$
test_set_nested_new	0.1082ms	37.7506μs	26.4897 KOps/s	27.0555 KOps/s	$\color{#d91a1a}-2.09\%$
test_select	0.1169ms	54.3558μs	18.3973 KOps/s	18.3752 KOps/s	$\color{#35bf28}+0.12\%$
test_select_nested	0.1372ms	59.1150μs	16.9162 KOps/s	17.1730 KOps/s	$\color{#d91a1a}-1.50\%$
test_exclude_nested	0.1657ms	75.7260μs	13.2055 KOps/s	13.7017 KOps/s	$\color{#d91a1a}-3.62\%$
test_empty[True]	0.5025ms	0.3117ms	3.2077 KOps/s	3.1991 KOps/s	$\color{#35bf28}+0.27\%$
test_empty[False]	10.2443μs	1.2561μs	796.1114 KOps/s	842.3419 KOps/s	$\textbf{\color{#d91a1a}-5.49\%}$
test_unbind_speed	0.6114ms	0.4057ms	2.4651 KOps/s	3.2963 KOps/s	$\textbf{\color{#d91a1a}-25.22\%}$
test_unbind_speed_stack0	0.6845ms	0.3565ms	2.8053 KOps/s	3.3780 KOps/s	$\textbf{\color{#d91a1a}-16.95\%}$
test_unbind_speed_stack1	0.1141s	0.9485ms	1.0543 KOps/s	1.3738 KOps/s	$\textbf{\color{#d91a1a}-23.26\%}$
test_split	86.8470ms	2.3396ms	427.4262 Ops/s	453.9179 Ops/s	$\textbf{\color{#d91a1a}-5.84\%}$
test_chunk	3.4056ms	2.1369ms	467.9657 Ops/s	447.6919 Ops/s	$\color{#35bf28}+4.53\%$
test_creation[device0]	0.2242ms	0.1149ms	8.7042 KOps/s	8.4284 KOps/s	$\color{#35bf28}+3.27\%$
test_creation_from_tensor	5.1334ms	0.1223ms	8.1761 KOps/s	8.3662 KOps/s	$\color{#d91a1a}-2.27\%$
test_add_one[memmap_tensor0]	0.2144ms	7.2817μs	137.3302 KOps/s	128.4618 KOps/s	$\textbf{\color{#35bf28}+6.90\%}$
test_contiguous[memmap_tensor0]	13.3450μs	1.9203μs	520.7414 KOps/s	512.6073 KOps/s	$\color{#35bf28}+1.59\%$
test_stack[memmap_tensor0]	32.1910μs	5.7487μs	173.9515 KOps/s	172.1151 KOps/s	$\color{#35bf28}+1.07\%$
test_memmaptd_index	0.7411ms	0.4037ms	2.4768 KOps/s	2.4617 KOps/s	$\color{#35bf28}+0.61\%$
test_memmaptd_index_astensor	0.7642ms	0.4830ms	2.0704 KOps/s	2.0484 KOps/s	$\color{#35bf28}+1.07\%$
test_memmaptd_index_op	1.6729ms	1.0546ms	948.2705 Ops/s	981.7813 Ops/s	$\color{#d91a1a}-3.41\%$
test_serialize_model	0.2062s	0.1305s	7.6648 Ops/s	8.6328 Ops/s	$\textbf{\color{#d91a1a}-11.21\%}$
test_serialize_model_pickle	0.4677s	0.3993s	2.5041 Ops/s	2.5661 Ops/s	$\color{#d91a1a}-2.42\%$
test_serialize_weights	0.1227s	0.1176s	8.5027 Ops/s	8.6556 Ops/s	$\color{#d91a1a}-1.77\%$
test_serialize_weights_returnearly	0.2105s	0.1898s	5.2695 Ops/s	6.2607 Ops/s	$\textbf{\color{#d91a1a}-15.83\%}$
test_serialize_weights_pickle	0.5567s	0.4405s	2.2703 Ops/s	2.5170 Ops/s	$\textbf{\color{#d91a1a}-9.80\%}$
test_serialize_weights_filesystem	0.2201s	0.1516s	6.5955 Ops/s	6.9735 Ops/s	$\textbf{\color{#d91a1a}-5.42\%}$
test_serialize_model_filesystem	0.1615s	0.1474s	6.7826 Ops/s	6.9781 Ops/s	$\color{#d91a1a}-2.80\%$
test_reshape_pytree	79.5690μs	38.4837μs	25.9850 KOps/s	24.9799 KOps/s	$\color{#35bf28}+4.02\%$
test_reshape_td	96.0900μs	46.4689μs	21.5198 KOps/s	20.7855 KOps/s	$\color{#35bf28}+3.53\%$
test_view_pytree	0.1179ms	38.0882μs	26.2548 KOps/s	25.3091 KOps/s	$\color{#35bf28}+3.74\%$
test_view_td	0.1293ms	51.2868μs	19.4982 KOps/s	18.2789 KOps/s	$\textbf{\color{#35bf28}+6.67\%}$
test_unbind_pytree	79.6400μs	35.4489μs	28.2096 KOps/s	27.3592 KOps/s	$\color{#35bf28}+3.11\%$
test_unbind_td	0.3070ms	45.0914μs	22.1772 KOps/s	21.5545 KOps/s	$\color{#35bf28}+2.89\%$
test_split_pytree	75.0710μs	37.5558μs	26.6271 KOps/s	25.8551 KOps/s	$\color{#35bf28}+2.99\%$
test_split_td	0.5059ms	55.8917μs	17.8917 KOps/s	17.3705 KOps/s	$\color{#35bf28}+3.00\%$
test_add_pytree	0.1308ms	44.6468μs	22.3980 KOps/s	21.6775 KOps/s	$\color{#35bf28}+3.32\%$
test_add_td	0.2988ms	85.8300μs	11.6509 KOps/s	12.5795 KOps/s	$\textbf{\color{#d91a1a}-7.38\%}$
test_compile_add_one_nested[tensordict-compile]	0.1185ms	57.8613μs	17.2827 KOps/s	15.9241 KOps/s	$\textbf{\color{#35bf28}+8.53\%}$
test_compile_add_one_nested[tensordict-eager]	0.2645ms	0.1781ms	5.6158 KOps/s	5.0154 KOps/s	$\textbf{\color{#35bf28}+11.97\%}$
test_compile_add_one_nested[pytree-compile]	0.1441ms	57.8683μs	17.2806 KOps/s	15.9920 KOps/s	$\textbf{\color{#35bf28}+8.06\%}$
test_compile_add_one_nested[pytree-eager]	0.2530ms	0.1384ms	7.2229 KOps/s	6.6722 KOps/s	$\textbf{\color{#35bf28}+8.25\%}$
test_compile_copy_nested[tensordict-compile]	61.5550μs	21.3230μs	46.8978 KOps/s	48.0556 KOps/s	$\color{#d91a1a}-2.41\%$
test_compile_copy_nested[tensordict-eager]	0.1569ms	68.4555μs	14.6080 KOps/s	14.6095 KOps/s	$-0.01\%$
test_compile_copy_nested[pytree-compile]	0.1619ms	74.6586μs	13.3943 KOps/s	13.2687 KOps/s	$\color{#35bf28}+0.95\%$
test_compile_copy_nested[pytree-eager]	0.1567ms	68.8357μs	14.5273 KOps/s	14.4625 KOps/s	$\color{#35bf28}+0.45\%$
test_compile_add_one_flat[tensordict-compile]	0.2925ms	0.1735ms	5.7637 KOps/s	5.6563 KOps/s	$\color{#35bf28}+1.90\%$
test_compile_add_one_flat[tensordict-eager]	0.3992ms	0.1970ms	5.0772 KOps/s	5.1969 KOps/s	$\color{#d91a1a}-2.30\%$
test_compile_add_one_flat[tensorclass-compile]	96.4310μs	46.0723μs	21.7050 KOps/s	21.2242 KOps/s	$\color{#35bf28}+2.27\%$
test_compile_add_one_flat[tensorclass-eager]	0.1723ms	68.4291μs	14.6137 KOps/s	13.3597 KOps/s	$\textbf{\color{#35bf28}+9.39\%}$
test_compile_add_one_flat[pytree-compile]	0.3691ms	0.1795ms	5.5697 KOps/s	5.1553 KOps/s	$\textbf{\color{#35bf28}+8.04\%}$
test_compile_add_one_flat[pytree-eager]	0.6184ms	0.2860ms	3.4968 KOps/s	3.1149 KOps/s	$\textbf{\color{#35bf28}+12.26\%}$
test_compile_add_self_flat[tensordict-eager]	0.3763ms	0.2072ms	4.8252 KOps/s	4.9234 KOps/s	$\color{#d91a1a}-1.99\%$
test_compile_add_self_flat[tensordict-compile]	0.4294ms	0.1778ms	5.6234 KOps/s	5.7167 KOps/s	$\color{#d91a1a}-1.63\%$
test_compile_add_self_flat[tensorclass-eager]	0.1206ms	61.2634μs	16.3230 KOps/s	15.8772 KOps/s	$\color{#35bf28}+2.81\%$
