[Feature] Densify lazy 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: $\large\color{#35bf28}7$. Worsened: $\large\color{#d91a1a}49$.

Expand to view detailed results

Name	Max	Mean	Ops	Ops on Repo HEAD	Change
test_plain_set_nested	49.5830μs	22.1152μs	45.2177 KOps/s	46.3063 KOps/s	$\color{#d91a1a}-2.35\%$
test_plain_set_stack_nested	57.2970μs	22.3371μs	44.7685 KOps/s	45.9930 KOps/s	$\color{#d91a1a}-2.66\%$
test_plain_set_nested_inplace	57.6380μs	25.0612μs	39.9023 KOps/s	42.0509 KOps/s	$\textbf{\color{#d91a1a}-5.11\%}$
test_plain_set_stack_nested_inplace	61.6250μs	24.3103μs	41.1348 KOps/s	42.4230 KOps/s	$\color{#d91a1a}-3.04\%$
test_items	23.0740μs	2.7052μs	369.6620 KOps/s	352.0356 KOps/s	$\textbf{\color{#35bf28}+5.01\%}$
test_items_nested	1.4549ms	0.3704ms	2.7001 KOps/s	2.9797 KOps/s	$\textbf{\color{#d91a1a}-9.38\%}$
test_items_nested_locked	0.5983ms	0.3380ms	2.9586 KOps/s	2.9583 KOps/s	$+0.01\%$
test_items_nested_leaf	0.1627ms	87.5593μs	11.4208 KOps/s	11.9237 KOps/s	$\color{#d91a1a}-4.22\%$
test_items_stack_nested	0.6323ms	0.3403ms	2.9387 KOps/s	2.9457 KOps/s	$\color{#d91a1a}-0.24\%$
test_items_stack_nested_leaf	0.1760ms	84.3198μs	11.8596 KOps/s	12.0171 KOps/s	$\color{#d91a1a}-1.31\%$
test_items_stack_nested_locked	0.6228ms	0.3391ms	2.9487 KOps/s	2.9423 KOps/s	$\color{#35bf28}+0.22\%$
test_keys	31.7600μs	3.9907μs	250.5809 KOps/s	257.4958 KOps/s	$\color{#d91a1a}-2.69\%$
test_keys_nested	0.2518ms	0.1454ms	6.8767 KOps/s	6.8638 KOps/s	$\color{#35bf28}+0.19\%$
test_keys_nested_locked	0.6694ms	0.1508ms	6.6331 KOps/s	6.6195 KOps/s	$\color{#35bf28}+0.21\%$
test_keys_nested_leaf	0.2136ms	0.1247ms	8.0210 KOps/s	8.1322 KOps/s	$\color{#d91a1a}-1.37\%$
test_keys_stack_nested	0.2790ms	0.1463ms	6.8330 KOps/s	6.9731 KOps/s	$\color{#d91a1a}-2.01\%$
test_keys_stack_nested_leaf	0.2383ms	0.1249ms	8.0033 KOps/s	8.1730 KOps/s	$\color{#d91a1a}-2.08\%$
test_keys_stack_nested_locked	0.2486ms	0.1502ms	6.6572 KOps/s	6.7425 KOps/s	$\color{#d91a1a}-1.27\%$
test_values	9.4803μs	1.4443μs	692.3908 KOps/s	856.5606 KOps/s	$\textbf{\color{#d91a1a}-19.17\%}$
test_values_nested	0.1217ms	51.3316μs	19.4812 KOps/s	19.3442 KOps/s	$\color{#35bf28}+0.71\%$
test_values_nested_locked	97.8930μs	50.5112μs	19.7976 KOps/s	19.9921 KOps/s	$\color{#d91a1a}-0.97\%$
test_values_nested_leaf	82.9350μs	46.0704μs	21.7059 KOps/s	22.4109 KOps/s	$\color{#d91a1a}-3.15\%$
test_values_stack_nested	96.8610μs	52.2080μs	19.1542 KOps/s	19.6169 KOps/s	$\color{#d91a1a}-2.36\%$
test_values_stack_nested_leaf	0.1051ms	45.5290μs	21.9640 KOps/s	22.5449 KOps/s	$\color{#d91a1a}-2.58\%$
test_values_stack_nested_locked	0.1010ms	52.0264μs	19.2210 KOps/s	19.7326 KOps/s	$\color{#d91a1a}-2.59\%$
test_membership	2.8734μs	0.7495μs	1.3342 MOps/s	1.0913 MOps/s	$\textbf{\color{#35bf28}+22.26\%}$
test_membership_nested	30.8980μs	2.6875μs	372.0961 KOps/s	367.0342 KOps/s	$\color{#35bf28}+1.38\%$
test_membership_nested_leaf	30.8280μs	2.6859μs	372.3194 KOps/s	383.6181 KOps/s	$\color{#d91a1a}-2.95\%$
test_membership_stacked_nested	38.5020μs	2.6692μs	374.6372 KOps/s	382.2072 KOps/s	$\color{#d91a1a}-1.98\%$
