[Doc] Streaming tensordicts

[vmoens]

Stack from ghstack (oldest at bottom):

• -> [Doc] Streaming tensordicts #956
• [Feature] Densify lazy tensordicts #955

[github-actions]

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

Total Benchmarks: 219. Improved: $\large\color{#35bf28}6$. Worsened: $\large\color{#d91a1a}38$.

Expand to view detailed results

Name	Max	Mean	Ops	Ops on Repo HEAD	Change
test_plain_set_nested	50.8050μs	22.1213μs	45.2054 KOps/s	46.9291 KOps/s	$\color{#d91a1a}-3.67\%$
test_plain_set_stack_nested	62.6370μs	22.2489μs	44.9461 KOps/s	47.6114 KOps/s	$\textbf{\color{#d91a1a}-5.60\%}$
test_plain_set_nested_inplace	75.1340μs	23.8617μs	41.9082 KOps/s	43.8111 KOps/s	$\color{#d91a1a}-4.34\%$
test_plain_set_stack_nested_inplace	76.0730μs	24.3317μs	41.0986 KOps/s	43.6710 KOps/s	$\textbf{\color{#d91a1a}-5.89\%}$
test_items	20.7090μs	2.7378μs	365.2526 KOps/s	379.3574 KOps/s	$\color{#d91a1a}-3.72\%$
test_items_nested	2.2305ms	0.3484ms	2.8699 KOps/s	3.0338 KOps/s	$\textbf{\color{#d91a1a}-5.40\%}$
test_items_nested_locked	0.6186ms	0.3504ms	2.8538 KOps/s	2.9996 KOps/s	$\color{#d91a1a}-4.86\%$
test_items_nested_leaf	0.1530ms	82.3176μs	12.1481 KOps/s	12.2048 KOps/s	$\color{#d91a1a}-0.46\%$
test_items_stack_nested	0.5789ms	0.3407ms	2.9349 KOps/s	2.9997 KOps/s	$\color{#d91a1a}-2.16\%$
test_items_stack_nested_leaf	0.1520ms	82.4611μs	12.1269 KOps/s	12.7122 KOps/s	$\color{#d91a1a}-4.60\%$
test_items_stack_nested_locked	0.5843ms	0.3404ms	2.9376 KOps/s	2.9455 KOps/s	$\color{#d91a1a}-0.27\%$
test_keys	23.0630μs	3.8171μs	261.9821 KOps/s	261.6522 KOps/s	$\color{#35bf28}+0.13\%$
test_keys_nested	0.2947ms	0.1401ms	7.1362 KOps/s	6.9494 KOps/s	$\color{#35bf28}+2.69\%$
test_keys_nested_locked	0.6989ms	0.1472ms	6.7927 KOps/s	6.6931 KOps/s	$\color{#35bf28}+1.49\%$
test_keys_nested_leaf	0.2426ms	0.1225ms	8.1605 KOps/s	8.1625 KOps/s	$\color{#d91a1a}-0.02\%$
test_keys_stack_nested	0.2539ms	0.1431ms	6.9874 KOps/s	7.1098 KOps/s	$\color{#d91a1a}-1.72\%$
test_keys_stack_nested_leaf	0.2249ms	0.1211ms	8.2581 KOps/s	8.2695 KOps/s	$\color{#d91a1a}-0.14\%$
test_keys_stack_nested_locked	0.2447ms	0.1496ms	6.6863 KOps/s	6.7931 KOps/s	$\color{#d91a1a}-1.57\%$
test_values	13.5205μs	2.2475μs	444.9459 KOps/s	851.4352 KOps/s	$\textbf{\color{#d91a1a}-47.74\%}$
test_values_nested	89.0170μs	48.6968μs	20.5352 KOps/s	20.3904 KOps/s	$\color{#35bf28}+0.71\%$
test_values_nested_locked	0.1064ms	48.9223μs	20.4406 KOps/s	20.4614 KOps/s	$\color{#d91a1a}-0.10\%$
test_values_nested_leaf	95.1990μs	44.5201μs	22.4618 KOps/s	22.3074 KOps/s	$\color{#35bf28}+0.69\%$
test_values_stack_nested	99.3060μs	49.6965μs	20.1222 KOps/s	19.6784 KOps/s	$\color{#35bf28}+2.25\%$
test_values_stack_nested_leaf	88.0250μs	43.9989μs	22.7278 KOps/s	23.1266 KOps/s	$\color{#d91a1a}-1.72\%$
test_values_stack_nested_locked	90.7200μs	49.3471μs	20.2646 KOps/s	19.7143 KOps/s	$\color{#35bf28}+2.79\%$
test_membership	2.5493μs	0.7238μs	1.3816 MOps/s	1.3453 MOps/s	$\color{#35bf28}+2.70\%$
test_membership_nested	30.4470μs	2.5934μs	385.5886 KOps/s	378.1200 KOps/s	$\color{#35bf28}+1.98\%$
test_membership_nested_leaf	42.2490μs	2.6357μs	379.4052 KOps/s	389.1424 KOps/s	$\color{#d91a1a}-2.50\%$
test_membership_stacked_nested	29.3860μs	2.6155μs	382.3300 KOps/s	401.7692 KOps/s	$\color{#d91a1a}-4.84\%$
test_membership_stacked_nested_leaf	25.8380μs	2.6390μs	378.9258 KOps/s	392.3195 KOps/s	$\color{#d91a1a}-3.41\%$
test_membership_nested_last	28.3330μs	3.9099μs	255.7634 KOps/s	263.5290 KOps/s	$\color{#d91a1a}-2.95\%$
test_membership_nested_leaf_last	30.4670μs	3.9042μs	256.1333 KOps/s	254.4892 KOps/s	$\color{#35bf28}+0.65\%$
test_membership_stacked_nested_last	24.4450μs	3.9255μs	254.7453 KOps/s	78.6358 KOps/s	$\textbf{\color{#35bf28}+223.96\%}$
test_membership_stacked_nested_leaf_last	43.0300μs	3.8070μs	262.6746 KOps/s	80.6033 KOps/s	$\textbf{\color{#35bf28}+225.89\%}$
test_nested_getleaf	41.2470μs	10.4037μs	96.1198 KOps/s	96.1789 KOps/s	$\color{#d91a1a}-0.06\%$
test_nested_get	50.8450μs	9.7485μs	102.5799 KOps/s	104.7909 KOps/s	$\color{#d91a1a}-2.11\%$
test_stacked_getleaf	39.5270μs	10.3153μs	96.9433 KOps/s	95.5553 KOps/s	$\color{#35bf28}+1.45\%$
test_stacked_get	32.7310μs	9.6172μs	103.9802 KOps/s	102.9688 KOps/s	$\color{#35bf28}+0.98\%$
test_nested_getitemleaf	52.1880μs	10.9828μs	91.0511 KOps/s	94.8358 KOps/s	$\color{#d91a1a}-3.99\%$
test_nested_getitem	53.4600μs	10.0788μs	99.2182 KOps/s	102.4214 KOps/s	$\color{#d91a1a}-3.13\%$
test_stacked_getitemleaf	44.2930μs	11.0666μs	90.3618 KOps/s	94.3194 KOps/s	$\color{#d91a1a}-4.20\%$
test_stacked_getitem	30.9990μs	10.0009μs	99.9913 KOps/s	100.8612 KOps/s	$\color{#d91a1a}-0.86\%$
test_lock_nested	80.1372ms	0.5698ms	1.7551 KOps/s	2.0350 KOps/s	$\textbf{\color{#d91a1a}-13.75\%}$
test_lock_stack_nested	0.7302ms	0.4648ms	2.1515 KOps/s	2.3230 KOps/s	$\textbf{\color{#d91a1a}-7.38\%}$
test_unlock_nested	80.1133ms	0.4919ms	2.0329 KOps/s	2.4290 KOps/s	$\textbf{\color{#d91a1a}-16.31\%}$
test_unlock_stack_nested	0.6147ms	0.3759ms	2.6599 KOps/s	2.7956 KOps/s	$\color{#d91a1a}-4.85\%$
test_flatten_speed	0.1961ms	0.1015ms	9.8502 KOps/s	9.9062 KOps/s	$\color{#d91a1a}-0.57\%$
test_unflatten_speed	0.8908ms	0.4550ms	2.1977 KOps/s	2.2592 KOps/s	$\color{#d91a1a}-2.72\%$
test_common_ops	3.8549ms	1.1478ms	871.2471 Ops/s	930.9555 Ops/s	$\textbf{\color{#d91a1a}-6.41\%}$
test_creation	21.0190μs	2.0107μs	497.3489 KOps/s	487.4781 KOps/s	$\color{#35bf28}+2.02\%$
test_creation_empty	47.0080μs	18.0837μs	55.2983 KOps/s	57.3083 KOps/s	$\color{#d91a1a}-3.51\%$
test_creation_nested_1	79.4090μs	21.6670μs	46.1532 KOps/s	48.0131 KOps/s	$\color{#d91a1a}-3.87\%$
test_creation_nested_2	60.8140μs	25.8720μs	38.6518 KOps/s	40.0496 KOps/s	$\color{#d91a1a}-3.49\%$
test_clone	63.7690μs	16.6489μs	60.0641 KOps/s	58.9920 KOps/s	$\color{#35bf28}+1.82\%$