test_compile_add_self_flat[tensorclass-compile]	0.1104ms	46.6363μs	21.4425 KOps/s	21.2740 KOps/s	$\color{#35bf28}+0.79\%$
test_compile_add_self_flat[pytree-eager]	0.4082ms	0.2321ms	4.3086 KOps/s	4.2064 KOps/s	$\color{#35bf28}+2.43\%$
test_compile_add_self_flat[pytree-compile]	0.2752ms	0.1762ms	5.6753 KOps/s	5.6036 KOps/s	$\color{#35bf28}+1.28\%$
test_compile_copy_flat[tensordict-compile]	0.1903ms	0.1039ms	9.6208 KOps/s	9.4875 KOps/s	$\color{#35bf28}+1.41\%$
test_compile_copy_flat[tensordict-eager]	0.1362ms	58.2443μs	17.1691 KOps/s	17.1406 KOps/s	$\color{#35bf28}+0.17\%$
test_compile_copy_flat[pytree-compile]	0.1753ms	77.8832μs	12.8397 KOps/s	12.7644 KOps/s	$\color{#35bf28}+0.59\%$
test_compile_copy_flat[pytree-eager]	0.1353ms	69.2336μs	14.4438 KOps/s	14.4815 KOps/s	$\color{#d91a1a}-0.26\%$
test_compile_assign_and_add[tensordict-compile]	0.4984ms	0.2319ms	4.3115 KOps/s	4.9605 KOps/s	$\textbf{\color{#d91a1a}-13.08\%}$
test_compile_assign_and_add[tensordict-eager]	3.2491ms	1.9750ms	506.3215 Ops/s	607.8096 Ops/s	$\textbf{\color{#d91a1a}-16.70\%}$
test_compile_assign_and_add[pytree-compile]	0.2935ms	0.1950ms	5.1288 KOps/s	4.9048 KOps/s	$\color{#35bf28}+4.57\%$
test_compile_assign_and_add[pytree-eager]	1.8538ms	1.1009ms	908.3140 Ops/s	898.1401 Ops/s	$\color{#35bf28}+1.13\%$
test_compile_assign_and_add_stack[compile]	0.5367ms	0.4267ms	2.3438 KOps/s	2.1862 KOps/s	$\textbf{\color{#35bf28}+7.21\%}$
test_compile_assign_and_add_stack[eager]	6.1418ms	3.9538ms	252.9212 Ops/s	263.3389 Ops/s	$\color{#d91a1a}-3.96\%$
test_compile_indexing[tensor-tensordict-compile]	95.5600μs	34.3930μs	29.0756 KOps/s	26.6013 KOps/s	$\textbf{\color{#35bf28}+9.30\%}$
test_compile_indexing[tensor-tensordict-eager]	1.0826ms	47.5016μs	21.0519 KOps/s	20.0337 KOps/s	$\textbf{\color{#35bf28}+5.08\%}$
test_compile_indexing[tensor-tensorclass-compile]	0.1071ms	29.7825μs	33.5767 KOps/s	32.7048 KOps/s	$\color{#35bf28}+2.67\%$
test_compile_indexing[tensor-tensorclass-eager]	69.8400μs	27.9961μs	35.7193 KOps/s	33.5615 KOps/s	$\textbf{\color{#35bf28}+6.43\%}$
test_compile_indexing[tensor-pytree-compile]	0.1037ms	30.3641μs	32.9336 KOps/s	32.8119 KOps/s	$\color{#35bf28}+0.37\%$
test_compile_indexing[tensor-pytree-eager]	0.1080ms	27.8343μs	35.9269 KOps/s	34.4605 KOps/s	$\color{#35bf28}+4.26\%$
test_compile_indexing[slice-tensordict-compile]	0.1525ms	74.8995μs	13.3512 KOps/s	12.8487 KOps/s	$\color{#35bf28}+3.91\%$
test_compile_indexing[slice-tensordict-eager]	0.5362ms	26.5341μs	37.6874 KOps/s	35.1544 KOps/s	$\textbf{\color{#35bf28}+7.21\%}$
test_compile_indexing[slice-tensorclass-compile]	0.1643ms	68.7645μs	14.5424 KOps/s	13.9402 KOps/s	$\color{#35bf28}+4.32\%$
test_compile_indexing[slice-tensorclass-eager]	97.9830μs	22.7241μs	44.0062 KOps/s	36.9017 KOps/s	$\textbf{\color{#35bf28}+19.25\%}$
test_compile_indexing[slice-pytree-compile]	0.1844ms	67.6443μs	14.7832 KOps/s	11.6915 KOps/s	$\textbf{\color{#35bf28}+26.44\%}$
test_compile_indexing[slice-pytree-eager]	63.6490μs	22.8397μs	43.7835 KOps/s	35.7917 KOps/s	$\textbf{\color{#35bf28}+22.33\%}$
test_compile_indexing[int-tensordict-compile]	0.1578ms	73.3223μs	13.6384 KOps/s	11.4425 KOps/s	$\textbf{\color{#35bf28}+19.19\%}$
test_compile_indexing[int-tensordict-eager]	0.9490ms	26.4064μs	37.8696 KOps/s	31.5860 KOps/s	$\textbf{\color{#35bf28}+19.89\%}$
test_compile_indexing[int-tensorclass-compile]	0.1537ms	68.1420μs	14.6752 KOps/s	13.2602 KOps/s	$\textbf{\color{#35bf28}+10.67\%}$
test_compile_indexing[int-tensorclass-eager]	63.8700μs	22.8524μs	43.7591 KOps/s	41.5358 KOps/s	$\textbf{\color{#35bf28}+5.35\%}$
test_compile_indexing[int-pytree-compile]	0.1487ms	67.7100μs	14.7689 KOps/s	14.1500 KOps/s	$\color{#35bf28}+4.37\%$
test_compile_indexing[int-pytree-eager]	92.7040μs	22.9731μs	43.5293 KOps/s	42.0989 KOps/s	$\color{#35bf28}+3.40\%$
test_mod_add[eager]	89.5790μs	27.0626μs	36.9514 KOps/s	40.3183 KOps/s	$\textbf{\color{#d91a1a}-8.35\%}$
test_mod_add[compile]	0.1406ms	39.6661μs	25.2104 KOps/s	25.8498 KOps/s	$\color{#d91a1a}-2.47\%$
test_mod_add[compile-overhead]	86.4520μs	38.8741μs	25.7241 KOps/s	25.1170 KOps/s	$\color{#35bf28}+2.42\%$
test_mod_wrap[eager]	0.3202ms	0.2092ms	4.7804 KOps/s	4.7674 KOps/s	$\color{#35bf28}+0.27\%$
test_mod_wrap[compile]	0.3873ms	0.2320ms	4.3111 KOps/s	4.2520 KOps/s	$\color{#35bf28}+1.39\%$
test_mod_wrap[compile-overhead]	0.3627ms	0.2310ms	4.3299 KOps/s	4.2879 KOps/s	$\color{#35bf28}+0.98\%$
test_mod_wrap_and_backward[eager]	14.2373ms	11.4446ms	87.3774 Ops/s	86.5955 Ops/s	$\color{#35bf28}+0.90\%$
test_mod_wrap_and_backward[compile]	13.9040ms	11.7950ms	84.7819 Ops/s	77.5904 Ops/s	$\textbf{\color{#35bf28}+9.27\%}$
test_mod_wrap_and_backward[compile-overhead]	14.1129ms	12.3865ms	80.7331 Ops/s	84.3387 Ops/s	$\color{#d91a1a}-4.28\%$
test_seq_add[eager]	0.2411ms	93.8521μs	10.6551 KOps/s	10.8211 KOps/s	$\color{#d91a1a}-1.53\%$
test_seq_add[compile]	0.1334ms	64.9264μs	15.4021 KOps/s	15.4690 KOps/s	$\color{#d91a1a}-0.43\%$
test_seq_add[compile-overhead]	0.1659ms	63.4129μs	15.7697 KOps/s	15.4418 KOps/s	$\color{#35bf28}+2.12\%$
test_seq_wrap[eager]	0.5044ms	0.3928ms	2.5458 KOps/s	2.6479 KOps/s	$\color{#d91a1a}-3.86\%$
test_seq_wrap[compile]	1.2061ms	0.2758ms	3.6253 KOps/s	3.6631 KOps/s	$\color{#d91a1a}-1.03\%$
test_seq_wrap[compile-overhead]	1.9764ms	0.3087ms	3.2390 KOps/s	3.6957 KOps/s	$\textbf{\color{#d91a1a}-12.36\%}$
test_func_call_runtime[False-eager]	1.6400ms	0.5866ms	1.7048 KOps/s	1.8431 KOps/s	$\textbf{\color{#d91a1a}-7.50\%}$