test_membership_stacked_nested_leaf	38.3710μs	2.6542μs	376.7670 KOps/s	382.0374 KOps/s	$\color{#d91a1a}-1.38\%$
test_membership_nested_last	31.2180μs	3.9769μs	251.4516 KOps/s	257.3222 KOps/s	$\color{#d91a1a}-2.28\%$
test_membership_nested_leaf_last	29.4050μs	3.9623μs	252.3776 KOps/s	258.6231 KOps/s	$\color{#d91a1a}-2.41\%$
test_membership_stacked_nested_last	57.0670μs	13.0736μs	76.4898 KOps/s	202.0362 KOps/s	$\textbf{\color{#d91a1a}-62.14\%}$
test_membership_stacked_nested_leaf_last	61.4650μs	12.5924μs	79.4128 KOps/s	203.7740 KOps/s	$\textbf{\color{#d91a1a}-61.03\%}$
test_nested_getleaf	40.3560μs	10.4307μs	95.8712 KOps/s	95.3393 KOps/s	$\color{#35bf28}+0.56\%$
test_nested_get	43.2010μs	9.9345μs	100.6596 KOps/s	101.8667 KOps/s	$\color{#d91a1a}-1.18\%$
test_stacked_getleaf	38.4720μs	10.4234μs	95.9377 KOps/s	95.8762 KOps/s	$\color{#35bf28}+0.06\%$
test_stacked_get	48.5820μs	9.7560μs	102.5009 KOps/s	101.7012 KOps/s	$\color{#35bf28}+0.79\%$
test_nested_getitemleaf	41.1770μs	10.8906μs	91.8224 KOps/s	91.8996 KOps/s	$\color{#d91a1a}-0.08\%$
test_nested_getitem	42.8400μs	10.0550μs	99.4530 KOps/s	99.7175 KOps/s	$\color{#d91a1a}-0.27\%$
test_stacked_getitemleaf	0.1888ms	10.9600μs	91.2406 KOps/s	93.0340 KOps/s	$\color{#d91a1a}-1.93\%$
test_stacked_getitem	42.4100μs	10.0555μs	99.4480 KOps/s	102.8447 KOps/s	$\color{#d91a1a}-3.30\%$
test_lock_nested	80.6786ms	0.5780ms	1.7301 KOps/s	1.9760 KOps/s	$\textbf{\color{#d91a1a}-12.45\%}$
test_lock_stack_nested	0.7115ms	0.4467ms	2.2384 KOps/s	2.1635 KOps/s	$\color{#35bf28}+3.46\%$
test_unlock_nested	88.5185ms	0.5178ms	1.9314 KOps/s	2.3908 KOps/s	$\textbf{\color{#d91a1a}-19.21\%}$
test_unlock_stack_nested	0.6307ms	0.3634ms	2.7521 KOps/s	2.6603 KOps/s	$\color{#35bf28}+3.45\%$
test_flatten_speed	0.5522ms	0.1073ms	9.3154 KOps/s	9.7910 KOps/s	$\color{#d91a1a}-4.86\%$
test_unflatten_speed	0.5984ms	0.4648ms	2.1513 KOps/s	2.1953 KOps/s	$\color{#d91a1a}-2.01\%$
test_common_ops	5.4312ms	1.1783ms	848.6853 Ops/s	916.8127 Ops/s	$\textbf{\color{#d91a1a}-7.43\%}$
test_creation	0.1375ms	2.2168μs	451.1008 KOps/s	495.2284 KOps/s	$\textbf{\color{#d91a1a}-8.91\%}$
test_creation_empty	53.5910μs	19.3333μs	51.7243 KOps/s	56.3804 KOps/s	$\textbf{\color{#d91a1a}-8.26\%}$
test_creation_nested_1	60.3430μs	22.5377μs	44.3702 KOps/s	47.0005 KOps/s	$\textbf{\color{#d91a1a}-5.60\%}$
test_creation_nested_2	1.3476ms	27.4419μs	36.4406 KOps/s	39.4596 KOps/s	$\textbf{\color{#d91a1a}-7.65\%}$
test_clone	0.2651ms	17.9813μs	55.6132 KOps/s	62.0612 KOps/s	$\textbf{\color{#d91a1a}-10.39\%}$
test_getitem[int]	0.8255ms	17.1901μs	58.1730 KOps/s	61.0094 KOps/s	$\color{#d91a1a}-4.65\%$
test_getitem[slice_int]	0.1356ms	33.6375μs	29.7287 KOps/s	32.4183 KOps/s	$\textbf{\color{#d91a1a}-8.30\%}$
test_getitem[range]	0.1623ms	58.1922μs	17.1844 KOps/s	17.9576 KOps/s	$\color{#d91a1a}-4.31\%$
test_getitem[tuple]	0.1388ms	26.1217μs	38.2824 KOps/s	40.2078 KOps/s	$\color{#d91a1a}-4.79\%$
test_getitem[list]	0.2893ms	52.9444μs	18.8877 KOps/s	19.5446 KOps/s	$\color{#d91a1a}-3.36\%$
test_setitem_dim[int]	88.1450μs	43.3310μs	23.0782 KOps/s	24.0836 KOps/s	$\color{#d91a1a}-4.18\%$