test_getitem[int]	1.1166ms	16.5893μs	60.2796 KOps/s	62.3522 KOps/s	$\color{#d91a1a}-3.32\%$
test_getitem[slice_int]	0.1382ms	32.1640μs	31.0907 KOps/s	31.9540 KOps/s	$\color{#d91a1a}-2.70\%$
test_getitem[range]	0.1904ms	58.3519μs	17.1374 KOps/s	17.4949 KOps/s	$\color{#d91a1a}-2.04\%$
test_getitem[tuple]	0.1242ms	26.6571μs	37.5134 KOps/s	39.7972 KOps/s	$\textbf{\color{#d91a1a}-5.74\%}$
test_getitem[list]	0.1808ms	53.2215μs	18.7894 KOps/s	19.3687 KOps/s	$\color{#d91a1a}-2.99\%$
test_setitem_dim[int]	64.1800μs	41.5662μs	24.0580 KOps/s	23.8293 KOps/s	$\color{#35bf28}+0.96\%$
test_setitem_dim[slice_int]	0.1166ms	72.0086μs	13.8872 KOps/s	14.2225 KOps/s	$\color{#d91a1a}-2.36\%$
test_setitem_dim[range]	0.1613ms	94.4621μs	10.5863 KOps/s	10.7491 KOps/s	$\color{#d91a1a}-1.51\%$
test_setitem_dim[tuple]	0.1319ms	63.3339μs	15.7893 KOps/s	16.9953 KOps/s	$\textbf{\color{#d91a1a}-7.10\%}$
test_setitem	88.5460μs	30.2292μs	33.0806 KOps/s	34.3680 KOps/s	$\color{#d91a1a}-3.75\%$
test_set	0.1194ms	29.3009μs	34.1286 KOps/s	35.7499 KOps/s	$\color{#d91a1a}-4.54\%$
test_set_shared	1.2410ms	0.2108ms	4.7437 KOps/s	4.6768 KOps/s	$\color{#35bf28}+1.43\%$
test_update	0.1368ms	38.0333μs	26.2927 KOps/s	27.6364 KOps/s	$\color{#d91a1a}-4.86\%$
test_update_nested	2.4328ms	49.1819μs	20.3327 KOps/s	22.3683 KOps/s	$\textbf{\color{#d91a1a}-9.10\%}$
test_update__nested	0.1367ms	35.9852μs	27.7892 KOps/s	30.1758 KOps/s	$\textbf{\color{#d91a1a}-7.91\%}$
test_set_nested	0.1062ms	33.6566μs	29.7119 KOps/s	32.1969 KOps/s	$\textbf{\color{#d91a1a}-7.72\%}$
test_set_nested_new	0.1072ms	37.8656μs	26.4092 KOps/s	28.4151 KOps/s	$\textbf{\color{#d91a1a}-7.06\%}$
test_select	0.1559ms	53.7072μs	18.6195 KOps/s	19.6196 KOps/s	$\textbf{\color{#d91a1a}-5.10\%}$
test_select_nested	0.1388ms	57.6372μs	17.3499 KOps/s	17.5861 KOps/s	$\color{#d91a1a}-1.34\%$
test_exclude_nested	0.1666ms	75.4828μs	13.2481 KOps/s	13.4105 KOps/s	$\color{#d91a1a}-1.21\%$
test_empty[True]	0.4533ms	0.3222ms	3.1033 KOps/s	3.2109 KOps/s	$\color{#d91a1a}-3.35\%$
test_empty[False]	5.8488μs	1.1318μs	883.5292 KOps/s	888.1983 KOps/s	$\color{#d91a1a}-0.53\%$
test_unbind_speed	0.3562ms	0.2889ms	3.4618 KOps/s	3.2825 KOps/s	$\textbf{\color{#35bf28}+5.46\%}$
test_unbind_speed_stack0	0.6814ms	0.2956ms	3.3831 KOps/s	3.5595 KOps/s	$\color{#d91a1a}-4.95\%$
test_unbind_speed_stack1	83.0676ms	0.7535ms	1.3271 KOps/s	1.4793 KOps/s	$\textbf{\color{#d91a1a}-10.29\%}$
test_split	81.8858ms	2.0206ms	494.8983 Ops/s	440.3197 Ops/s	$\textbf{\color{#35bf28}+12.40\%}$
test_chunk	83.8380ms	2.0211ms	494.7700 Ops/s	512.7088 Ops/s	$\color{#d91a1a}-3.50\%$
test_creation[device0]	4.2665ms	0.1177ms	8.4954 KOps/s	8.2454 KOps/s	$\color{#35bf28}+3.03\%$
test_creation_from_tensor	0.2283ms	0.1219ms	8.2051 KOps/s	8.6709 KOps/s	$\textbf{\color{#d91a1a}-5.37\%}$
test_add_one[memmap_tensor0]	0.2274ms	7.7485μs	129.0572 KOps/s	129.6854 KOps/s	$\color{#d91a1a}-0.48\%$
test_contiguous[memmap_tensor0]	29.6660μs	2.0095μs	497.6379 KOps/s	501.0831 KOps/s	$\color{#d91a1a}-0.69\%$
test_stack[memmap_tensor0]	56.7360μs	5.7693μs	173.3322 KOps/s	179.4097 KOps/s	$\color{#d91a1a}-3.39\%$
test_memmaptd_index	1.0497ms	0.3944ms	2.5352 KOps/s	2.5582 KOps/s	$\color{#d91a1a}-0.90\%$
test_memmaptd_index_astensor	0.9891ms	0.4763ms	2.0996 KOps/s	2.1319 KOps/s	$\color{#d91a1a}-1.52\%$
test_memmaptd_index_op	1.3412ms	1.0304ms	970.4619 Ops/s	986.1767 Ops/s	$\color{#d91a1a}-1.59\%$
test_serialize_model	0.1227s	0.1154s	8.6689 Ops/s	8.8360 Ops/s	$\color{#d91a1a}-1.89\%$
test_serialize_model_pickle	0.4691s	0.3933s	2.5427 Ops/s	2.5751 Ops/s	$\color{#d91a1a}-1.26\%$
test_serialize_weights	0.1255s	0.1148s	8.7130 Ops/s	8.6048 Ops/s	$\color{#35bf28}+1.26\%$
test_serialize_weights_returnearly	0.1842s	0.1607s	6.2234 Ops/s	6.6857 Ops/s	$\textbf{\color{#d91a1a}-6.91\%}$
test_serialize_weights_pickle	0.4935s	0.4097s	2.4409 Ops/s	2.4326 Ops/s	$\color{#35bf28}+0.34\%$
test_serialize_weights_filesystem	0.2159s	0.1472s	6.7943 Ops/s	6.6155 Ops/s	$\color{#35bf28}+2.70\%$
test_serialize_model_filesystem	0.1626s	0.1488s	6.7211 Ops/s	6.5693 Ops/s	$\color{#35bf28}+2.31\%$
test_reshape_pytree	99.7560μs	39.9408μs	25.0371 KOps/s	25.7887 KOps/s	$\color{#d91a1a}-2.91\%$
test_reshape_td	0.1491ms	46.1178μs	21.6836 KOps/s	22.6204 KOps/s	$\color{#d91a1a}-4.14\%$
test_view_pytree	85.7200μs	37.3977μs	26.7396 KOps/s	25.8662 KOps/s	$\color{#35bf28}+3.38\%$
test_view_td	0.1178ms	52.8992μs	18.9039 KOps/s	19.4729 KOps/s	$\color{#d91a1a}-2.92\%$
test_unbind_pytree	71.7140μs	36.5132μs	27.3874 KOps/s	26.4886 KOps/s	$\color{#35bf28}+3.39\%$
test_unbind_td	0.3475ms	45.7989μs	21.8346 KOps/s	22.0983 KOps/s	$\color{#d91a1a}-1.19\%$
test_split_pytree	0.1266ms	40.7171μs	24.5597 KOps/s	25.0395 KOps/s	$\color{#d91a1a}-1.92\%$
test_split_td	0.4835ms	57.3181μs	17.4465 KOps/s	17.4714 KOps/s	$\color{#d91a1a}-0.14\%$
test_add_pytree	0.1327ms	46.3715μs	21.5650 KOps/s	20.9437 KOps/s	$\color{#35bf28}+2.97\%$
test_add_td	0.2147ms	82.9381μs	12.0572 KOps/s	12.0761 KOps/s	$\color{#d91a1a}-0.16\%$
test_compile_add_one_nested[tensordict-compile]	0.1247ms	55.0544μs	18.1638 KOps/s	19.2303 KOps/s	$\textbf{\color{#d91a1a}-5.55\%}$
test_compile_add_one_nested[tensordict-eager]	0.4262ms	0.1871ms	5.3437 KOps/s	5.2553 KOps/s	$\color{#35bf28}+1.68\%$
test_compile_add_one_nested[pytree-compile]	0.1672ms	55.6856μs	17.9580 KOps/s	19.3185 KOps/s	$\textbf{\color{#d91a1a}-7.04\%}$
test_compile_add_one_nested[pytree-eager]	0.2574ms	0.1452ms	6.8856 KOps/s	6.9254 KOps/s	$\color{#d91a1a}-0.57\%$
test_compile_copy_nested[tensordict-compile]	92.6930μs	20.0171μs	49.9573 KOps/s	51.1677 KOps/s	$\color{#d91a1a}-2.37\%$
test_compile_copy_nested[tensordict-eager]	0.1392ms	62.9847μs	15.8769 KOps/s	15.6960 KOps/s	$\color{#35bf28}+1.15\%$
test_compile_copy_nested[pytree-compile]	0.1639ms	78.3114μs	12.7695 KOps/s	12.5430 KOps/s	$\color{#35bf28}+1.81\%$