test_func_call_runtime[False-compile]	1.0058ms	0.5365ms	1.8640 KOps/s	1.9669 KOps/s	$\textbf{\color{#d91a1a}-5.23\%}$
test_func_call_runtime[False-compile-overhead]	0.6746ms	0.5007ms	1.9970 KOps/s	1.9628 KOps/s	$\color{#35bf28}+1.74\%$
test_func_call_runtime[True-eager]	0.8620ms	0.7449ms	1.3425 KOps/s	1.3195 KOps/s	$\color{#35bf28}+1.74\%$
test_func_call_runtime[True-compile]	0.6764ms	0.5151ms	1.9412 KOps/s	1.7085 KOps/s	$\textbf{\color{#35bf28}+13.62\%}$
test_func_call_runtime[True-compile-overhead]	0.6309ms	0.5114ms	1.9554 KOps/s	1.9159 KOps/s	$\color{#35bf28}+2.06\%$
test_func_call_cm_runtime[False-eager]	1.4874ms	0.6272ms	1.5945 KOps/s	1.8681 KOps/s	$\textbf{\color{#d91a1a}-14.65\%}$
test_func_call_cm_runtime[False-compile]	1.1813ms	0.5233ms	1.9109 KOps/s	1.9500 KOps/s	$\color{#d91a1a}-2.01\%$
test_func_call_cm_runtime[False-compile-overhead]	0.9501ms	0.5072ms	1.9715 KOps/s	1.9592 KOps/s	$\color{#35bf28}+0.63\%$
test_func_call_cm_runtime[True-eager]	1.0119ms	0.8664ms	1.1542 KOps/s	1.1217 KOps/s	$\color{#35bf28}+2.90\%$
test_func_call_cm_runtime[True-compile]	1.0509ms	0.7300ms	1.3699 KOps/s	1.1478 KOps/s	$\textbf{\color{#35bf28}+19.35\%}$
test_func_call_cm_runtime[True-compile-overhead]	1.1779ms	0.7360ms	1.3587 KOps/s	1.3190 KOps/s	$\color{#35bf28}+3.01\%$
test_vmap_func_call_cm_runtime[eager]	2.6206ms	1.8649ms	536.2259 Ops/s	530.0275 Ops/s	$\color{#35bf28}+1.17\%$
test_vmap_func_call_cm_runtime[compile]	2.4900ms	1.8917ms	528.6301 Ops/s	514.2960 Ops/s	$\color{#35bf28}+2.79\%$
test_vmap_func_call_cm_runtime[compile-overhead]	2.7097ms	1.9254ms	519.3708 Ops/s	508.7570 Ops/s	$\color{#35bf28}+2.09\%$
test_distributed	0.2521ms	0.1259ms	7.9403 KOps/s	7.6638 KOps/s	$\color{#35bf28}+3.61\%$
test_tdmodule	46.8680μs	20.2493μs	49.3845 KOps/s	57.8097 KOps/s	$\textbf{\color{#d91a1a}-14.57\%}$
test_tdmodule_dispatch	72.4250μs	39.3421μs	25.4181 KOps/s	29.2806 KOps/s	$\textbf{\color{#d91a1a}-13.19\%}$
test_tdseq	42.4200μs	22.8410μs	43.7809 KOps/s	51.2625 KOps/s	$\textbf{\color{#d91a1a}-14.59\%}$
test_tdseq_dispatch	68.6090μs	44.7773μs	22.3327 KOps/s	25.3506 KOps/s	$\textbf{\color{#d91a1a}-11.90\%}$
test_instantiation_functorch	2.5220ms	1.5832ms	631.6355 Ops/s	618.5979 Ops/s	$\color{#35bf28}+2.11\%$
test_instantiation_td	1.9181ms	1.1624ms	860.2716 Ops/s	847.2131 Ops/s	$\color{#35bf28}+1.54\%$
test_exec_functorch	0.3502ms	0.1862ms	5.3705 KOps/s	5.4202 KOps/s	$\color{#d91a1a}-0.92\%$
test_exec_functional_call	0.3433ms	0.1703ms	5.8732 KOps/s	5.5886 KOps/s	$\textbf{\color{#35bf28}+5.09\%}$
test_exec_td	0.3052ms	0.1689ms	5.9208 KOps/s	5.8982 KOps/s	$\color{#35bf28}+0.38\%$
test_exec_td_decorator	0.3462ms	0.2223ms	4.4986 KOps/s	4.4003 KOps/s	$\color{#35bf28}+2.23\%$
test_vmap_mlp_speed[True-True]	0.8025ms	0.6509ms	1.5364 KOps/s	1.4939 KOps/s	$\color{#35bf28}+2.85\%$
test_vmap_mlp_speed[True-False]	0.8657ms	0.6497ms	1.5391 KOps/s	1.5251 KOps/s	$\color{#35bf28}+0.92\%$
test_vmap_mlp_speed[False-True]	0.8067ms	0.4952ms	2.0194 KOps/s	1.9964 KOps/s	$\color{#35bf28}+1.15\%$
test_vmap_mlp_speed[False-False]	0.6713ms	0.4944ms	2.0225 KOps/s	1.9879 KOps/s	$\color{#35bf28}+1.74\%$
test_vmap_mlp_speed_decorator[True-True]	1.3925ms	0.6225ms	1.6065 KOps/s	1.5919 KOps/s	$\color{#35bf28}+0.92\%$
test_vmap_mlp_speed_decorator[True-False]	0.9574ms	0.6232ms	1.6047 KOps/s	1.5784 KOps/s	$\color{#35bf28}+1.67\%$
test_vmap_mlp_speed_decorator[False-True]	0.7795ms	0.5092ms	1.9637 KOps/s	1.9100 KOps/s	$\color{#35bf28}+2.81\%$
test_vmap_mlp_speed_decorator[False-False]	0.7310ms	0.5082ms	1.9679 KOps/s	1.9270 KOps/s	$\color{#35bf28}+2.12\%$
test_to_module_speed[True]	2.0503ms	1.3158ms	759.9883 Ops/s	765.3669 Ops/s	$\color{#d91a1a}-0.70\%$
test_to_module_speed[False]	1.9691ms	1.2763ms	783.4966 Ops/s	790.3830 Ops/s	$\color{#d91a1a}-0.87\%$
test_tc_init	0.1076ms	47.5272μs	21.0406 KOps/s	23.5330 KOps/s	$\textbf{\color{#d91a1a}-10.59\%}$
test_tc_init_nested	0.1678ms	94.2758μs	10.6072 KOps/s	11.6543 KOps/s	$\textbf{\color{#d91a1a}-8.99\%}$
test_tc_first_layer_tensor	16.1900μs	1.5548μs	643.1893 KOps/s	651.5511 KOps/s	$\color{#d91a1a}-1.28\%$
test_tc_first_layer_nontensor	28.9640μs	4.7885μs	208.8345 KOps/s	216.6973 KOps/s	$\color{#d91a1a}-3.63\%$
test_tc_second_layer_tensor	38.6330μs	2.8415μs	351.9266 KOps/s	359.1640 KOps/s	$\color{#d91a1a}-2.02\%$
test_tc_second_layer_nontensor	43.7920μs	6.1898μs	161.5550 KOps/s	167.1201 KOps/s	$\color{#d91a1a}-3.33\%$
test_unbind	0.4589s	14.9766ms	66.7708 Ops/s	77.5994 Ops/s	$\textbf{\color{#d91a1a}-13.95\%}$
test_full_like	7.7109ms	6.8349ms	146.3078 Ops/s	142.0332 Ops/s	$\color{#35bf28}+3.01\%$
test_zeros_like	2.9933ms	2.6335ms	379.7248 Ops/s	372.2304 Ops/s	$\color{#35bf28}+2.01\%$
test_ones_like	10.0016ms	5.8865ms	169.8805 Ops/s	326.0933 Ops/s	$\textbf{\color{#d91a1a}-47.90\%}$
test_clone	13.9636ms	7.7088ms	129.7223 Ops/s	207.5325 Ops/s	$\textbf{\color{#d91a1a}-37.49\%}$
test_squeeze	75.4920μs	13.0917μs	76.3840 KOps/s	79.2947 KOps/s	$\color{#d91a1a}-3.67\%$
test_unsqueeze	0.1694ms	90.0702μs	11.1025 KOps/s	10.4247 KOps/s	$\textbf{\color{#35bf28}+6.50\%}$
test_split	0.5553ms	0.1893ms	5.2828 KOps/s	4.9494 KOps/s	$\textbf{\color{#35bf28}+6.74\%}$
test_permute	0.3666ms	0.2171ms	4.6071 KOps/s	4.3818 KOps/s	$\textbf{\color{#35bf28}+5.14\%}$
test_stack	26.1090ms	24.1343ms	41.4349 Ops/s	41.8202 Ops/s	$\color{#d91a1a}-0.92\%$
test_cat	26.2814ms	23.8884ms	41.8613 Ops/s	41.8339 Ops/s	$\color{#35bf28}+0.07\%$

[github-actions]

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

Total Benchmarks: 228. Improved: $\large\color{#35bf28}14$. Worsened: $\large\color{#d91a1a}16$.