test_setitem_dim[slice_int]	0.1258ms	74.4821μs	13.4260 KOps/s	13.5735 KOps/s	$\color{#d91a1a}-1.09\%$
test_setitem_dim[range]	0.2011ms	96.7834μs	10.3324 KOps/s	10.7608 KOps/s	$\color{#d91a1a}-3.98\%$
test_setitem_dim[tuple]	0.1097ms	61.0217μs	16.3876 KOps/s	16.9919 KOps/s	$\color{#d91a1a}-3.56\%$
test_setitem	0.1675ms	31.8820μs	31.3657 KOps/s	35.4565 KOps/s	$\textbf{\color{#d91a1a}-11.54\%}$
test_set	0.3143ms	31.4596μs	31.7868 KOps/s	36.0598 KOps/s	$\textbf{\color{#d91a1a}-11.85\%}$
test_set_shared	1.2976ms	0.2204ms	4.5382 KOps/s	4.6820 KOps/s	$\color{#d91a1a}-3.07\%$
test_update	0.3441ms	38.8928μs	25.7117 KOps/s	28.5334 KOps/s	$\textbf{\color{#d91a1a}-9.89\%}$
test_update_nested	0.2483ms	49.6538μs	20.1394 KOps/s	22.3366 KOps/s	$\textbf{\color{#d91a1a}-9.84\%}$
test_update__nested	0.1282ms	35.3600μs	28.2806 KOps/s	30.4925 KOps/s	$\textbf{\color{#d91a1a}-7.25\%}$
test_set_nested	0.1469ms	34.3228μs	29.1351 KOps/s	33.2697 KOps/s	$\textbf{\color{#d91a1a}-12.43\%}$
test_set_nested_new	0.1339ms	38.5237μs	25.9580 KOps/s	28.8362 KOps/s	$\textbf{\color{#d91a1a}-9.98\%}$
test_select	0.1679ms	56.2533μs	17.7767 KOps/s	19.0070 KOps/s	$\textbf{\color{#d91a1a}-6.47\%}$
test_select_nested	0.1172ms	59.5073μs	16.8047 KOps/s	16.9734 KOps/s	$\color{#d91a1a}-0.99\%$
test_exclude_nested	0.1627ms	78.0333μs	12.8150 KOps/s	13.1391 KOps/s	$\color{#d91a1a}-2.47\%$
test_empty[True]	0.7217ms	0.3252ms	3.0748 KOps/s	3.0950 KOps/s	$\color{#d91a1a}-0.65\%$
test_empty[False]	11.8497μs	1.1605μs	861.6754 KOps/s	862.6629 KOps/s	$\color{#d91a1a}-0.11\%$
test_unbind_speed	0.5256ms	0.3110ms	3.2155 KOps/s	3.2230 KOps/s	$\color{#d91a1a}-0.23\%$
test_unbind_speed_stack0	0.5736ms	0.2953ms	3.3865 KOps/s	3.3675 KOps/s	$\color{#35bf28}+0.56\%$
test_unbind_speed_stack1	81.9724ms	0.7647ms	1.3077 KOps/s	1.3964 KOps/s	$\textbf{\color{#d91a1a}-6.36\%}$
test_split	89.4074ms	2.1988ms	454.7959 Ops/s	469.0029 Ops/s	$\color{#d91a1a}-3.03\%$
test_chunk	83.8390ms	2.1872ms	457.2155 Ops/s	467.7117 Ops/s	$\color{#d91a1a}-2.24\%$
test_creation[device0]	0.2177ms	0.1216ms	8.2223 KOps/s	8.2205 KOps/s	$\color{#35bf28}+0.02\%$
test_creation_from_tensor	4.1673ms	0.1232ms	8.1155 KOps/s	8.3875 KOps/s	$\color{#d91a1a}-3.24\%$
test_add_one[memmap_tensor0]	0.2251ms	8.0027μs	124.9575 KOps/s	133.2127 KOps/s	$\textbf{\color{#d91a1a}-6.20\%}$
test_contiguous[memmap_tensor0]	18.3140μs	2.0929μs	477.7989 KOps/s	500.9327 KOps/s	$\color{#d91a1a}-4.62\%$
test_stack[memmap_tensor0]	50.8450μs	5.9258μs	168.7541 KOps/s	179.8554 KOps/s	$\textbf{\color{#d91a1a}-6.17\%}$
test_memmaptd_index	1.2011ms	0.4244ms	2.3564 KOps/s	2.4935 KOps/s	$\textbf{\color{#d91a1a}-5.50\%}$
test_memmaptd_index_astensor	1.0639ms	0.5008ms	1.9967 KOps/s	2.0933 KOps/s	$\color{#d91a1a}-4.61\%$
test_memmaptd_index_op	1.8573ms	1.0897ms	917.7253 Ops/s	984.1313 Ops/s	$\textbf{\color{#d91a1a}-6.75\%}$
test_serialize_model	0.1281s	0.1211s	8.2589 Ops/s	7.4242 Ops/s	$\textbf{\color{#35bf28}+11.24\%}$
test_serialize_model_pickle	0.4711s	0.3881s	2.5769 Ops/s	2.5405 Ops/s	$\color{#35bf28}+1.44\%$