test_compile_copy_nested[pytree-eager]	0.1568ms	70.1176μs	14.2617 KOps/s	14.3399 KOps/s	$\color{#d91a1a}-0.54\%$
test_compile_add_one_flat[tensordict-compile]	0.4073ms	0.1735ms	5.7636 KOps/s	5.9750 KOps/s	$\color{#d91a1a}-3.54\%$
test_compile_add_one_flat[tensordict-eager]	0.3427ms	0.1876ms	5.3302 KOps/s	5.2359 KOps/s	$\color{#35bf28}+1.80\%$
test_compile_add_one_flat[tensorclass-compile]	79.7790μs	38.3436μs	26.0800 KOps/s	26.0357 KOps/s	$\color{#35bf28}+0.17\%$
test_compile_add_one_flat[tensorclass-eager]	0.8747ms	71.5792μs	13.9705 KOps/s	14.2801 KOps/s	$\color{#d91a1a}-2.17\%$
test_compile_add_one_flat[pytree-compile]	0.4451ms	0.1708ms	5.8556 KOps/s	5.7002 KOps/s	$\color{#35bf28}+2.73\%$
test_compile_add_one_flat[pytree-eager]	0.5843ms	0.2873ms	3.4802 KOps/s	3.3867 KOps/s	$\color{#35bf28}+2.76\%$
test_compile_add_self_flat[tensordict-eager]	0.3513ms	0.2047ms	4.8851 KOps/s	4.9339 KOps/s	$\color{#d91a1a}-0.99\%$
test_compile_add_self_flat[tensordict-compile]	0.5942ms	0.1849ms	5.4091 KOps/s	5.7868 KOps/s	$\textbf{\color{#d91a1a}-6.53\%}$
test_compile_add_self_flat[tensorclass-eager]	0.7223ms	61.9265μs	16.1482 KOps/s	15.4179 KOps/s	$\color{#35bf28}+4.74\%$
test_compile_add_self_flat[tensorclass-compile]	0.1222ms	40.4746μs	24.7068 KOps/s	26.1352 KOps/s	$\textbf{\color{#d91a1a}-5.47\%}$
test_compile_add_self_flat[pytree-eager]	0.4585ms	0.2387ms	4.1892 KOps/s	4.1298 KOps/s	$\color{#35bf28}+1.44\%$
test_compile_add_self_flat[pytree-compile]	0.5730ms	0.1784ms	5.6040 KOps/s	5.9170 KOps/s	$\textbf{\color{#d91a1a}-5.29\%}$
test_compile_copy_flat[tensordict-compile]	0.1890ms	0.1092ms	9.1544 KOps/s	9.3946 KOps/s	$\color{#d91a1a}-2.56\%$
test_compile_copy_flat[tensordict-eager]	0.1187ms	55.5192μs	18.0118 KOps/s	18.2197 KOps/s	$\color{#d91a1a}-1.14\%$
test_compile_copy_flat[pytree-compile]	0.1923ms	79.5288μs	12.5741 KOps/s	12.4543 KOps/s	$\color{#35bf28}+0.96\%$
test_compile_copy_flat[pytree-eager]	0.1596ms	70.5934μs	14.1656 KOps/s	14.1226 KOps/s	$\color{#35bf28}+0.30\%$
test_compile_assign_and_add[tensordict-compile]	0.3744ms	0.1892ms	5.2863 KOps/s	5.6742 KOps/s	$\textbf{\color{#d91a1a}-6.84\%}$
test_compile_assign_and_add[tensordict-eager]	1.7978ms	1.5941ms	627.3206 Ops/s	629.4596 Ops/s	$\color{#d91a1a}-0.34\%$
test_compile_assign_and_add[pytree-compile]	0.2589ms	0.1841ms	5.4322 KOps/s	5.4736 KOps/s	$\color{#d91a1a}-0.76\%$
test_compile_assign_and_add[pytree-eager]	1.2500ms	1.0946ms	913.5807 Ops/s	906.4682 Ops/s	$\color{#35bf28}+0.78\%$
test_compile_assign_and_add_stack[compile]	0.5032ms	0.4024ms	2.4849 KOps/s	2.4481 KOps/s	$\color{#35bf28}+1.50\%$
test_compile_assign_and_add_stack[eager]	6.0025ms	3.8352ms	260.7445 Ops/s	261.5586 Ops/s	$\color{#d91a1a}-0.31\%$
test_compile_indexing[tensor-tensordict-compile]	0.1027ms	34.5615μs	28.9339 KOps/s	31.1249 KOps/s	$\textbf{\color{#d91a1a}-7.04\%}$
test_compile_indexing[tensor-tensordict-eager]	0.6876ms	49.0686μs	20.3796 KOps/s	20.7403 KOps/s	$\color{#d91a1a}-1.74\%$
test_compile_indexing[tensor-tensorclass-compile]	0.1014ms	29.6494μs	33.7275 KOps/s	35.5073 KOps/s	$\textbf{\color{#d91a1a}-5.01\%}$
test_compile_indexing[tensor-tensorclass-eager]	90.9610μs	30.6873μs	32.5868 KOps/s	32.2175 KOps/s	$\color{#35bf28}+1.15\%$
test_compile_indexing[tensor-pytree-compile]	89.4570μs	29.1739μs	34.2772 KOps/s	36.1667 KOps/s	$\textbf{\color{#d91a1a}-5.22\%}$
test_compile_indexing[tensor-pytree-eager]	74.3390μs	30.2095μs	33.1022 KOps/s	32.0791 KOps/s	$\color{#35bf28}+3.19\%$
test_compile_indexing[slice-tensordict-compile]	0.1493ms	74.1263μs	13.4905 KOps/s	13.9865 KOps/s	$\color{#d91a1a}-3.55\%$
test_compile_indexing[slice-tensordict-eager]	0.4810ms	27.5960μs	36.2371 KOps/s	35.6280 KOps/s	$\color{#35bf28}+1.71\%$
test_compile_indexing[slice-tensorclass-compile]	0.1362ms	67.5965μs	14.7937 KOps/s	14.9875 KOps/s	$\color{#d91a1a}-1.29\%$
test_compile_indexing[slice-tensorclass-eager]	80.9320μs	24.6755μs	40.5260 KOps/s	40.0183 KOps/s	$\color{#35bf28}+1.27\%$
test_compile_indexing[slice-pytree-compile]	0.1366ms	68.7803μs	14.5390 KOps/s	14.8858 KOps/s	$\color{#d91a1a}-2.33\%$
test_compile_indexing[slice-pytree-eager]	87.4100μs	23.7950μs	42.0256 KOps/s	40.1481 KOps/s	$\color{#35bf28}+4.68\%$
test_compile_indexing[int-tensordict-compile]	0.1630ms	74.5150μs	13.4201 KOps/s	14.1689 KOps/s	$\textbf{\color{#d91a1a}-5.28\%}$
test_compile_indexing[int-tensordict-eager]	0.8566ms	28.0015μs	35.7123 KOps/s	35.5151 KOps/s	$\color{#35bf28}+0.56\%$
test_compile_indexing[int-tensorclass-compile]	0.1373ms	66.6155μs	15.0115 KOps/s	14.9792 KOps/s	$\color{#35bf28}+0.22\%$
test_compile_indexing[int-tensorclass-eager]	76.8640μs	23.9733μs	41.7130 KOps/s	39.7755 KOps/s	$\color{#35bf28}+4.87\%$
test_compile_indexing[int-pytree-compile]	0.1410ms	68.1062μs	14.6830 KOps/s	14.7989 KOps/s	$\color{#d91a1a}-0.78\%$
test_compile_indexing[int-pytree-eager]	96.7370μs	23.9240μs	41.7991 KOps/s	40.1230 KOps/s	$\color{#35bf28}+4.18\%$
test_mod_add[eager]	79.3420μs	25.9158μs	38.5865 KOps/s	40.4335 KOps/s	$\color{#d91a1a}-4.57\%$
test_mod_add[compile]	0.1008ms	38.1308μs	26.2255 KOps/s	27.9114 KOps/s	$\textbf{\color{#d91a1a}-6.04\%}$
test_mod_add[compile-overhead]	0.1206ms	38.0815μs	26.2594 KOps/s	27.9664 KOps/s	$\textbf{\color{#d91a1a}-6.10\%}$
test_mod_wrap[eager]	0.4360ms	0.2160ms	4.6297 KOps/s	4.7913 KOps/s	$\color{#d91a1a}-3.37\%$
test_mod_wrap[compile]	1.7432ms	0.2336ms	4.2808 KOps/s	4.2914 KOps/s	$\color{#d91a1a}-0.25\%$
test_mod_wrap[compile-overhead]	0.4201ms	0.2265ms	4.4151 KOps/s	4.4128 KOps/s	$\color{#35bf28}+0.05\%$
test_mod_wrap_and_backward[eager]	17.3096ms	12.7882ms	78.1971 Ops/s	92.2091 Ops/s	$\textbf{\color{#d91a1a}-15.20\%}$
test_mod_wrap_and_backward[compile]	14.7260ms	11.5685ms	86.4416 Ops/s	85.7060 Ops/s	$\color{#35bf28}+0.86\%$
test_mod_wrap_and_backward[compile-overhead]	18.0968ms	11.7087ms	85.4063 Ops/s	85.4381 Ops/s	$\color{#d91a1a}-0.04\%$