Expand to view detailed results

Name	Max	Mean	Ops	Ops on Repo HEAD	Change
test_plain_set_nested	0.1423ms	13.4781μs	74.1944 KOps/s	73.2826 KOps/s	$\color{#35bf28}+1.24\%$
test_plain_set_stack_nested	36.0900μs	13.6054μs	73.5004 KOps/s	73.0380 KOps/s	$\color{#35bf28}+0.63\%$
test_plain_set_nested_inplace	64.2610μs	14.5842μs	68.5676 KOps/s	67.6228 KOps/s	$\color{#35bf28}+1.40\%$
test_plain_set_stack_nested_inplace	49.4900μs	14.6887μs	68.0794 KOps/s	67.9551 KOps/s	$\color{#35bf28}+0.18\%$
test_items	36.9110μs	2.9916μs	334.2701 KOps/s	348.2568 KOps/s	$\color{#d91a1a}-4.02\%$
test_items_nested	0.3791ms	0.3251ms	3.0762 KOps/s	3.0654 KOps/s	$\color{#35bf28}+0.35\%$
test_items_nested_locked	0.3684ms	0.3235ms	3.0911 KOps/s	3.0427 KOps/s	$\color{#35bf28}+1.59\%$
test_items_nested_leaf	79.2720μs	55.4711μs	18.0274 KOps/s	17.8418 KOps/s	$\color{#35bf28}+1.04\%$
test_items_stack_nested	0.3734ms	0.3241ms	3.0859 KOps/s	3.0214 KOps/s	$\color{#35bf28}+2.14\%$
test_items_stack_nested_leaf	83.2910μs	55.9829μs	17.8626 KOps/s	17.4104 KOps/s	$\color{#35bf28}+2.60\%$
test_items_stack_nested_locked	0.3757ms	0.3246ms	3.0804 KOps/s	2.9910 KOps/s	$\color{#35bf28}+2.99\%$
test_keys	39.0510μs	3.4001μs	294.1129 KOps/s	294.5316 KOps/s	$\color{#d91a1a}-0.14\%$
test_keys_nested	87.0010μs	56.6905μs	17.6396 KOps/s	17.9659 KOps/s	$\color{#d91a1a}-1.82\%$
test_keys_nested_locked	2.5389ms	62.6206μs	15.9692 KOps/s	15.9527 KOps/s	$\color{#35bf28}+0.10\%$
test_keys_nested_leaf	74.1610μs	46.0415μs	21.7195 KOps/s	21.0003 KOps/s	$\color{#35bf28}+3.43\%$
test_keys_stack_nested	89.3110μs	56.4150μs	17.7258 KOps/s	17.5387 KOps/s	$\color{#35bf28}+1.07\%$
test_keys_stack_nested_leaf	79.6110μs	47.5032μs	21.0512 KOps/s	21.1425 KOps/s	$\color{#d91a1a}-0.43\%$
test_keys_stack_nested_locked	90.1420μs	62.1923μs	16.0792 KOps/s	15.9650 KOps/s	$\color{#35bf28}+0.71\%$
test_values	3.5830μs	0.8362μs	1.1958 MOps/s	1.1847 MOps/s	$\color{#35bf28}+0.94\%$
test_values_nested	70.4610μs	41.2954μs	24.2158 KOps/s	24.1881 KOps/s	$\color{#35bf28}+0.11\%$
test_values_nested_locked	77.5310μs	43.2008μs	23.1477 KOps/s	23.2035 KOps/s	$\color{#d91a1a}-0.24\%$
test_values_nested_leaf	64.2810μs	35.8045μs	27.9295 KOps/s	27.8958 KOps/s	$\color{#35bf28}+0.12\%$
test_values_stack_nested	76.2720μs	41.4895μs	24.1025 KOps/s	23.6376 KOps/s	$\color{#35bf28}+1.97\%$
test_values_stack_nested_leaf	69.0610μs	36.3362μs	27.5207 KOps/s	27.4096 KOps/s	$\color{#35bf28}+0.41\%$
test_values_stack_nested_locked	76.3910μs	43.3493μs	23.0684 KOps/s	22.7081 KOps/s	$\color{#35bf28}+1.59\%$
test_membership	1.9535μs	0.5007μs	1.9973 MOps/s	1.9903 MOps/s	$\color{#35bf28}+0.35\%$
test_membership_nested	15.2805μs	1.8690μs	535.0445 KOps/s	514.6768 KOps/s	$\color{#35bf28}+3.96\%$
test_membership_nested_leaf	11.7800μs	1.8409μs	543.2007 KOps/s	515.2955 KOps/s	$\textbf{\color{#35bf28}+5.42\%}$
test_membership_stacked_nested	48.2700μs	1.9137μs	522.5501 KOps/s	513.5835 KOps/s	$\color{#35bf28}+1.75\%$
test_membership_stacked_nested_leaf	27.0300μs	1.9453μs	514.0492 KOps/s	514.4970 KOps/s	$\color{#d91a1a}-0.09\%$
test_membership_nested_last	34.0000μs	2.7752μs	360.3346 KOps/s	358.3612 KOps/s	$\color{#35bf28}+0.55\%$
test_membership_nested_leaf_last	36.4400μs	2.7970μs	357.5280 KOps/s	359.3850 KOps/s	$\color{#d91a1a}-0.52\%$
test_membership_stacked_nested_last	33.6210μs	2.7900μs	358.4272 KOps/s	305.3557 KOps/s	$\textbf{\color{#35bf28}+17.38\%}$
test_membership_stacked_nested_leaf_last	44.4010μs	2.8015μs	356.9518 KOps/s	316.6232 KOps/s	$\textbf{\color{#35bf28}+12.74\%}$
test_nested_getleaf	53.5610μs	6.0785μs	164.5154 KOps/s	164.8446 KOps/s	$\color{#d91a1a}-0.20\%$
test_nested_get	30.1400μs	5.6843μs	175.9231 KOps/s	175.7009 KOps/s	$\color{#35bf28}+0.13\%$
test_stacked_getleaf	49.0810μs	6.0303μs	165.8291 KOps/s	165.0051 KOps/s	$\color{#35bf28}+0.50\%$
test_stacked_get	39.1100μs	5.6837μs	175.9431 KOps/s	177.7379 KOps/s	$\color{#d91a1a}-1.01\%$
test_nested_getitemleaf	30.5700μs	6.1464μs	162.6968 KOps/s	163.3697 KOps/s	$\color{#d91a1a}-0.41\%$
test_nested_getitem	36.0600μs	5.7376μs	174.2878 KOps/s	175.3553 KOps/s	$\color{#d91a1a}-0.61\%$
test_stacked_getitemleaf	47.1010μs	6.1323μs	163.0696 KOps/s	164.5605 KOps/s	$\color{#d91a1a}-0.91\%$
test_stacked_getitem	37.8700μs	5.7496μs	173.9244 KOps/s	176.9446 KOps/s	$\color{#d91a1a}-1.71\%$
test_lock_nested	4.6849ms	0.4198ms	2.3822 KOps/s	2.3853 KOps/s	$\color{#d91a1a}-0.13\%$
test_lock_stack_nested	0.4349ms	0.3783ms	2.6434 KOps/s	2.6540 KOps/s	$\color{#d91a1a}-0.40\%$
test_unlock_nested	0.7371ms	0.3542ms	2.8236 KOps/s	2.8306 KOps/s	$\color{#d91a1a}-0.25\%$
test_unlock_stack_nested	0.3784ms	0.3186ms	3.1383 KOps/s	3.1490 KOps/s	$\color{#d91a1a}-0.34\%$
test_flatten_speed	0.1185ms	68.8184μs	14.5310 KOps/s	14.5185 KOps/s	$\color{#35bf28}+0.09\%$