test_serialize_weights	0.1313s	0.1213s	8.2471 Ops/s	8.7053 Ops/s	$\textbf{\color{#d91a1a}-5.26\%}$
test_serialize_weights_returnearly	0.1774s	0.1607s	6.2244 Ops/s	6.2154 Ops/s	$\color{#35bf28}+0.15\%$
test_serialize_weights_pickle	0.4415s	0.4019s	2.4884 Ops/s	2.4186 Ops/s	$\color{#35bf28}+2.88\%$
test_serialize_weights_filesystem	0.1465s	0.1416s	7.0640 Ops/s	6.5801 Ops/s	$\textbf{\color{#35bf28}+7.35\%}$
test_serialize_model_filesystem	0.2288s	0.1603s	6.2378 Ops/s	6.6309 Ops/s	$\textbf{\color{#d91a1a}-5.93\%}$
test_reshape_pytree	0.1031ms	41.0704μs	24.3484 KOps/s	25.2112 KOps/s	$\color{#d91a1a}-3.42\%$
test_reshape_td	96.0500μs	47.7815μs	20.9286 KOps/s	20.6710 KOps/s	$\color{#35bf28}+1.25\%$
test_view_pytree	94.9480μs	39.7903μs	25.1317 KOps/s	25.0476 KOps/s	$\color{#35bf28}+0.34\%$
test_view_td	0.1070ms	54.1267μs	18.4752 KOps/s	18.9574 KOps/s	$\color{#d91a1a}-2.54\%$
test_unbind_pytree	78.5570μs	37.9453μs	26.3538 KOps/s	27.2752 KOps/s	$\color{#d91a1a}-3.38\%$
test_unbind_td	0.3646ms	47.0479μs	21.2550 KOps/s	21.8943 KOps/s	$\color{#d91a1a}-2.92\%$
test_split_pytree	80.5700μs	41.8738μs	23.8813 KOps/s	25.1560 KOps/s	$\textbf{\color{#d91a1a}-5.07\%}$
test_split_td	0.5653ms	60.8105μs	16.4445 KOps/s	17.2782 KOps/s	$\color{#d91a1a}-4.83\%$
test_add_pytree	0.1250ms	50.0850μs	19.9661 KOps/s	21.8675 KOps/s	$\textbf{\color{#d91a1a}-8.70\%}$
test_add_td	0.1890ms	89.8703μs	11.1272 KOps/s	11.9204 KOps/s	$\textbf{\color{#d91a1a}-6.65\%}$
test_compile_add_one_nested[tensordict-compile]	0.1217ms	55.3685μs	18.0608 KOps/s	18.4384 KOps/s	$\color{#d91a1a}-2.05\%$
test_compile_add_one_nested[tensordict-eager]	0.4200ms	0.1934ms	5.1713 KOps/s	5.2596 KOps/s	$\color{#d91a1a}-1.68\%$
test_compile_add_one_nested[pytree-compile]	0.2098ms	55.9354μs	17.8778 KOps/s	18.8698 KOps/s	$\textbf{\color{#d91a1a}-5.26\%}$
test_compile_add_one_nested[pytree-eager]	0.2963ms	0.1487ms	6.7263 KOps/s	7.0572 KOps/s	$\color{#d91a1a}-4.69\%$
test_compile_copy_nested[tensordict-compile]	66.2840μs	20.9232μs	47.7938 KOps/s	48.1187 KOps/s	$\color{#d91a1a}-0.68\%$
test_compile_copy_nested[tensordict-eager]	0.1893ms	65.2790μs	15.3189 KOps/s	15.7408 KOps/s	$\color{#d91a1a}-2.68\%$
test_compile_copy_nested[pytree-compile]	0.1803ms	79.7734μs	12.5355 KOps/s	12.7461 KOps/s	$\color{#d91a1a}-1.65\%$
test_compile_copy_nested[pytree-eager]	0.1349ms	70.0495μs	14.2756 KOps/s	14.2735 KOps/s	$\color{#35bf28}+0.01\%$
test_compile_add_one_flat[tensordict-compile]	0.2966ms	0.1800ms	5.5562 KOps/s	5.7898 KOps/s	$\color{#d91a1a}-4.03\%$
test_compile_add_one_flat[tensordict-eager]	0.2966ms	0.1977ms	5.0582 KOps/s	5.1643 KOps/s	$\color{#d91a1a}-2.06\%$
test_compile_add_one_flat[tensorclass-compile]	0.1101ms	40.2565μs	24.8407 KOps/s	26.2943 KOps/s	$\textbf{\color{#d91a1a}-5.53\%}$
test_compile_add_one_flat[tensorclass-eager]	1.1929ms	72.3399μs	13.8236 KOps/s	14.2006 KOps/s	$\color{#d91a1a}-2.65\%$
test_compile_add_one_flat[pytree-compile]	0.3034ms	0.1773ms	5.6393 KOps/s	5.8604 KOps/s	$\color{#d91a1a}-3.77\%$
test_compile_add_one_flat[pytree-eager]	0.3842ms	0.3038ms	3.2912 KOps/s	3.4493 KOps/s	$\color{#d91a1a}-4.58\%$