test_seq_add[eager]	0.1774ms	91.0680μs	10.9808 KOps/s	11.7742 KOps/s	$\textbf{\color{#d91a1a}-6.74\%}$
test_seq_add[compile]	0.1596ms	62.4547μs	16.0116 KOps/s	16.2754 KOps/s	$\color{#d91a1a}-1.62\%$
test_seq_add[compile-overhead]	0.1544ms	61.9439μs	16.1436 KOps/s	16.8881 KOps/s	$\color{#d91a1a}-4.41\%$
test_seq_wrap[eager]	0.4958ms	0.3825ms	2.6146 KOps/s	2.6358 KOps/s	$\color{#d91a1a}-0.80\%$
test_seq_wrap[compile]	0.4023ms	0.2639ms	3.7895 KOps/s	3.7617 KOps/s	$\color{#35bf28}+0.74\%$
test_seq_wrap[compile-overhead]	0.6233ms	0.2623ms	3.8129 KOps/s	3.7737 KOps/s	$\color{#35bf28}+1.04\%$
test_func_call_runtime[False-eager]	0.9264ms	0.5242ms	1.9076 KOps/s	1.9085 KOps/s	$\color{#d91a1a}-0.05\%$
test_func_call_runtime[False-compile]	0.8347ms	0.4969ms	2.0124 KOps/s	1.9780 KOps/s	$\color{#35bf28}+1.74\%$
test_func_call_runtime[False-compile-overhead]	0.6087ms	0.4902ms	2.0401 KOps/s	2.0360 KOps/s	$\color{#35bf28}+0.20\%$
test_func_call_runtime[True-eager]	1.2069ms	0.7424ms	1.3471 KOps/s	1.3162 KOps/s	$\color{#35bf28}+2.35\%$
test_func_call_runtime[True-compile]	0.6987ms	0.5031ms	1.9877 KOps/s	1.9695 KOps/s	$\color{#35bf28}+0.92\%$
test_func_call_runtime[True-compile-overhead]	1.0138ms	0.5137ms	1.9465 KOps/s	1.9544 KOps/s	$\color{#d91a1a}-0.40\%$
test_func_call_cm_runtime[False-eager]	0.8895ms	0.5367ms	1.8633 KOps/s	1.9010 KOps/s	$\color{#d91a1a}-1.98\%$
test_func_call_cm_runtime[False-compile]	0.8583ms	0.4966ms	2.0137 KOps/s	2.0222 KOps/s	$\color{#d91a1a}-0.42\%$
test_func_call_cm_runtime[False-compile-overhead]	0.9194ms	0.5124ms	1.9517 KOps/s	2.0070 KOps/s	$\color{#d91a1a}-2.75\%$
test_func_call_cm_runtime[True-eager]	1.0309ms	0.8888ms	1.1251 KOps/s	1.1357 KOps/s	$\color{#d91a1a}-0.93\%$
test_func_call_cm_runtime[True-compile]	0.9591ms	0.8364ms	1.1956 KOps/s	1.1889 KOps/s	$\color{#35bf28}+0.57\%$
test_func_call_cm_runtime[True-compile-overhead]	1.4083ms	0.8485ms	1.1785 KOps/s	1.2178 KOps/s	$\color{#d91a1a}-3.23\%$
test_distributed	0.2309ms	0.1291ms	7.7470 KOps/s	7.7164 KOps/s	$\color{#35bf28}+0.40\%$
test_tdmodule	0.1118ms	18.7839μs	53.2371 KOps/s	58.0495 KOps/s	$\textbf{\color{#d91a1a}-8.29\%}$
test_tdmodule_dispatch	63.7500μs	36.9392μs	27.0715 KOps/s	27.9120 KOps/s	$\color{#d91a1a}-3.01\%$
test_tdseq	38.7530μs	19.3335μs	51.7237 KOps/s	55.6705 KOps/s	$\textbf{\color{#d91a1a}-7.09\%}$
test_tdseq_dispatch	72.9370μs	41.1859μs	24.2802 KOps/s	25.9842 KOps/s	$\textbf{\color{#d91a1a}-6.56\%}$
test_instantiation_functorch	2.5988ms	1.6440ms	608.2579 Ops/s	610.4911 Ops/s	$\color{#d91a1a}-0.37\%$
test_instantiation_td	1.8002ms	1.1711ms	853.8966 Ops/s	867.0962 Ops/s	$\color{#d91a1a}-1.52\%$
test_exec_functorch	0.3081ms	0.1834ms	5.4537 KOps/s	5.5605 KOps/s	$\color{#d91a1a}-1.92\%$
test_exec_functional_call	0.4337ms	0.1769ms	5.6525 KOps/s	5.9190 KOps/s	$\color{#d91a1a}-4.50\%$
test_exec_td	0.2740ms	0.1783ms	5.6072 KOps/s	6.1399 KOps/s	$\textbf{\color{#d91a1a}-8.68\%}$
test_exec_td_decorator	0.4364ms	0.2352ms	4.2523 KOps/s	4.5975 KOps/s	$\textbf{\color{#d91a1a}-7.51\%}$
test_vmap_mlp_speed[True-True]	0.9979ms	0.5837ms	1.7132 KOps/s	1.7217 KOps/s	$\color{#d91a1a}-0.49\%$
test_vmap_mlp_speed[True-False]	0.8548ms	0.5859ms	1.7067 KOps/s	1.7744 KOps/s	$\color{#d91a1a}-3.82\%$
test_vmap_mlp_speed[False-True]	0.5987ms	0.4809ms	2.0795 KOps/s	2.0966 KOps/s	$\color{#d91a1a}-0.82\%$
test_vmap_mlp_speed[False-False]	0.6836ms	0.4761ms	2.1006 KOps/s	2.1051 KOps/s	$\color{#d91a1a}-0.22\%$
test_vmap_mlp_speed_decorator[True-True]	1.4037ms	0.6276ms	1.5934 KOps/s	1.5680 KOps/s	$\color{#35bf28}+1.62\%$
test_vmap_mlp_speed_decorator[True-False]	0.8114ms	0.6323ms	1.5814 KOps/s	1.5815 KOps/s	$-0.00\%$
test_vmap_mlp_speed_decorator[False-True]	0.7655ms	0.5099ms	1.9612 KOps/s	1.9208 KOps/s	$\color{#35bf28}+2.10\%$
test_vmap_mlp_speed_decorator[False-False]	0.8735ms	0.5230ms	1.9120 KOps/s	1.9137 KOps/s	$\color{#d91a1a}-0.09\%$
test_to_module_speed[True]	2.1547ms	1.3257ms	754.3450 Ops/s	762.3985 Ops/s	$\color{#d91a1a}-1.06\%$
test_to_module_speed[False]	1.9147ms	1.2922ms	773.8783 Ops/s	803.3222 Ops/s	$\color{#d91a1a}-3.67\%$
test_tc_init	75.0100μs	43.7404μs	22.8622 KOps/s	24.0358 KOps/s	$\color{#d91a1a}-4.88\%$
test_tc_init_nested	0.1526ms	81.6331μs	12.2499 KOps/s	12.2176 KOps/s	$\color{#35bf28}+0.26\%$
test_tc_first_layer_tensor	22.8130μs	1.4896μs	671.3391 KOps/s	681.7609 KOps/s	$\color{#d91a1a}-1.53\%$
test_tc_first_layer_nontensor	31.2080μs	4.3057μs	232.2523 KOps/s	241.0800 KOps/s	$\color{#d91a1a}-3.66\%$
test_tc_second_layer_tensor	39.9240μs	2.6911μs	371.5917 KOps/s	376.6347 KOps/s	$\color{#d91a1a}-1.34\%$
test_tc_second_layer_nontensor	23.8040μs	5.3613μs	186.5202 KOps/s	189.8637 KOps/s	$\color{#d91a1a}-1.76\%$
test_unbind	0.4499s	13.5578ms	73.7582 Ops/s	77.8211 Ops/s	$\textbf{\color{#d91a1a}-5.22\%}$
test_full_like	9.1547ms	7.1727ms	139.4179 Ops/s	141.2983 Ops/s	$\color{#d91a1a}-1.33\%$
test_zeros_like	10.5868ms	6.4270ms	155.5925 Ops/s	134.0806 Ops/s	$\textbf{\color{#35bf28}+16.04\%}$
test_ones_like	13.0813ms	7.6339ms	130.9942 Ops/s	123.6974 Ops/s	$\textbf{\color{#35bf28}+5.90\%}$
test_clone	15.4455ms	9.0470ms	110.5342 Ops/s	110.4748 Ops/s	$\color{#35bf28}+0.05\%$
test_squeeze	73.1260μs	13.3245μs	75.0500 KOps/s	77.5264 KOps/s	$\color{#d91a1a}-3.19\%$
test_unsqueeze	0.1682ms	93.5248μs	10.6924 KOps/s	10.8410 KOps/s	$\color{#d91a1a}-1.37\%$
test_split	0.4576ms	0.1997ms	5.0085 KOps/s	5.0036 KOps/s	$\color{#35bf28}+0.10\%$
test_permute	0.4866ms	0.2221ms	4.5027 KOps/s	4.5921 KOps/s	$\color{#d91a1a}-1.95\%$
test_stack	29.4656ms	24.0781ms	41.5315 Ops/s	40.7821 Ops/s	$\color{#35bf28}+1.84\%$
test_cat	32.6967ms	24.1859ms	41.3464 Ops/s	40.8805 Ops/s	$\color{#35bf28}+1.14\%$

[github-actions]

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

Total Benchmarks: 225. Improved: $\large\color{#35bf28}47$. Worsened: $\large\color{#d91a1a}4$.