test_unflatten_speed	0.3390ms	0.2869ms	3.4851 KOps/s	3.4913 KOps/s	$\color{#d91a1a}-0.18\%$
test_common_ops	1.4942ms	1.2040ms	830.5443 Ops/s	815.1284 Ops/s	$\color{#35bf28}+1.89\%$
test_creation	29.7910μs	1.4717μs	679.5032 KOps/s	694.4175 KOps/s	$\color{#d91a1a}-2.15\%$
test_creation_empty	51.3410μs	14.7177μs	67.9454 KOps/s	68.6038 KOps/s	$\color{#d91a1a}-0.96\%$
test_creation_nested_1	49.1710μs	16.5256μs	60.5121 KOps/s	60.2004 KOps/s	$\color{#35bf28}+0.52\%$
test_creation_nested_2	51.5700μs	18.9357μs	52.8104 KOps/s	52.1240 KOps/s	$\color{#35bf28}+1.32\%$
test_clone	61.1110μs	27.9360μs	35.7961 KOps/s	33.1360 KOps/s	$\textbf{\color{#35bf28}+8.03\%}$
test_getitem[int]	91.6347ms	22.8879μs	43.6913 KOps/s	64.0822 KOps/s	$\textbf{\color{#d91a1a}-31.82\%}$
test_getitem[slice_int]	0.1158ms	27.2425μs	36.7073 KOps/s	36.7136 KOps/s	$\color{#d91a1a}-0.02\%$
test_getitem[range]	0.2442ms	0.1077ms	9.2888 KOps/s	9.2332 KOps/s	$\color{#35bf28}+0.60\%$
test_getitem[tuple]	0.1230ms	23.3292μs	42.8647 KOps/s	42.6426 KOps/s	$\color{#35bf28}+0.52\%$
test_getitem[list]	0.1894ms	96.7958μs	10.3310 KOps/s	10.2136 KOps/s	$\color{#35bf28}+1.15\%$
test_setitem_dim[int]	67.9210μs	43.5903μs	22.9409 KOps/s	22.5965 KOps/s	$\color{#35bf28}+1.52\%$
test_setitem_dim[slice_int]	92.6110μs	66.5557μs	15.0250 KOps/s	15.1524 KOps/s	$\color{#d91a1a}-0.84\%$
test_setitem_dim[range]	0.1698ms	0.1246ms	8.0238 KOps/s	7.9956 KOps/s	$\color{#35bf28}+0.35\%$
test_setitem_dim[tuple]	95.7410μs	59.9529μs	16.6798 KOps/s	16.5007 KOps/s	$\color{#35bf28}+1.09\%$
test_setitem	78.6520μs	40.1610μs	24.8998 KOps/s	24.3346 KOps/s	$\color{#35bf28}+2.32\%$
test_set	70.6310μs	39.2072μs	25.5055 KOps/s	24.0163 KOps/s	$\textbf{\color{#35bf28}+6.20\%}$
test_set_shared	0.3532ms	49.0214μs	20.3992 KOps/s	19.6759 KOps/s	$\color{#35bf28}+3.68\%$
test_update	86.9520μs	46.8548μs	21.3425 KOps/s	20.7506 KOps/s	$\color{#35bf28}+2.85\%$
test_update_nested	81.2620μs	54.0440μs	18.5035 KOps/s	18.1390 KOps/s	$\color{#35bf28}+2.01\%$
test_update__nested	95.8710μs	56.7743μs	17.6136 KOps/s	17.3217 KOps/s	$\color{#35bf28}+1.69\%$
test_set_nested	72.6110μs	41.4062μs	24.1510 KOps/s	23.0828 KOps/s	$\color{#35bf28}+4.63\%$
test_set_nested_new	83.5210μs	44.8928μs	22.2753 KOps/s	21.5610 KOps/s	$\color{#35bf28}+3.31\%$
test_select	0.5408ms	58.4989μs	17.0943 KOps/s	16.8132 KOps/s	$\color{#35bf28}+1.67\%$
test_select_nested	79.8220μs	41.7820μs	23.9337 KOps/s	23.7951 KOps/s	$\color{#35bf28}+0.58\%$
test_exclude_nested	84.0010μs	58.5599μs	17.0765 KOps/s	17.3399 KOps/s	$\color{#d91a1a}-1.52\%$
test_empty[True]	0.2837ms	0.2428ms	4.1182 KOps/s	4.0786 KOps/s	$\color{#35bf28}+0.97\%$
test_empty[False]	3.4431μs	0.7447μs	1.3428 MOps/s	1.3458 MOps/s	$\color{#d91a1a}-0.22\%$
test_to	56.1210μs	25.6997μs	38.9109 KOps/s	38.4271 KOps/s	$\color{#35bf28}+1.26\%$
test_to_nonblocking	60.7810μs	24.3804μs	41.0166 KOps/s	39.9256 KOps/s	$\color{#35bf28}+2.73\%$
test_unbind_speed	0.3062ms	0.2779ms	3.5981 KOps/s	3.6549 KOps/s	$\color{#d91a1a}-1.55\%$
test_unbind_speed_stack0	0.3674ms	0.2788ms	3.5865 KOps/s	3.6690 KOps/s	$\color{#d91a1a}-2.25\%$
test_unbind_speed_stack1	90.6798ms	0.7130ms	1.4026 KOps/s	1.4024 KOps/s	$\color{#35bf28}+0.01\%$
test_split	93.3989ms	2.1838ms	457.9092 Ops/s	469.7142 Ops/s	$\color{#d91a1a}-2.51\%$
test_chunk	93.5502ms	2.1884ms	456.9505 Ops/s	469.1022 Ops/s	$\color{#d91a1a}-2.59\%$
test_creation[device0]	0.3892ms	0.1255ms	7.9687 KOps/s	8.0338 KOps/s	$\color{#d91a1a}-0.81\%$
test_creation_from_tensor	0.4277ms	0.1282ms	7.7985 KOps/s	7.5976 KOps/s	$\color{#35bf28}+2.64\%$
test_add_one[memmap_tensor0]	0.2908ms	8.5004μs	117.6422 KOps/s	116.0175 KOps/s	$\color{#35bf28}+1.40\%$
test_contiguous[memmap_tensor0]	31.4810μs	2.1005μs	476.0829 KOps/s	474.1897 KOps/s	$\color{#35bf28}+0.40\%$
test_stack[memmap_tensor0]	37.3400μs	6.6454μs	150.4806 KOps/s	156.7164 KOps/s	$\color{#d91a1a}-3.98\%$
test_memmaptd_index	1.1140ms	0.4074ms	2.4549 KOps/s	2.4488 KOps/s	$\color{#35bf28}+0.25\%$
test_memmaptd_index_astensor	0.9738ms	0.4675ms	2.1389 KOps/s	2.1507 KOps/s	$\color{#d91a1a}-0.55\%$
test_memmaptd_index_op	1.3989ms	0.9843ms	1.0159 KOps/s	1.0195 KOps/s	$\color{#d91a1a}-0.35\%$
test_serialize_model	0.1300s	0.1291s	7.7456 Ops/s	7.6917 Ops/s	$\color{#35bf28}+0.70\%$
test_serialize_model_pickle	1.3765s	1.2175s	0.8213 Ops/s	0.8241 Ops/s	$\color{#d91a1a}-0.33\%$
test_serialize_weights	0.2173s	0.1416s	7.0630 Ops/s	7.0187 Ops/s	$\color{#35bf28}+0.63\%$
test_serialize_weights_returnearly	0.2168s	54.3828ms	18.3882 Ops/s	17.9311 Ops/s	$\color{#35bf28}+2.55\%$
test_serialize_weights_pickle	1.3767s	1.2172s	0.8216 Ops/s	0.8254 Ops/s	$\color{#d91a1a}-0.46\%$
test_reshape_pytree	67.9810μs	35.2877μs	28.3384 KOps/s	28.9935 KOps/s	$\color{#d91a1a}-2.26\%$