test_compile_add_self_flat[tensordict-eager]	0.2989ms	0.2111ms	4.7371 KOps/s	4.8729 KOps/s	$\color{#d91a1a}-2.79\%$
test_compile_add_self_flat[tensordict-compile]	0.2833ms	0.1803ms	5.5471 KOps/s	5.7067 KOps/s	$\color{#d91a1a}-2.80\%$
test_compile_add_self_flat[tensorclass-eager]	0.1932ms	63.7792μs	15.6791 KOps/s	15.8233 KOps/s	$\color{#d91a1a}-0.91\%$
test_compile_add_self_flat[tensorclass-compile]	79.1590μs	41.0246μs	24.3756 KOps/s	25.0798 KOps/s	$\color{#d91a1a}-2.81\%$
test_compile_add_self_flat[pytree-eager]	0.3642ms	0.2513ms	3.9795 KOps/s	4.2061 KOps/s	$\textbf{\color{#d91a1a}-5.39\%}$
test_compile_add_self_flat[pytree-compile]	0.3010ms	0.1780ms	5.6187 KOps/s	5.8169 KOps/s	$\color{#d91a1a}-3.41\%$
test_compile_copy_flat[tensordict-compile]	0.2701ms	0.1101ms	9.0828 KOps/s	9.3687 KOps/s	$\color{#d91a1a}-3.05\%$
test_compile_copy_flat[tensordict-eager]	0.1192ms	56.5721μs	17.6766 KOps/s	18.1291 KOps/s	$\color{#d91a1a}-2.50\%$
test_compile_copy_flat[pytree-compile]	0.1531ms	79.2012μs	12.6261 KOps/s	11.9595 KOps/s	$\textbf{\color{#35bf28}+5.57\%}$
test_compile_copy_flat[pytree-eager]	0.1380ms	70.2825μs	14.2283 KOps/s	14.0584 KOps/s	$\color{#35bf28}+1.21\%$
test_compile_assign_and_add[tensordict-compile]	0.2950ms	0.1918ms	5.2150 KOps/s	5.2640 KOps/s	$\color{#d91a1a}-0.93\%$
test_compile_assign_and_add[tensordict-eager]	3.8921ms	1.6936ms	590.4551 Ops/s	613.2494 Ops/s	$\color{#d91a1a}-3.72\%$
test_compile_assign_and_add[pytree-compile]	0.3763ms	0.1881ms	5.3155 KOps/s	5.3266 KOps/s	$\color{#d91a1a}-0.21\%$
test_compile_assign_and_add[pytree-eager]	1.4652ms	1.1297ms	885.1958 Ops/s	935.5714 Ops/s	$\textbf{\color{#d91a1a}-5.38\%}$
test_compile_assign_and_add_stack[compile]	0.6248ms	0.4055ms	2.4662 KOps/s	2.4109 KOps/s	$\color{#35bf28}+2.29\%$
test_compile_assign_and_add_stack[eager]	4.3552ms	4.0027ms	249.8329 Ops/s	260.9257 Ops/s	$\color{#d91a1a}-4.25\%$
test_compile_indexing[tensor-tensordict-compile]	91.5020μs	33.2263μs	30.0966 KOps/s	30.6204 KOps/s	$\color{#d91a1a}-1.71\%$
test_compile_indexing[tensor-tensordict-eager]	1.4968ms	50.0832μs	19.9668 KOps/s	21.0963 KOps/s	$\textbf{\color{#d91a1a}-5.35\%}$
test_compile_indexing[tensor-tensorclass-compile]	86.2520μs	29.6650μs	33.7098 KOps/s	36.9389 KOps/s	$\textbf{\color{#d91a1a}-8.74\%}$
test_compile_indexing[tensor-tensorclass-eager]	92.1730μs	30.7161μs	32.5562 KOps/s	32.7913 KOps/s	$\color{#d91a1a}-0.72\%$
test_compile_indexing[tensor-pytree-compile]	64.4310μs	29.8210μs	33.5334 KOps/s	36.7996 KOps/s	$\textbf{\color{#d91a1a}-8.88\%}$
test_compile_indexing[tensor-pytree-eager]	83.2860μs	31.0768μs	32.1783 KOps/s	32.7270 KOps/s	$\color{#d91a1a}-1.68\%$
test_compile_indexing[slice-tensordict-compile]	0.1963ms	73.2297μs	13.6557 KOps/s	13.7970 KOps/s	$\color{#d91a1a}-1.02\%$
test_compile_indexing[slice-tensordict-eager]	0.7271ms	28.9713μs	34.5169 KOps/s	36.0616 KOps/s	$\color{#d91a1a}-4.28\%$
test_compile_indexing[slice-tensorclass-compile]	0.1366ms	69.3495μs	14.4197 KOps/s	14.9320 KOps/s	$\color{#d91a1a}-3.43\%$