Expand to view detailed results

Name	Max	Mean	Ops	Ops on Repo HEAD	Change
test_plain_set_nested	88.7720μs	16.4811μs	60.6757 KOps/s	57.4391 KOps/s	$\textbf{\color{#35bf28}+5.63\%}$
test_plain_set_stack_nested	42.1510μs	16.5407μs	60.4570 KOps/s	57.0619 KOps/s	$\textbf{\color{#35bf28}+5.95\%}$
test_plain_set_nested_inplace	36.5210μs	17.5936μs	56.8390 KOps/s	53.9215 KOps/s	$\textbf{\color{#35bf28}+5.41\%}$
test_plain_set_stack_nested_inplace	36.7500μs	17.5913μs	56.8462 KOps/s	54.0373 KOps/s	$\textbf{\color{#35bf28}+5.20\%}$
test_items	21.4710μs	4.7121μs	212.2187 KOps/s	212.9091 KOps/s	$\color{#d91a1a}-0.32\%$
test_items_nested	0.3992ms	0.3605ms	2.7740 KOps/s	2.7604 KOps/s	$\color{#35bf28}+0.49\%$
test_items_nested_locked	0.4018ms	0.3605ms	2.7742 KOps/s	2.7133 KOps/s	$\color{#35bf28}+2.25\%$
test_items_nested_leaf	0.1090ms	86.9033μs	11.5070 KOps/s	11.8862 KOps/s	$\color{#d91a1a}-3.19\%$
test_items_stack_nested	0.4021ms	0.3610ms	2.7701 KOps/s	2.7128 KOps/s	$\color{#35bf28}+2.11\%$
test_items_stack_nested_leaf	0.1091ms	84.2656μs	11.8672 KOps/s	11.7385 KOps/s	$\color{#35bf28}+1.10\%$
test_items_stack_nested_locked	0.3993ms	0.3647ms	2.7417 KOps/s	2.7169 KOps/s	$\color{#35bf28}+0.91\%$
test_keys	26.0790μs	4.3448μs	230.1590 KOps/s	228.3803 KOps/s	$\color{#35bf28}+0.78\%$
test_keys_nested	90.0610μs	67.4461μs	14.8267 KOps/s	14.9945 KOps/s	$\color{#d91a1a}-1.12\%$
test_keys_nested_locked	2.4560ms	72.2595μs	13.8390 KOps/s	13.9412 KOps/s	$\color{#d91a1a}-0.73\%$
test_keys_nested_leaf	76.1110μs	57.0947μs	17.5148 KOps/s	17.6999 KOps/s	$\color{#d91a1a}-1.05\%$
test_keys_stack_nested	87.7400μs	67.5043μs	14.8139 KOps/s	15.1523 KOps/s	$\color{#d91a1a}-2.23\%$
test_keys_stack_nested_leaf	79.4010μs	57.5890μs	17.3644 KOps/s	17.7699 KOps/s	$\color{#d91a1a}-2.28\%$
test_keys_stack_nested_locked	88.8120μs	72.3060μs	13.8301 KOps/s	14.0993 KOps/s	$\color{#d91a1a}-1.91\%$
test_values	10.5705μs	1.7791μs	562.0664 KOps/s	569.3669 KOps/s	$\color{#d91a1a}-1.28\%$
test_values_nested	58.7410μs	33.7805μs	29.6029 KOps/s	29.5694 KOps/s	$\color{#35bf28}+0.11\%$
test_values_nested_locked	56.1620μs	35.5087μs	28.1621 KOps/s	27.8615 KOps/s	$\color{#35bf28}+1.08\%$
test_values_nested_leaf	52.2900μs	29.9729μs	33.3635 KOps/s	32.9954 KOps/s	$\color{#35bf28}+1.12\%$
test_values_stack_nested	58.6010μs	34.0386μs	29.3784 KOps/s	29.1563 KOps/s	$\color{#35bf28}+0.76\%$
test_values_stack_nested_leaf	50.9910μs	30.2465μs	33.0616 KOps/s	32.6745 KOps/s	$\color{#35bf28}+1.18\%$
test_values_stack_nested_locked	56.6510μs	35.9864μs	27.7883 KOps/s	27.7168 KOps/s	$\color{#35bf28}+0.26\%$
test_membership	17.2810μs	0.6593μs	1.5168 MOps/s	1.8595 MOps/s	$\textbf{\color{#d91a1a}-18.43\%}$
test_membership_nested	29.9510μs	2.0171μs	495.7548 KOps/s	516.6183 KOps/s	$\color{#d91a1a}-4.04\%$
test_membership_nested_leaf	13.4555μs	1.9497μs	512.8884 KOps/s	522.9499 KOps/s	$\color{#d91a1a}-1.92\%$
test_membership_stacked_nested	22.6610μs	2.0384μs	490.5771 KOps/s	506.4207 KOps/s	$\color{#d91a1a}-3.13\%$
test_membership_stacked_nested_leaf	31.7200μs	1.9888μs	502.8214 KOps/s	513.0577 KOps/s	$\color{#d91a1a}-2.00\%$
test_membership_nested_last	34.4420μs	2.9392μs	340.2255 KOps/s	338.0408 KOps/s	$\color{#35bf28}+0.65\%$
test_membership_nested_leaf_last	24.2600μs	2.9788μs	335.7051 KOps/s	340.2811 KOps/s	$\color{#d91a1a}-1.34\%$
test_membership_stacked_nested_last	38.5110μs	2.9359μs	340.6112 KOps/s	334.7437 KOps/s	$\color{#35bf28}+1.75\%$
test_membership_stacked_nested_leaf_last	18.9600μs	2.9028μs	344.4967 KOps/s	341.7779 KOps/s	$\color{#35bf28}+0.80\%$
test_nested_getleaf	34.7600μs	7.8620μs	127.1943 KOps/s	127.6049 KOps/s	$\color{#d91a1a}-0.32\%$
test_nested_get	29.6810μs	7.3846μs	135.4164 KOps/s	136.3190 KOps/s	$\color{#d91a1a}-0.66\%$
test_stacked_getleaf	36.3400μs	7.8493μs	127.3996 KOps/s	127.8768 KOps/s	$\color{#d91a1a}-0.37\%$
test_stacked_get	23.9910μs	7.3897μs	135.3233 KOps/s	136.5724 KOps/s	$\color{#d91a1a}-0.91\%$
test_nested_getitemleaf	23.6110μs	8.0835μs	123.7095 KOps/s	123.4178 KOps/s	$\color{#35bf28}+0.24\%$
test_nested_getitem	25.1000μs	7.6567μs	130.6047 KOps/s	130.8080 KOps/s	$\color{#d91a1a}-0.16\%$
test_stacked_getitemleaf	34.2110μs	8.0971μs	123.5005 KOps/s	123.4172 KOps/s	$\color{#35bf28}+0.07\%$
test_stacked_getitem	30.0500μs	7.6103μs	131.4012 KOps/s	131.8058 KOps/s	$\color{#d91a1a}-0.31\%$
test_lock_nested	7.4554ms	0.4725ms	2.1166 KOps/s	2.0975 KOps/s	$\color{#35bf28}+0.91\%$
test_lock_stack_nested	0.4842ms	0.4321ms	2.3141 KOps/s	2.2640 KOps/s	$\color{#35bf28}+2.21\%$
test_unlock_nested	0.8820ms	0.3815ms	2.6212 KOps/s	2.5091 KOps/s	$\color{#35bf28}+4.47\%$
test_unlock_stack_nested	0.3844ms	0.3497ms	2.8592 KOps/s	2.7583 KOps/s	$\color{#35bf28}+3.66\%$
test_flatten_speed	0.5054ms	0.1038ms	9.6312 KOps/s	9.5787 KOps/s	$\color{#35bf28}+0.55\%$
test_unflatten_speed	0.3638ms	0.3169ms	3.1552 KOps/s	3.1430 KOps/s	$\color{#35bf28}+0.39\%$
test_common_ops	1.5976ms	1.3559ms	737.4926 Ops/s	709.6307 Ops/s	$\color{#35bf28}+3.93\%$
test_creation	20.0100μs	1.6344μs	611.8436 KOps/s	605.7707 KOps/s	$\color{#35bf28}+1.00\%$
test_creation_empty	38.2810μs	16.1998μs	61.7290 KOps/s	54.4259 KOps/s	$\textbf{\color{#35bf28}+13.42\%}$
test_creation_nested_1	1.0858ms	18.3865μs	54.3877 KOps/s	50.2701 KOps/s	$\textbf{\color{#35bf28}+8.19\%}$
test_creation_nested_2	47.5410μs	21.2142μs	47.1383 KOps/s	42.8106 KOps/s	$\textbf{\color{#35bf28}+10.11\%}$
test_clone	54.9620μs	29.8092μs	33.5467 KOps/s	30.3126 KOps/s	$\textbf{\color{#35bf28}+10.67\%}$
test_getitem[int]	1.1760ms	17.6444μs	56.6752 KOps/s	53.1296 KOps/s	$\textbf{\color{#35bf28}+6.67\%}$
test_getitem[slice_int]	0.1520ms	29.7492μs	33.6143 KOps/s	31.9003 KOps/s	$\textbf{\color{#35bf28}+5.37\%}$
test_getitem[range]	0.2947ms	0.1200ms	8.3366 KOps/s	8.5380 KOps/s	$\color{#d91a1a}-2.36\%$
test_getitem[tuple]	0.1508ms	26.8672μs	37.2201 KOps/s	36.7824 KOps/s	$\color{#35bf28}+1.19\%$