test_reshape_td	69.7510μs	41.9600μs	23.8322 KOps/s	23.6955 KOps/s	$\color{#35bf28}+0.58\%$
test_view_pytree	0.4254ms	34.9150μs	28.6410 KOps/s	28.9056 KOps/s	$\color{#d91a1a}-0.92\%$
test_view_td	81.4410μs	47.6700μs	20.9776 KOps/s	21.3273 KOps/s	$\color{#d91a1a}-1.64\%$
test_unbind_pytree	0.4226ms	33.3162μs	30.0154 KOps/s	29.5940 KOps/s	$\color{#35bf28}+1.42\%$
test_unbind_td	0.5281ms	42.0003μs	23.8093 KOps/s	23.5115 KOps/s	$\color{#35bf28}+1.27\%$
test_split_pytree	0.4154ms	46.3190μs	21.5894 KOps/s	22.3112 KOps/s	$\color{#d91a1a}-3.24\%$
test_split_td	96.7141ms	68.4361μs	14.6122 KOps/s	18.2464 KOps/s	$\textbf{\color{#d91a1a}-19.92\%}$
test_add_pytree	92.4020μs	53.0035μs	18.8667 KOps/s	18.1045 KOps/s	$\color{#35bf28}+4.21\%$
test_add_td	0.4709ms	86.0537μs	11.6207 KOps/s	11.0315 KOps/s	$\textbf{\color{#35bf28}+5.34\%}$
test_compile_add_one_nested[tensordict-compile]	0.4039ms	0.2057ms	4.8609 KOps/s	4.7091 KOps/s	$\color{#35bf28}+3.22\%$
test_compile_add_one_nested[tensordict-eager]	0.2255ms	0.1505ms	6.6427 KOps/s	6.6105 KOps/s	$\color{#35bf28}+0.49\%$
test_compile_add_one_nested[pytree-compile]	0.2029ms	0.1414ms	7.0717 KOps/s	7.0729 KOps/s	$\color{#d91a1a}-0.02\%$
test_compile_add_one_nested[pytree-eager]	0.2379ms	0.1776ms	5.6309 KOps/s	5.5503 KOps/s	$\color{#35bf28}+1.45\%$
test_compile_copy_nested[tensordict-compile]	57.3210μs	21.4823μs	46.5500 KOps/s	47.0058 KOps/s	$\color{#d91a1a}-0.97\%$
test_compile_copy_nested[tensordict-eager]	80.6010μs	43.6085μs	22.9313 KOps/s	22.8701 KOps/s	$\color{#35bf28}+0.27\%$
test_compile_copy_nested[pytree-compile]	0.2058ms	63.5024μs	15.7474 KOps/s	15.6644 KOps/s	$\color{#35bf28}+0.53\%$
test_compile_copy_nested[pytree-eager]	98.7120μs	49.1645μs	20.3399 KOps/s	20.5237 KOps/s	$\color{#d91a1a}-0.90\%$
test_compile_add_one_flat[tensordict-compile]	0.3993ms	0.3088ms	3.2382 KOps/s	3.2327 KOps/s	$\color{#35bf28}+0.17\%$
test_compile_add_one_flat[tensordict-eager]	0.2695ms	0.2082ms	4.8027 KOps/s	4.7757 KOps/s	$\color{#35bf28}+0.57\%$
test_compile_add_one_flat[tensorclass-compile]	0.1654ms	0.1246ms	8.0266 KOps/s	7.6707 KOps/s	$\color{#35bf28}+4.64\%$
test_compile_add_one_flat[tensorclass-eager]	0.1099ms	59.7617μs	16.7331 KOps/s	16.1501 KOps/s	$\color{#35bf28}+3.61\%$
test_compile_add_one_flat[pytree-compile]	0.4120ms	0.3069ms	3.2585 KOps/s	3.2076 KOps/s	$\color{#35bf28}+1.59\%$
test_compile_add_one_flat[pytree-eager]	0.6422ms	0.5982ms	1.6716 KOps/s	1.6327 KOps/s	$\color{#35bf28}+2.38\%$
test_compile_add_self_flat[tensordict-eager]	0.3228ms	0.2478ms	4.0355 KOps/s	4.0178 KOps/s	$\color{#35bf28}+0.44\%$
test_compile_add_self_flat[tensordict-compile]	0.3732ms	0.3094ms	3.2317 KOps/s	3.2326 KOps/s	$\color{#d91a1a}-0.03\%$
test_compile_add_self_flat[tensorclass-eager]	0.1279ms	69.6745μs	14.3525 KOps/s	14.2795 KOps/s	$\color{#35bf28}+0.51\%$
test_compile_add_self_flat[tensorclass-compile]	0.1802ms	0.1251ms	7.9936 KOps/s	7.7907 KOps/s	$\color{#35bf28}+2.60\%$
test_compile_add_self_flat[pytree-eager]	0.6173ms	0.5167ms	1.9353 KOps/s	1.8992 KOps/s	$\color{#35bf28}+1.90\%$
test_compile_add_self_flat[pytree-compile]	0.3728ms	0.3071ms	3.2560 KOps/s	3.2115 KOps/s	$\color{#35bf28}+1.39\%$
test_compile_copy_flat[tensordict-compile]	75.6610μs	17.6717μs	56.5875 KOps/s	50.9541 KOps/s	$\textbf{\color{#35bf28}+11.06\%}$
test_compile_copy_flat[tensordict-eager]	65.5700μs	27.1981μs	36.7672 KOps/s	36.6189 KOps/s	$\color{#35bf28}+0.41\%$
test_compile_copy_flat[pytree-compile]	98.6310μs	68.2908μs	14.6433 KOps/s	14.6635 KOps/s	$\color{#d91a1a}-0.14\%$
test_compile_copy_flat[pytree-eager]	84.6010μs	51.7496μs	19.3238 KOps/s	19.4489 KOps/s	$\color{#d91a1a}-0.64\%$
test_compile_assign_and_add[tensordict-compile]	2.3322ms	0.8269ms	1.2094 KOps/s	1.1420 KOps/s	$\textbf{\color{#35bf28}+5.90\%}$
test_compile_assign_and_add[tensordict-eager]	3.3595ms	3.1376ms	318.7135 Ops/s	317.2517 Ops/s	$\color{#35bf28}+0.46\%$
test_compile_assign_and_add[pytree-compile]	2.3156ms	0.8071ms	1.2389 KOps/s	1.1363 KOps/s	$\textbf{\color{#35bf28}+9.04\%}$
test_compile_assign_and_add[pytree-eager]	3.3668ms	3.1833ms	314.1382 Ops/s	308.4233 Ops/s	$\color{#35bf28}+1.85\%$
test_compile_indexing[tensor-tensordict-compile]	0.1627ms	0.1124ms	8.8947 KOps/s	9.4667 KOps/s	$\textbf{\color{#d91a1a}-6.04\%}$
test_compile_indexing[tensor-tensordict-eager]	0.2040ms	66.0920μs	15.1304 KOps/s	16.6217 KOps/s	$\textbf{\color{#d91a1a}-8.97\%}$
test_compile_indexing[tensor-tensorclass-compile]	0.1535ms	0.1005ms	9.9527 KOps/s	9.9968 KOps/s	$\color{#d91a1a}-0.44\%$
test_compile_indexing[tensor-tensorclass-eager]	0.1212ms	44.7222μs	22.3602 KOps/s	23.6388 KOps/s	$\textbf{\color{#d91a1a}-5.41\%}$
test_compile_indexing[tensor-pytree-compile]	0.1465ms	0.1072ms	9.3321 KOps/s	9.4625 KOps/s	$\color{#d91a1a}-1.38\%$