test_compile_indexing[slice-tensorclass-eager]	67.2360μs	25.3465μs	39.4532 KOps/s	40.8123 KOps/s	$\color{#d91a1a}-3.33\%$
test_compile_indexing[slice-pytree-compile]	0.1310ms	68.7588μs	14.5436 KOps/s	15.1598 KOps/s	$\color{#d91a1a}-4.06\%$
test_compile_indexing[slice-pytree-eager]	72.6960μs	25.0241μs	39.9615 KOps/s	39.7813 KOps/s	$\color{#35bf28}+0.45\%$
test_compile_indexing[int-tensordict-compile]	0.1385ms	73.6514μs	13.5775 KOps/s	13.9787 KOps/s	$\color{#d91a1a}-2.87\%$
test_compile_indexing[int-tensordict-eager]	1.0381ms	29.0201μs	34.4589 KOps/s	36.4577 KOps/s	$\textbf{\color{#d91a1a}-5.48\%}$
test_compile_indexing[int-tensorclass-compile]	0.1635ms	68.9071μs	14.5123 KOps/s	14.6545 KOps/s	$\color{#d91a1a}-0.97\%$
test_compile_indexing[int-tensorclass-eager]	0.1408ms	25.5335μs	39.1643 KOps/s	40.5168 KOps/s	$\color{#d91a1a}-3.34\%$
test_compile_indexing[int-pytree-compile]	0.1701ms	69.3363μs	14.4225 KOps/s	14.9700 KOps/s	$\color{#d91a1a}-3.66\%$
test_compile_indexing[int-pytree-eager]	73.6760μs	24.6729μs	40.5303 KOps/s	39.5369 KOps/s	$\color{#35bf28}+2.51\%$
test_mod_add[eager]	76.8140μs	27.0835μs	36.9228 KOps/s	40.0176 KOps/s	$\textbf{\color{#d91a1a}-7.73\%}$
test_mod_add[compile]	87.4340μs	38.0226μs	26.3002 KOps/s	27.6859 KOps/s	$\textbf{\color{#d91a1a}-5.01\%}$
test_mod_add[compile-overhead]	81.1020μs	37.3596μs	26.7669 KOps/s	27.3214 KOps/s	$\color{#d91a1a}-2.03\%$
test_mod_wrap[eager]	0.4138ms	0.2143ms	4.6659 KOps/s	4.4457 KOps/s	$\color{#35bf28}+4.95\%$
test_mod_wrap[compile]	2.0368ms	0.2322ms	4.3073 KOps/s	4.2833 KOps/s	$\color{#35bf28}+0.56\%$
test_mod_wrap[compile-overhead]	0.4546ms	0.2266ms	4.4136 KOps/s	4.2088 KOps/s	$\color{#35bf28}+4.87\%$
test_mod_wrap_and_backward[eager]	13.8402ms	11.4367ms	87.4379 Ops/s	91.3787 Ops/s	$\color{#d91a1a}-4.31\%$
test_mod_wrap_and_backward[compile]	13.6393ms	11.6260ms	86.0140 Ops/s	92.0014 Ops/s	$\textbf{\color{#d91a1a}-6.51\%}$
test_mod_wrap_and_backward[compile-overhead]	14.9062ms	12.2259ms	81.7937 Ops/s	91.7629 Ops/s	$\textbf{\color{#d91a1a}-10.86\%}$
test_seq_add[eager]	0.1702ms	92.2231μs	10.8433 KOps/s	11.5547 KOps/s	$\textbf{\color{#d91a1a}-6.16\%}$
test_seq_add[compile]	0.1560ms	59.7094μs	16.7478 KOps/s	16.6438 KOps/s	$\color{#35bf28}+0.62\%$
test_seq_add[compile-overhead]	0.1742ms	59.7875μs	16.7259 KOps/s	16.8632 KOps/s	$\color{#d91a1a}-0.81\%$
test_seq_wrap[eager]	0.5076ms	0.3840ms	2.6044 KOps/s	2.6840 KOps/s	$\color{#d91a1a}-2.97\%$
test_seq_wrap[compile]	0.5511ms	0.2626ms	3.8076 KOps/s	3.7932 KOps/s	$\color{#35bf28}+0.38\%$
test_seq_wrap[compile-overhead]	0.5157ms	0.2630ms	3.8025 KOps/s	3.7904 KOps/s	$\color{#35bf28}+0.32\%$
test_func_call_runtime[False-eager]	0.7349ms	0.5507ms	1.8159 KOps/s	1.9015 KOps/s	$\color{#d91a1a}-4.50\%$
test_func_call_runtime[False-compile]	0.9221ms	0.4982ms	2.0074 KOps/s	2.0418 KOps/s	$\color{#d91a1a}-1.69\%$
test_func_call_runtime[False-compile-overhead]	0.6113ms	0.4929ms	2.0288 KOps/s	2.0451 KOps/s	$\color{#d91a1a}-0.80\%$
test_func_call_runtime[True-eager]	1.2247ms	0.7743ms	1.2914 KOps/s	1.3263 KOps/s	$\color{#d91a1a}-2.63\%$