test_getitem[list]	0.2300ms	0.1078ms	9.2804 KOps/s	9.3285 KOps/s	$\color{#d91a1a}-0.52\%$
test_setitem_dim[int]	76.6220μs	55.0845μs	18.1539 KOps/s	16.9365 KOps/s	$\textbf{\color{#35bf28}+7.19\%}$
test_setitem_dim[slice_int]	0.1042ms	80.1279μs	12.4801 KOps/s	12.0336 KOps/s	$\color{#35bf28}+3.71\%$
test_setitem_dim[range]	0.1899ms	0.1453ms	6.8832 KOps/s	6.8413 KOps/s	$\color{#35bf28}+0.61\%$
test_setitem_dim[tuple]	0.1017ms	77.0709μs	12.9751 KOps/s	13.1575 KOps/s	$\color{#d91a1a}-1.39\%$
test_setitem	79.9320μs	45.1690μs	22.1391 KOps/s	21.4140 KOps/s	$\color{#35bf28}+3.39\%$
test_set	78.4220μs	43.8273μs	22.8168 KOps/s	21.7069 KOps/s	$\textbf{\color{#35bf28}+5.11\%}$
test_set_shared	92.8241ms	63.5737μs	15.7298 KOps/s	17.3249 KOps/s	$\textbf{\color{#d91a1a}-9.21\%}$
test_update	89.5320μs	53.0263μs	18.8586 KOps/s	17.7105 KOps/s	$\textbf{\color{#35bf28}+6.48\%}$
test_update_nested	90.4520μs	59.8253μs	16.7153 KOps/s	15.1220 KOps/s	$\textbf{\color{#35bf28}+10.54\%}$
test_update__nested	98.7710μs	61.0891μs	16.3695 KOps/s	14.6318 KOps/s	$\textbf{\color{#35bf28}+11.88\%}$
test_set_nested	88.5910μs	46.4047μs	21.5495 KOps/s	20.4075 KOps/s	$\textbf{\color{#35bf28}+5.60\%}$
test_set_nested_new	85.2110μs	47.9680μs	20.8472 KOps/s	17.9946 KOps/s	$\textbf{\color{#35bf28}+15.85\%}$
test_select	0.1227ms	64.7519μs	15.4436 KOps/s	14.6730 KOps/s	$\textbf{\color{#35bf28}+5.25\%}$
test_select_nested	0.4925ms	52.4386μs	19.0699 KOps/s	19.3167 KOps/s	$\color{#d91a1a}-1.28\%$
test_exclude_nested	98.8910μs	69.3698μs	14.4155 KOps/s	14.3987 KOps/s	$\color{#35bf28}+0.12\%$
test_empty[True]	0.3315ms	0.2829ms	3.5352 KOps/s	3.5181 KOps/s	$\color{#35bf28}+0.48\%$
test_empty[False]	2.9071μs	0.8555μs	1.1689 MOps/s	1.1538 MOps/s	$\color{#35bf28}+1.31\%$
test_to	55.5110μs	27.0267μs	37.0005 KOps/s	35.3081 KOps/s	$\color{#35bf28}+4.79\%$
test_to_nonblocking	54.2900μs	26.3348μs	37.9725 KOps/s	37.7068 KOps/s	$\color{#35bf28}+0.70\%$
test_unbind_speed	1.2973ms	0.2979ms	3.3573 KOps/s	3.2047 KOps/s	$\color{#35bf28}+4.76\%$
test_unbind_speed_stack0	0.3442ms	0.2958ms	3.3805 KOps/s	3.2325 KOps/s	$\color{#35bf28}+4.58\%$
test_unbind_speed_stack1	91.7662ms	0.7696ms	1.2994 KOps/s	1.2594 KOps/s	$\color{#35bf28}+3.17\%$
test_split	92.9094ms	2.3549ms	424.6432 Ops/s	411.7045 Ops/s	$\color{#35bf28}+3.14\%$
test_chunk	2.2977ms	2.1641ms	462.0962 Ops/s	411.2591 Ops/s	$\textbf{\color{#35bf28}+12.36\%}$
test_creation[device0]	0.1639ms	0.1104ms	9.0606 KOps/s	9.2341 KOps/s	$\color{#d91a1a}-1.88\%$
test_creation_from_tensor	0.1664ms	0.1082ms	9.2400 KOps/s	9.4295 KOps/s	$\color{#d91a1a}-2.01\%$
test_add_one[memmap_tensor0]	0.1512ms	9.4687μs	105.6106 KOps/s	99.7784 KOps/s	$\textbf{\color{#35bf28}+5.85\%}$
test_contiguous[memmap_tensor0]	24.4200μs	2.2703μs	440.4676 KOps/s	424.9915 KOps/s	$\color{#35bf28}+3.64\%$
test_stack[memmap_tensor0]	37.6600μs	6.9396μs	144.1005 KOps/s	135.5578 KOps/s	$\textbf{\color{#35bf28}+6.30\%}$
test_memmaptd_index	1.2806ms	0.4546ms	2.1999 KOps/s	2.1374 KOps/s	$\color{#35bf28}+2.92\%$
test_memmaptd_index_astensor	97.1417ms	0.6090ms	1.6419 KOps/s	1.8697 KOps/s	$\textbf{\color{#d91a1a}-12.18\%}$
test_memmaptd_index_op	1.4804ms	1.0898ms	917.5655 Ops/s	867.4427 Ops/s	$\textbf{\color{#35bf28}+5.78\%}$
test_serialize_model	91.9756ms	88.9556ms	11.2416 Ops/s	10.8865 Ops/s	$\color{#35bf28}+3.26\%$
test_serialize_model_pickle	1.3488s	1.2366s	0.8087 Ops/s	0.8080 Ops/s	$\color{#35bf28}+0.08\%$
test_serialize_weights	89.8306ms	85.7310ms	11.6644 Ops/s	11.0766 Ops/s	$\textbf{\color{#35bf28}+5.31\%}$
test_serialize_weights_returnearly	58.5255ms	53.0578ms	18.8474 Ops/s	14.9228 Ops/s	$\textbf{\color{#35bf28}+26.30\%}$
test_serialize_weights_pickle	1.3508s	1.2375s	0.8081 Ops/s	0.8083 Ops/s	$\color{#d91a1a}-0.03\%$
test_reshape_pytree	63.3320μs	38.3906μs	26.0481 KOps/s	24.7657 KOps/s	$\textbf{\color{#35bf28}+5.18\%}$
test_reshape_td	0.2486ms	43.6455μs	22.9119 KOps/s	21.8843 KOps/s	$\color{#35bf28}+4.70\%$
test_view_pytree	61.7710μs	37.6279μs	26.5761 KOps/s	25.3308 KOps/s	$\color{#35bf28}+4.92\%$
test_view_td	0.2593ms	47.9099μs	20.8725 KOps/s	18.9791 KOps/s	$\textbf{\color{#35bf28}+9.98\%}$
test_unbind_pytree	68.4820μs	37.0180μs	27.0139 KOps/s	25.8622 KOps/s	$\color{#35bf28}+4.45\%$
test_unbind_td	0.3654ms	45.9114μs	21.7811 KOps/s	20.6732 KOps/s	$\textbf{\color{#35bf28}+5.36\%}$
test_split_pytree	76.9310μs	50.3647μs	19.8552 KOps/s	18.2497 KOps/s	$\textbf{\color{#35bf28}+8.80\%}$
test_split_td	0.4517ms	58.4309μs	17.1142 KOps/s	13.4623 KOps/s	$\textbf{\color{#35bf28}+27.13\%}$
test_add_pytree	89.4720μs	59.1530μs	16.9053 KOps/s	15.3729 KOps/s	$\textbf{\color{#35bf28}+9.97\%}$
test_add_td	0.3129ms	95.9315μs	10.4241 KOps/s	9.4713 KOps/s	$\textbf{\color{#35bf28}+10.06\%}$
test_compile_add_one_nested[tensordict-compile]	0.4244ms	0.2203ms	4.5390 KOps/s	4.4948 KOps/s	$\color{#35bf28}+0.98\%$
test_compile_add_one_nested[tensordict-eager]	0.2664ms	0.1746ms	5.7283 KOps/s	5.5315 KOps/s	$\color{#35bf28}+3.56\%$
test_compile_add_one_nested[pytree-compile]	0.1936ms	0.1552ms	6.4437 KOps/s	6.3200 KOps/s	$\color{#35bf28}+1.96\%$
test_compile_add_one_nested[pytree-eager]	0.2466ms	0.1929ms	5.1848 KOps/s	4.9598 KOps/s	$\color{#35bf28}+4.54\%$
test_compile_copy_nested[tensordict-compile]	59.1220μs	22.6339μs	44.1815 KOps/s	43.7496 KOps/s	$\color{#35bf28}+0.99\%$
test_compile_copy_nested[tensordict-eager]	79.8410μs	48.0632μs	20.8059 KOps/s	19.9822 KOps/s	$\color{#35bf28}+4.12\%$
test_compile_copy_nested[pytree-compile]	0.1108ms	74.7653μs	13.3752 KOps/s	13.5879 KOps/s	$\color{#d91a1a}-1.57\%$
test_compile_copy_nested[pytree-eager]	84.0420μs	59.3263μs	16.8559 KOps/s	16.8789 KOps/s	$\color{#d91a1a}-0.14\%$
test_compile_add_one_flat[tensordict-compile]	0.5053ms	0.3463ms	2.8873 KOps/s	2.9084 KOps/s	$\color{#d91a1a}-0.73\%$
test_compile_add_one_flat[tensordict-eager]	0.2803ms	0.2219ms	4.5055 KOps/s	4.3551 KOps/s	$\color{#35bf28}+3.45\%$