test_compile_indexing[tensor-pytree-eager]	93.1620μs	45.3779μs	22.0372 KOps/s	22.1496 KOps/s	$\color{#d91a1a}-0.51\%$
test_compile_indexing[slice-tensordict-compile]	0.2134ms	0.1374ms	7.2798 KOps/s	7.4917 KOps/s	$\color{#d91a1a}-2.83\%$
test_compile_indexing[slice-tensordict-eager]	0.2048ms	25.8398μs	38.7000 KOps/s	40.2366 KOps/s	$\color{#d91a1a}-3.82\%$
test_compile_indexing[slice-tensorclass-compile]	0.1936ms	0.1355ms	7.3815 KOps/s	7.8344 KOps/s	$\textbf{\color{#d91a1a}-5.78\%}$
test_compile_indexing[slice-tensorclass-eager]	65.0010μs	21.9263μs	45.6073 KOps/s	49.5386 KOps/s	$\textbf{\color{#d91a1a}-7.94\%}$
test_compile_indexing[slice-pytree-compile]	0.2656ms	0.1307ms	7.6498 KOps/s	7.8626 KOps/s	$\color{#d91a1a}-2.71\%$
test_compile_indexing[slice-pytree-eager]	61.6500μs	20.7714μs	48.1431 KOps/s	49.3627 KOps/s	$\color{#d91a1a}-2.47\%$
test_compile_indexing[int-tensordict-compile]	0.1968ms	0.1357ms	7.3690 KOps/s	7.4501 KOps/s	$\color{#d91a1a}-1.09\%$
test_compile_indexing[int-tensordict-eager]	0.4767ms	25.2125μs	39.6628 KOps/s	40.0491 KOps/s	$\color{#d91a1a}-0.96\%$
test_compile_indexing[int-tensorclass-compile]	0.2586ms	0.1360ms	7.3519 KOps/s	7.8455 KOps/s	$\textbf{\color{#d91a1a}-6.29\%}$
test_compile_indexing[int-tensorclass-eager]	0.2149ms	27.5826μs	36.2547 KOps/s	50.1291 KOps/s	$\textbf{\color{#d91a1a}-27.68\%}$
test_compile_indexing[int-pytree-compile]	0.1839ms	0.1368ms	7.3087 KOps/s	7.7988 KOps/s	$\textbf{\color{#d91a1a}-6.29\%}$
test_compile_indexing[int-pytree-eager]	55.0310μs	22.0881μs	45.2733 KOps/s	50.7824 KOps/s	$\textbf{\color{#d91a1a}-10.85\%}$
test_mod_add[eager]	67.2910μs	33.0604μs	30.2477 KOps/s	32.3842 KOps/s	$\textbf{\color{#d91a1a}-6.60\%}$
test_mod_add[compile]	0.3469ms	69.3055μs	14.4289 KOps/s	14.6400 KOps/s	$\color{#d91a1a}-1.44\%$
test_mod_add[compile-overhead]	0.2637ms	0.1347ms	7.4265 KOps/s	6.8346 KOps/s	$\textbf{\color{#35bf28}+8.66\%}$
test_mod_wrap[eager]	0.3451ms	0.2356ms	4.2448 KOps/s	4.1297 KOps/s	$\color{#35bf28}+2.79\%$
test_mod_wrap[compile]	0.3576ms	0.2866ms	3.4886 KOps/s	3.4189 KOps/s	$\color{#35bf28}+2.04\%$
test_mod_wrap[compile-overhead]	7.7496ms	4.0984ms	244.0001 Ops/s	246.0708 Ops/s	$\color{#d91a1a}-0.84\%$
test_mod_wrap_and_backward[eager]	1.4370ms	1.3399ms	746.3274 Ops/s	700.4672 Ops/s	$\textbf{\color{#35bf28}+6.55\%}$
test_mod_wrap_and_backward[compile]	1.5196ms	1.2918ms	774.1379 Ops/s	708.7194 Ops/s	$\textbf{\color{#35bf28}+9.23\%}$
test_mod_wrap_and_backward[compile-overhead]	1.3396ms	0.8930ms	1.1198 KOps/s	998.3503 Ops/s	$\textbf{\color{#35bf28}+12.17\%}$
test_seq_add[eager]	0.2208ms	98.0737μs	10.1964 KOps/s	10.2953 KOps/s	$\color{#d91a1a}-0.96\%$
test_seq_add[compile]	0.1812ms	81.4467μs	12.2780 KOps/s	12.5827 KOps/s	$\color{#d91a1a}-2.42\%$
test_seq_add[compile-overhead]	0.1708ms	0.1145ms	8.7346 KOps/s	8.8979 KOps/s	$\color{#d91a1a}-1.84\%$
test_seq_wrap[eager]	0.5127ms	0.3908ms	2.5592 KOps/s	2.5929 KOps/s	$\color{#d91a1a}-1.30\%$
test_seq_wrap[compile]	0.7132ms	0.3049ms	3.2797 KOps/s	3.2296 KOps/s	$\color{#35bf28}+1.55\%$
test_seq_wrap[compile-overhead]	0.2642ms	0.2157ms	4.6364 KOps/s	4.6080 KOps/s	$\color{#35bf28}+0.62\%$
test_func_call_runtime[False-eager]	1.1382ms	0.7265ms	1.3764 KOps/s	1.3753 KOps/s	$\color{#35bf28}+0.08\%$
test_func_call_runtime[False-compile]	1.1739ms	0.7582ms	1.3190 KOps/s	1.3025 KOps/s	$\color{#35bf28}+1.27\%$
test_func_call_runtime[False-compile-overhead]	0.7536ms	0.3552ms	2.8153 KOps/s	2.8393 KOps/s	$\color{#d91a1a}-0.84\%$
test_func_call_runtime[True-eager]	1.3272ms	0.8939ms	1.1187 KOps/s	1.1260 KOps/s	$\color{#d91a1a}-0.65\%$
test_func_call_runtime[True-compile]	1.1878ms	0.7839ms	1.2757 KOps/s	1.2685 KOps/s	$\color{#35bf28}+0.57\%$
test_func_call_runtime[True-compile-overhead]	0.4801ms	0.3760ms	2.6597 KOps/s	2.6663 KOps/s	$\color{#d91a1a}-0.25\%$
test_func_call_cm_runtime[False-eager]	0.8648ms	0.7202ms	1.3885 KOps/s	1.3694 KOps/s	$\color{#35bf28}+1.40\%$
test_func_call_cm_runtime[False-compile]	0.9046ms	0.7613ms	1.3135 KOps/s	1.2941 KOps/s	$\color{#35bf28}+1.50\%$
test_func_call_cm_runtime[False-compile-overhead]	0.4481ms	0.3541ms	2.8243 KOps/s	2.8142 KOps/s	$\color{#35bf28}+0.36\%$
test_func_call_cm_runtime[True-eager]	1.1103ms	0.9794ms	1.0211 KOps/s	1.0053 KOps/s	$\color{#35bf28}+1.57\%$
test_func_call_cm_runtime[True-compile]	1.2193ms	0.8125ms	1.2307 KOps/s	1.2202 KOps/s	$\color{#35bf28}+0.86\%$
test_func_call_cm_runtime[True-compile-overhead]	0.5348ms	0.3978ms	2.5141 KOps/s	2.4910 KOps/s	$\color{#35bf28}+0.93\%$
test_vmap_func_call_cm_runtime[eager]	2.5006ms	2.0566ms	486.2302 Ops/s	484.5449 Ops/s	$\color{#35bf28}+0.35\%$
test_vmap_func_call_cm_runtime[compile]	0.9187ms	0.8241ms	1.2135 KOps/s	1.1996 KOps/s	$\color{#35bf28}+1.16\%$