test_func_call_runtime[True-compile]	0.5936ms	0.5094ms	1.9631 KOps/s	1.9936 KOps/s	$\color{#d91a1a}-1.53\%$
test_func_call_runtime[True-compile-overhead]	0.9091ms	0.5112ms	1.9562 KOps/s	1.9680 KOps/s	$\color{#d91a1a}-0.60\%$
test_func_call_cm_runtime[False-eager]	1.5901ms	0.5640ms	1.7730 KOps/s	1.9219 KOps/s	$\textbf{\color{#d91a1a}-7.75\%}$
test_func_call_cm_runtime[False-compile]	0.5994ms	0.4930ms	2.0282 KOps/s	2.0502 KOps/s	$\color{#d91a1a}-1.07\%$
test_func_call_cm_runtime[False-compile-overhead]	1.0161ms	0.4957ms	2.0175 KOps/s	2.0359 KOps/s	$\color{#d91a1a}-0.90\%$
test_func_call_cm_runtime[True-eager]	1.1310ms	0.8992ms	1.1121 KOps/s	1.1133 KOps/s	$\color{#d91a1a}-0.11\%$
test_func_call_cm_runtime[True-compile]	1.1447ms	0.8482ms	1.1790 KOps/s	1.1823 KOps/s	$\color{#d91a1a}-0.28\%$
test_func_call_cm_runtime[True-compile-overhead]	1.0676ms	0.8460ms	1.1821 KOps/s	1.1892 KOps/s	$\color{#d91a1a}-0.60\%$
test_distributed	0.2667ms	0.1294ms	7.7303 KOps/s	7.4154 KOps/s	$\color{#35bf28}+4.25\%$
test_tdmodule	93.4250μs	18.2857μs	54.6875 KOps/s	57.9267 KOps/s	$\textbf{\color{#d91a1a}-5.59\%}$
test_tdmodule_dispatch	62.8180μs	37.4287μs	26.7175 KOps/s	27.9936 KOps/s	$\color{#d91a1a}-4.56\%$
test_tdseq	35.6060μs	20.0617μs	49.8462 KOps/s	52.3524 KOps/s	$\color{#d91a1a}-4.79\%$
test_tdseq_dispatch	84.7080μs	41.8683μs	23.8844 KOps/s	24.9460 KOps/s	$\color{#d91a1a}-4.26\%$
test_instantiation_functorch	1.9260ms	1.6675ms	599.6831 Ops/s	602.4616 Ops/s	$\color{#d91a1a}-0.46\%$
test_instantiation_td	1.9740ms	1.2058ms	829.3365 Ops/s	827.4871 Ops/s	$\color{#35bf28}+0.22\%$
test_exec_functorch	0.4057ms	0.1843ms	5.4260 KOps/s	5.5778 KOps/s	$\color{#d91a1a}-2.72\%$
test_exec_functional_call	0.3280ms	0.1737ms	5.7556 KOps/s	5.8513 KOps/s	$\color{#d91a1a}-1.64\%$
test_exec_td	0.2827ms	0.1745ms	5.7307 KOps/s	5.7649 KOps/s	$\color{#d91a1a}-0.59\%$
test_exec_td_decorator	0.8630ms	0.2285ms	4.3754 KOps/s	4.4504 KOps/s	$\color{#d91a1a}-1.68\%$
test_vmap_mlp_speed[True-True]	0.8393ms	0.5939ms	1.6839 KOps/s	1.7322 KOps/s	$\color{#d91a1a}-2.79\%$
test_vmap_mlp_speed[True-False]	0.8583ms	0.5863ms	1.7056 KOps/s	1.7624 KOps/s	$\color{#d91a1a}-3.22\%$
test_vmap_mlp_speed[False-True]	0.7927ms	0.4878ms	2.0499 KOps/s	2.1271 KOps/s	$\color{#d91a1a}-3.63\%$
test_vmap_mlp_speed[False-False]	0.6706ms	0.4882ms	2.0484 KOps/s	2.1316 KOps/s	$\color{#d91a1a}-3.90\%$
test_vmap_mlp_speed_decorator[True-True]	0.9491ms	0.6457ms	1.5488 KOps/s	1.5985 KOps/s	$\color{#d91a1a}-3.11\%$
test_vmap_mlp_speed_decorator[True-False]	1.0580ms	0.6451ms	1.5501 KOps/s	1.6083 KOps/s	$\color{#d91a1a}-3.62\%$
test_vmap_mlp_speed_decorator[False-True]	0.8574ms	0.5314ms	1.8819 KOps/s	1.9345 KOps/s	$\color{#d91a1a}-2.72\%$
test_vmap_mlp_speed_decorator[False-False]	0.7380ms	0.5316ms	1.8811 KOps/s	1.9440 KOps/s	$\color{#d91a1a}-3.24\%$
test_to_module_speed[True]	2.2819ms	1.3249ms	754.7476 Ops/s	754.8753 Ops/s	$\color{#d91a1a}-0.02\%$
test_to_module_speed[False]	2.1091ms	1.2911ms	774.5321 Ops/s	770.4757 Ops/s	$\color{#35bf28}+0.53\%$
test_tc_init	77.6560μs	46.1553μs	21.6660 KOps/s	22.0187 KOps/s	$\color{#d91a1a}-1.60\%$