test_compile_add_one_flat[tensorclass-compile]	0.1873ms	0.1384ms	7.2261 KOps/s	7.2020 KOps/s	$\color{#35bf28}+0.34\%$
test_compile_add_one_flat[tensorclass-eager]	0.1230ms	64.6867μs	15.4591 KOps/s	15.2609 KOps/s	$\color{#35bf28}+1.30\%$
test_compile_add_one_flat[pytree-compile]	0.3958ms	0.3449ms	2.8994 KOps/s	2.9019 KOps/s	$\color{#d91a1a}-0.09\%$
test_compile_add_one_flat[pytree-eager]	0.7147ms	0.6400ms	1.5625 KOps/s	1.5011 KOps/s	$\color{#35bf28}+4.09\%$
test_compile_add_self_flat[tensordict-eager]	0.3122ms	0.2691ms	3.7167 KOps/s	3.6184 KOps/s	$\color{#35bf28}+2.71\%$
test_compile_add_self_flat[tensordict-compile]	0.3945ms	0.3468ms	2.8834 KOps/s	2.8866 KOps/s	$\color{#d91a1a}-0.11\%$
test_compile_add_self_flat[tensorclass-eager]	0.1751ms	76.9582μs	12.9941 KOps/s	12.8248 KOps/s	$\color{#35bf28}+1.32\%$
test_compile_add_self_flat[tensorclass-compile]	0.1934ms	0.1440ms	6.9435 KOps/s	7.1761 KOps/s	$\color{#d91a1a}-3.24\%$
test_compile_add_self_flat[pytree-eager]	0.6705ms	0.5424ms	1.8436 KOps/s	1.7574 KOps/s	$\color{#35bf28}+4.91\%$
test_compile_add_self_flat[pytree-compile]	0.4340ms	0.3445ms	2.9030 KOps/s	2.8976 KOps/s	$\color{#35bf28}+0.19\%$
test_compile_copy_flat[tensordict-compile]	48.3510μs	19.9274μs	50.1822 KOps/s	49.0181 KOps/s	$\color{#35bf28}+2.37\%$
test_compile_copy_flat[tensordict-eager]	64.7410μs	31.6887μs	31.5570 KOps/s	30.4326 KOps/s	$\color{#35bf28}+3.69\%$
test_compile_copy_flat[pytree-compile]	0.1136ms	76.8512μs	13.0122 KOps/s	12.9645 KOps/s	$\color{#35bf28}+0.37\%$
test_compile_copy_flat[pytree-eager]	87.4720μs	60.5115μs	16.5258 KOps/s	16.4745 KOps/s	$\color{#35bf28}+0.31\%$
test_compile_assign_and_add[tensordict-compile]	2.4807ms	0.8786ms	1.1382 KOps/s	1.0578 KOps/s	$\textbf{\color{#35bf28}+7.60\%}$
test_compile_assign_and_add[tensordict-eager]	3.6026ms	3.3976ms	294.3232 Ops/s	285.1702 Ops/s	$\color{#35bf28}+3.21\%$
test_compile_assign_and_add[pytree-compile]	2.5032ms	0.8813ms	1.1347 KOps/s	1.0700 KOps/s	$\textbf{\color{#35bf28}+6.05\%}$
test_compile_assign_and_add[pytree-eager]	3.4935ms	3.4550ms	289.4320 Ops/s	281.3667 Ops/s	$\color{#35bf28}+2.87\%$
test_compile_indexing[tensor-tensordict-compile]	0.1814ms	0.1184ms	8.4445 KOps/s	8.3998 KOps/s	$\color{#35bf28}+0.53\%$
test_compile_indexing[tensor-tensordict-eager]	0.2250ms	63.6984μs	15.6990 KOps/s	15.1795 KOps/s	$\color{#35bf28}+3.42\%$
test_compile_indexing[tensor-tensorclass-compile]	0.1440ms	0.1110ms	9.0052 KOps/s	8.9459 KOps/s	$\color{#35bf28}+0.66\%$
test_compile_indexing[tensor-tensorclass-eager]	0.1837ms	46.5963μs	21.4609 KOps/s	20.9719 KOps/s	$\color{#35bf28}+2.33\%$
test_compile_indexing[tensor-pytree-compile]	0.1373ms	0.1107ms	9.0308 KOps/s	8.6678 KOps/s	$\color{#35bf28}+4.19\%$
test_compile_indexing[tensor-pytree-eager]	90.1720μs	47.3374μs	21.1249 KOps/s	19.4909 KOps/s	$\textbf{\color{#35bf28}+8.38\%}$
test_compile_indexing[slice-tensordict-compile]	0.1880ms	0.1481ms	6.7537 KOps/s	6.7029 KOps/s	$\color{#35bf28}+0.76\%$
test_compile_indexing[slice-tensordict-eager]	0.1727ms	26.5483μs	37.6672 KOps/s	35.7572 KOps/s	$\textbf{\color{#35bf28}+5.34\%}$
test_compile_indexing[slice-tensorclass-compile]	0.1705ms	0.1404ms	7.1220 KOps/s	7.1204 KOps/s	$\color{#35bf28}+0.02\%$
test_compile_indexing[slice-tensorclass-eager]	46.6810μs	22.9661μs	43.5424 KOps/s	41.9504 KOps/s	$\color{#35bf28}+3.80\%$
test_compile_indexing[slice-pytree-compile]	0.2396ms	0.1400ms	7.1403 KOps/s	7.0138 KOps/s	$\color{#35bf28}+1.80\%$
test_compile_indexing[slice-pytree-eager]	51.5500μs	22.7697μs	43.9180 KOps/s	42.4636 KOps/s	$\color{#35bf28}+3.43\%$
test_compile_indexing[int-tensordict-compile]	0.1838ms	0.1488ms	6.7221 KOps/s	6.7328 KOps/s	$\color{#d91a1a}-0.16\%$
test_compile_indexing[int-tensordict-eager]	0.4862ms	26.0994μs	38.3151 KOps/s	37.1112 KOps/s	$\color{#35bf28}+3.24\%$
test_compile_indexing[int-tensorclass-compile]	0.2480ms	0.1397ms	7.1558 KOps/s	7.1301 KOps/s	$\color{#35bf28}+0.36\%$
test_compile_indexing[int-tensorclass-eager]	47.0810μs	23.0887μs	43.3112 KOps/s	42.7477 KOps/s	$\color{#35bf28}+1.32\%$
test_compile_indexing[int-pytree-compile]	0.1830ms	0.1397ms	7.1571 KOps/s	7.1386 KOps/s	$\color{#35bf28}+0.26\%$
test_compile_indexing[int-pytree-eager]	45.1910μs	22.6792μs	44.0933 KOps/s	42.2548 KOps/s	$\color{#35bf28}+4.35\%$
test_mod_add[eager]	67.8810μs	33.0774μs	30.2321 KOps/s	28.8900 KOps/s	$\color{#35bf28}+4.65\%$
test_mod_add[compile]	0.1034ms	73.9123μs	13.5296 KOps/s	12.9670 KOps/s	$\color{#35bf28}+4.34\%$
test_mod_add[compile-overhead]	0.2676ms	0.1447ms	6.9091 KOps/s	6.3987 KOps/s	$\textbf{\color{#35bf28}+7.98\%}$
test_mod_wrap[eager]	0.3517ms	0.2510ms	3.9834 KOps/s	3.6498 KOps/s	$\textbf{\color{#35bf28}+9.14\%}$
test_mod_wrap[compile]	1.2023ms	0.2995ms	3.3393 KOps/s	3.1106 KOps/s	$\textbf{\color{#35bf28}+7.35\%}$
test_mod_wrap[compile-overhead]	8.2204ms	4.3193ms	231.5172 Ops/s	233.6075 Ops/s	$\color{#d91a1a}-0.89\%$
test_mod_wrap_and_backward[eager]	1.5236ms	1.3898ms	719.5095 Ops/s	674.1418 Ops/s	$\textbf{\color{#35bf28}+6.73\%}$
test_mod_wrap_and_backward[compile]	1.7028ms	1.3714ms	729.1751 Ops/s	665.9277 Ops/s	$\textbf{\color{#35bf28}+9.50\%}$
test_mod_wrap_and_backward[compile-overhead]	1.4380ms	0.9983ms	1.0017 KOps/s	959.7093 Ops/s	$\color{#35bf28}+4.37\%$
test_seq_add[eager]	0.1594ms	0.1032ms	9.6885 KOps/s	8.8873 KOps/s	$\textbf{\color{#35bf28}+9.02\%}$
test_seq_add[compile]	0.1456ms	87.0193μs	11.4917 KOps/s	11.3781 KOps/s	$\color{#35bf28}+1.00\%$
test_seq_add[compile-overhead]	0.1595ms	0.1235ms	8.0952 KOps/s	7.9855 KOps/s	$\color{#35bf28}+1.37\%$
test_seq_wrap[eager]	0.4671ms	0.3976ms	2.5149 KOps/s	2.4534 KOps/s	$\color{#35bf28}+2.51\%$
test_seq_wrap[compile]	0.3811ms	0.3246ms	3.0808 KOps/s	2.9971 KOps/s	$\color{#35bf28}+2.79\%$
test_seq_wrap[compile-overhead]	0.2989ms	0.2375ms	4.2097 KOps/s	4.1618 KOps/s	$\color{#35bf28}+1.15\%$
test_func_call_runtime[False-eager]	0.8490ms	0.7582ms	1.3189 KOps/s	1.2462 KOps/s	$\textbf{\color{#35bf28}+5.83\%}$