test_vmap_func_call_cm_runtime[compile-overhead]	0.5394ms	0.4023ms	2.4857 KOps/s	2.4795 KOps/s	$\color{#35bf28}+0.25\%$
test_distributed	3.2547ms	0.1730ms	5.7802 KOps/s	8.9335 KOps/s	$\textbf{\color{#d91a1a}-35.30\%}$
test_tdmodule	24.5300μs	14.2490μs	70.1805 KOps/s	66.9259 KOps/s	$\color{#35bf28}+4.86\%$
test_tdmodule_dispatch	49.3610μs	27.8747μs	35.8749 KOps/s	35.6520 KOps/s	$\color{#35bf28}+0.63\%$
test_tdseq	43.8900μs	15.2888μs	65.4076 KOps/s	65.0597 KOps/s	$\color{#35bf28}+0.53\%$
test_tdseq_dispatch	53.0710μs	30.4529μs	32.8376 KOps/s	32.8196 KOps/s	$\color{#35bf28}+0.05\%$
test_instantiation_functorch	1.9859ms	1.8259ms	547.6637 Ops/s	543.9815 Ops/s	$\color{#35bf28}+0.68\%$
test_instantiation_td	1.8240ms	1.1813ms	846.5333 Ops/s	843.9704 Ops/s	$\color{#35bf28}+0.30\%$
test_exec_functorch	0.2716ms	0.2131ms	4.6929 KOps/s	4.8045 KOps/s	$\color{#d91a1a}-2.32\%$
test_exec_functional_call	0.2947ms	0.2191ms	4.5639 KOps/s	4.8396 KOps/s	$\textbf{\color{#d91a1a}-5.70\%}$
test_exec_td	0.2745ms	0.2282ms	4.3820 KOps/s	4.4992 KOps/s	$\color{#d91a1a}-2.60\%$
test_exec_td_decorator	0.7420ms	0.2690ms	3.7171 KOps/s	3.7538 KOps/s	$\color{#d91a1a}-0.98\%$
test_vmap_mlp_speed[True-True]	0.7893ms	0.6726ms	1.4867 KOps/s	1.4454 KOps/s	$\color{#35bf28}+2.86\%$
test_vmap_mlp_speed[True-False]	0.8065ms	0.6702ms	1.4920 KOps/s	1.4535 KOps/s	$\color{#35bf28}+2.65\%$
test_vmap_mlp_speed[False-True]	0.6852ms	0.5638ms	1.7736 KOps/s	1.6896 KOps/s	$\color{#35bf28}+4.97\%$
test_vmap_mlp_speed[False-False]	0.6811ms	0.5622ms	1.7788 KOps/s	1.6746 KOps/s	$\textbf{\color{#35bf28}+6.22\%}$
test_vmap_mlp_speed_decorator[True-True]	1.1304ms	0.6855ms	1.4588 KOps/s	1.4663 KOps/s	$\color{#d91a1a}-0.51\%$
test_vmap_mlp_speed_decorator[True-False]	0.8242ms	0.6941ms	1.4407 KOps/s	1.4759 KOps/s	$\color{#d91a1a}-2.38\%$
test_vmap_mlp_speed_decorator[False-True]	0.7708ms	0.6081ms	1.6444 KOps/s	1.6413 KOps/s	$\color{#35bf28}+0.19\%$
test_vmap_mlp_speed_decorator[False-False]	0.7134ms	0.6089ms	1.6423 KOps/s	1.6762 KOps/s	$\color{#d91a1a}-2.03\%$
test_vmap_transformer_speed[True-True]	8.3545ms	8.2208ms	121.6427 Ops/s	120.6177 Ops/s	$\color{#35bf28}+0.85\%$
test_vmap_transformer_speed[True-False]	8.5564ms	8.2000ms	121.9518 Ops/s	120.4110 Ops/s	$\color{#35bf28}+1.28\%$
test_vmap_transformer_speed[False-True]	8.1616ms	8.0031ms	124.9516 Ops/s	123.7449 Ops/s	$\color{#35bf28}+0.98\%$
test_vmap_transformer_speed[False-False]	8.1863ms	8.0438ms	124.3200 Ops/s	123.4753 Ops/s	$\color{#35bf28}+0.68\%$
test_vmap_transformer_speed_decorator[True-True]	19.4271ms	19.2599ms	51.9213 Ops/s	51.7660 Ops/s	$\color{#35bf28}+0.30\%$
test_vmap_transformer_speed_decorator[True-False]	20.2922ms	19.2877ms	51.8465 Ops/s	51.7339 Ops/s	$\color{#35bf28}+0.22\%$
test_vmap_transformer_speed_decorator[False-True]	20.0554ms	19.7287ms	50.6874 Ops/s	52.1247 Ops/s	$\color{#d91a1a}-2.76\%$
test_vmap_transformer_speed_decorator[False-False]	20.1971ms	19.4122ms	51.5141 Ops/s	52.1079 Ops/s	$\color{#d91a1a}-1.14\%$
test_to_module_speed[True]	2.0420ms	0.9491ms	1.0536 KOps/s	1.0726 KOps/s	$\color{#d91a1a}-1.77\%$
test_to_module_speed[False]	1.1293ms	0.9207ms	1.0861 KOps/s	1.0841 KOps/s	$\color{#35bf28}+0.19\%$
test_tc_init	61.9210μs	33.3983μs	29.9417 KOps/s	31.5204 KOps/s	$\textbf{\color{#d91a1a}-5.01\%}$
test_tc_init_nested	0.1102ms	66.9339μs	14.9401 KOps/s	15.1638 KOps/s	$\color{#d91a1a}-1.48\%$
test_tc_first_layer_tensor	14.6246μs	0.6621μs	1.5104 MOps/s	1.5164 MOps/s	$\color{#d91a1a}-0.40\%$
test_tc_first_layer_nontensor	25.9300μs	2.2092μs	452.6463 KOps/s	454.0020 KOps/s	$\color{#d91a1a}-0.30\%$
test_tc_second_layer_tensor	10.2275μs	1.3684μs	730.7956 KOps/s	742.5225 KOps/s	$\color{#d91a1a}-1.58\%$
test_tc_second_layer_nontensor	0.1051ms	2.8973μs	345.1443 KOps/s	349.0280 KOps/s	$\color{#d91a1a}-1.11\%$
test_unbind	0.1944s	12.2077ms	81.9153 Ops/s	92.7359 Ops/s	$\textbf{\color{#d91a1a}-11.67\%}$
test_full_like	0.6736ms	0.5761ms	1.7357 KOps/s	1.7400 KOps/s	$\color{#d91a1a}-0.25\%$
test_zeros_like	0.2610ms	0.1979ms	5.0532 KOps/s	5.0519 KOps/s	$\color{#35bf28}+0.03\%$
test_ones_like	0.2403ms	0.1980ms	5.0517 KOps/s	5.0559 KOps/s	$\color{#d91a1a}-0.08\%$
test_clone	0.4445ms	0.4146ms	2.4121 KOps/s	2.4117 KOps/s	$\color{#35bf28}+0.01\%$
test_squeeze	33.3410μs	9.9399μs	100.6051 KOps/s	98.7886 KOps/s	$\color{#35bf28}+1.84\%$
test_unsqueeze	0.2212ms	76.3118μs	13.1041 KOps/s	13.4892 KOps/s	$\color{#d91a1a}-2.85\%$
test_split	0.4340ms	0.1582ms	6.3223 KOps/s	6.3652 KOps/s	$\color{#d91a1a}-0.67\%$
test_permute	0.2225ms	0.1767ms	5.6588 KOps/s	5.6490 KOps/s	$\color{#35bf28}+0.17\%$
test_stack	1.2550ms	0.8700ms	1.1495 KOps/s	1.1713 KOps/s	$\color{#d91a1a}-1.86\%$
test_cat	1.2521ms	1.2320ms	811.6905 Ops/s	811.9882 Ops/s	$\color{#d91a1a}-0.04\%$