test_func_call_runtime[False-compile]	0.9139ms	0.8030ms	1.2453 KOps/s	1.1902 KOps/s	$\color{#35bf28}+4.63\%$
test_func_call_runtime[False-compile-overhead]	0.4217ms	0.3828ms	2.6121 KOps/s	2.5870 KOps/s	$\color{#35bf28}+0.97\%$
test_func_call_runtime[True-eager]	1.0709ms	0.9401ms	1.0637 KOps/s	1.0366 KOps/s	$\color{#35bf28}+2.62\%$
test_func_call_runtime[True-compile]	0.9291ms	0.8509ms	1.1752 KOps/s	1.1305 KOps/s	$\color{#35bf28}+3.96\%$
test_func_call_runtime[True-compile-overhead]	0.4777ms	0.4240ms	2.3587 KOps/s	2.3184 KOps/s	$\color{#35bf28}+1.73\%$
test_func_call_cm_runtime[False-eager]	0.8739ms	0.7522ms	1.3295 KOps/s	1.3048 KOps/s	$\color{#35bf28}+1.89\%$
test_func_call_cm_runtime[False-compile]	0.8843ms	0.8035ms	1.2445 KOps/s	1.1867 KOps/s	$\color{#35bf28}+4.87\%$
test_func_call_cm_runtime[False-compile-overhead]	0.4291ms	0.3845ms	2.6007 KOps/s	2.5950 KOps/s	$\color{#35bf28}+0.22\%$
test_func_call_cm_runtime[True-eager]	1.1784ms	1.0482ms	954.0519 Ops/s	925.8179 Ops/s	$\color{#35bf28}+3.05\%$
test_func_call_cm_runtime[True-compile]	1.1282ms	1.0275ms	973.2354 Ops/s	947.3746 Ops/s	$\color{#35bf28}+2.73\%$
test_func_call_cm_runtime[True-compile-overhead]	1.1373ms	1.0334ms	967.6626 Ops/s	946.6131 Ops/s	$\color{#35bf28}+2.22\%$
test_distributed	2.2169ms	71.0611μs	14.0724 KOps/s	14.3982 KOps/s	$\color{#d91a1a}-2.26\%$
test_tdmodule	0.1306ms	16.1199μs	62.0352 KOps/s	57.3812 KOps/s	$\textbf{\color{#35bf28}+8.11\%}$
test_tdmodule_dispatch	51.1400μs	33.1497μs	30.1661 KOps/s	29.3308 KOps/s	$\color{#35bf28}+2.85\%$
test_tdseq	31.4410μs	16.3116μs	61.3059 KOps/s	55.3768 KOps/s	$\textbf{\color{#35bf28}+10.71\%}$
test_tdseq_dispatch	62.4510μs	34.5900μs	28.9101 KOps/s	27.4430 KOps/s	$\textbf{\color{#35bf28}+5.35\%}$
test_instantiation_functorch	2.1484ms	2.0231ms	494.2865 Ops/s	480.7596 Ops/s	$\color{#35bf28}+2.81\%$
test_instantiation_td	1.9956ms	1.3137ms	761.2191 Ops/s	741.9893 Ops/s	$\color{#35bf28}+2.59\%$
test_exec_functorch	0.2882ms	0.2332ms	4.2877 KOps/s	4.3083 KOps/s	$\color{#d91a1a}-0.48\%$
test_exec_functional_call	0.2756ms	0.2245ms	4.4544 KOps/s	4.4359 KOps/s	$\color{#35bf28}+0.41\%$
test_exec_td	0.2863ms	0.2322ms	4.3066 KOps/s	4.2257 KOps/s	$\color{#35bf28}+1.91\%$
test_exec_td_decorator	0.4203ms	0.2842ms	3.5184 KOps/s	3.5008 KOps/s	$\color{#35bf28}+0.50\%$
test_vmap_mlp_speed[True-True]	1.0818ms	0.6871ms	1.4555 KOps/s	1.4842 KOps/s	$\color{#d91a1a}-1.94\%$
test_vmap_mlp_speed[True-False]	0.7460ms	0.6777ms	1.4757 KOps/s	1.5217 KOps/s	$\color{#d91a1a}-3.03\%$
test_vmap_mlp_speed[False-True]	0.7069ms	0.6034ms	1.6573 KOps/s	1.6750 KOps/s	$\color{#d91a1a}-1.05\%$
test_vmap_mlp_speed[False-False]	0.7562ms	0.6046ms	1.6539 KOps/s	1.7382 KOps/s	$\color{#d91a1a}-4.85\%$
test_vmap_mlp_speed_decorator[True-True]	1.2762ms	0.7253ms	1.3787 KOps/s	1.3920 KOps/s	$\color{#d91a1a}-0.96\%$
test_vmap_mlp_speed_decorator[True-False]	0.8631ms	0.7230ms	1.3831 KOps/s	1.4004 KOps/s	$\color{#d91a1a}-1.24\%$
test_vmap_mlp_speed_decorator[False-True]	0.8084ms	0.6420ms	1.5576 KOps/s	1.6131 KOps/s	$\color{#d91a1a}-3.44\%$
test_vmap_mlp_speed_decorator[False-False]	0.7681ms	0.6418ms	1.5580 KOps/s	1.6130 KOps/s	$\color{#d91a1a}-3.41\%$
test_vmap_transformer_speed[True-True]	9.7319ms	9.0996ms	109.8953 Ops/s	112.4520 Ops/s	$\color{#d91a1a}-2.27\%$
test_vmap_transformer_speed[True-False]	9.3235ms	9.1003ms	109.8863 Ops/s	112.5415 Ops/s	$\color{#d91a1a}-2.36\%$
test_vmap_transformer_speed[False-True]	9.3839ms	9.0476ms	110.5263 Ops/s	113.6750 Ops/s	$\color{#d91a1a}-2.77\%$
test_vmap_transformer_speed[False-False]	9.4642ms	9.0727ms	110.2211 Ops/s	113.6027 Ops/s	$\color{#d91a1a}-2.98\%$
test_vmap_transformer_speed_decorator[True-True]	21.9784ms	21.2502ms	47.0584 Ops/s	47.8920 Ops/s	$\color{#d91a1a}-1.74\%$
test_vmap_transformer_speed_decorator[True-False]	22.2636ms	21.2681ms	47.0188 Ops/s	47.7093 Ops/s	$\color{#d91a1a}-1.45\%$
test_vmap_transformer_speed_decorator[False-True]	21.9077ms	21.5709ms	46.3589 Ops/s	48.1983 Ops/s	$\color{#d91a1a}-3.82\%$
test_vmap_transformer_speed_decorator[False-False]	22.4343ms	21.5105ms	46.4889 Ops/s	48.0699 Ops/s	$\color{#d91a1a}-3.29\%$
test_to_module_speed[True]	1.2477ms	1.1494ms	870.0229 Ops/s	864.9787 Ops/s	$\color{#35bf28}+0.58\%$
test_to_module_speed[False]	1.6208ms	1.1216ms	891.6029 Ops/s	890.6113 Ops/s	$\color{#35bf28}+0.11\%$
test_tc_init	93.0110μs	38.3116μs	26.1018 KOps/s	25.4783 KOps/s	$\color{#35bf28}+2.45\%$
test_tc_init_nested	0.1114ms	80.5705μs	12.4115 KOps/s	12.6318 KOps/s	$\color{#d91a1a}-1.74\%$
test_tc_first_layer_tensor	4.7368μs	0.7915μs	1.2633 MOps/s	1.2772 MOps/s	$\color{#d91a1a}-1.09\%$
test_tc_first_layer_nontensor	15.6300μs	2.5609μs	390.4902 KOps/s	392.9770 KOps/s	$\color{#d91a1a}-0.63\%$
test_tc_second_layer_tensor	6.4033μs	1.6170μs	618.4151 KOps/s	619.2610 KOps/s	$\color{#d91a1a}-0.14\%$
test_tc_second_layer_nontensor	16.9910μs	3.3483μs	298.6625 KOps/s	295.8309 KOps/s	$\color{#35bf28}+0.96\%$
test_unbind	0.1813s	10.5059ms	95.1843 Ops/s	64.0069 Ops/s	$\textbf{\color{#35bf28}+48.71\%}$
test_full_like	0.1691s	0.6604ms	1.5142 KOps/s	1.7340 KOps/s	$\textbf{\color{#d91a1a}-12.68\%}$
test_zeros_like	0.2650ms	0.1974ms	5.0652 KOps/s	5.0590 KOps/s	$\color{#35bf28}+0.12\%$
test_ones_like	0.2334ms	0.1975ms	5.0642 KOps/s	5.0626 KOps/s	$\color{#35bf28}+0.03\%$
test_clone	0.4427ms	0.4144ms	2.4130 KOps/s	2.4134 KOps/s	$\color{#d91a1a}-0.02\%$
test_squeeze	29.2110μs	10.6579μs	93.8269 KOps/s	88.6054 KOps/s	$\textbf{\color{#35bf28}+5.89\%}$
test_unsqueeze	0.2146ms	80.3084μs	12.4520 KOps/s	11.9587 KOps/s	$\color{#35bf28}+4.13\%$
test_split	0.4437ms	0.1734ms	5.7681 KOps/s	5.6765 KOps/s	$\color{#35bf28}+1.61\%$
test_permute	0.2507ms	0.1896ms	5.2755 KOps/s	5.2027 KOps/s	$\color{#35bf28}+1.40\%$
test_stack	1.2481ms	0.9055ms	1.1043 KOps/s	1.0808 KOps/s	$\color{#35bf28}+2.18\%$
test_cat	1.2476ms	1.2316ms	811.9537 Ops/s	811.8915 Ops/s	$+0.01\%$