test_tc_init_nested	0.1738ms	94.8299μs	10.5452 KOps/s	10.9411 KOps/s	$\color{#d91a1a}-3.62\%$
test_tc_first_layer_tensor	23.8140μs	1.4859μs	672.9845 KOps/s	690.6866 KOps/s	$\color{#d91a1a}-2.56\%$
test_tc_first_layer_nontensor	28.7830μs	4.2970μs	232.7203 KOps/s	236.0247 KOps/s	$\color{#d91a1a}-1.40\%$
test_tc_second_layer_tensor	39.8640μs	2.6909μs	371.6186 KOps/s	369.1007 KOps/s	$\color{#35bf28}+0.68\%$
test_tc_second_layer_nontensor	29.0140μs	5.5165μs	181.2747 KOps/s	181.5853 KOps/s	$\color{#d91a1a}-0.17\%$
test_unbind	0.4549s	17.2802ms	57.8698 Ops/s	67.5783 Ops/s	$\textbf{\color{#d91a1a}-14.37\%}$
test_full_like	17.3406ms	11.3495ms	88.1096 Ops/s	141.6104 Ops/s	$\textbf{\color{#d91a1a}-37.78\%}$
test_zeros_like	15.4088ms	7.1767ms	139.3399 Ops/s	131.9552 Ops/s	$\textbf{\color{#35bf28}+5.60\%}$
test_ones_like	12.1370ms	7.4477ms	134.2704 Ops/s	128.3583 Ops/s	$\color{#35bf28}+4.61\%$
test_clone	16.9052ms	8.8800ms	112.6130 Ops/s	102.5625 Ops/s	$\textbf{\color{#35bf28}+9.80\%}$
test_squeeze	73.4470μs	13.4700μs	74.2388 KOps/s	76.6820 KOps/s	$\color{#d91a1a}-3.19\%$
test_unsqueeze	0.3276ms	96.2925μs	10.3850 KOps/s	10.7905 KOps/s	$\color{#d91a1a}-3.76\%$
test_split	0.3920ms	0.2039ms	4.9050 KOps/s	4.9867 KOps/s	$\color{#d91a1a}-1.64\%$
test_permute	0.3700ms	0.2205ms	4.5355 KOps/s	4.5835 KOps/s	$\color{#d91a1a}-1.05\%$
test_stack	29.5706ms	24.1686ms	41.3760 Ops/s	41.6282 Ops/s	$\color{#d91a1a}-0.61\%$
test_cat	29.5340ms	24.2424ms	41.2500 Ops/s	41.5835 Ops/s	$\color{#d91a1a}-0.80\%$

[github-actions]

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

Total Benchmarks: 225. Improved: $\large\color{#35bf28}34$. Worsened: $\large\color{#d91a1a}9$.

Expand to view detailed results

Name	Max	Mean	Ops	Ops on Repo HEAD	Change
test_plain_set_nested	0.2985ms	16.5736μs	60.3371 KOps/s	55.5278 KOps/s	$\textbf{\color{#35bf28}+8.66\%}$
test_plain_set_stack_nested	34.9100μs	16.6564μs	60.0369 KOps/s	54.9337 KOps/s	$\textbf{\color{#35bf28}+9.29\%}$
test_plain_set_nested_inplace	0.2094ms	17.8321μs	56.0788 KOps/s	52.2699 KOps/s	$\textbf{\color{#35bf28}+7.29\%}$
test_plain_set_stack_nested_inplace	43.7110μs	17.7074μs	56.4734 KOps/s	50.8548 KOps/s	$\textbf{\color{#35bf28}+11.05\%}$
test_items	0.1936ms	4.7056μs	212.5136 KOps/s	212.4430 KOps/s	$\color{#35bf28}+0.03\%$
test_items_nested	0.5555ms	0.3620ms	2.7621 KOps/s	2.6930 KOps/s	$\color{#35bf28}+2.56\%$
test_items_nested_locked	0.5455ms	0.3636ms	2.7501 KOps/s	2.6494 KOps/s	$\color{#35bf28}+3.80\%$
test_items_nested_leaf	0.1123ms	84.2212μs	11.8735 KOps/s	11.6745 KOps/s	$\color{#35bf28}+1.70\%$
test_items_stack_nested	0.5583ms	0.3606ms	2.7731 KOps/s	2.7039 KOps/s	$\color{#35bf28}+2.56\%$
test_items_stack_nested_leaf	0.2812ms	85.7132μs	11.6668 KOps/s	11.5298 KOps/s	$\color{#35bf28}+1.19\%$
test_items_stack_nested_locked	0.5491ms	0.3649ms	2.7402 KOps/s	2.7010 KOps/s	$\color{#35bf28}+1.45\%$
test_keys	0.1883ms	4.3702μs	228.8200 KOps/s	227.7542 KOps/s	$\color{#35bf28}+0.47\%$
test_keys_nested	0.2577ms	66.8962μs	14.9485 KOps/s	15.2018 KOps/s	$\color{#d91a1a}-1.67\%$
test_keys_nested_locked	0.9129ms	72.3059μs	13.8301 KOps/s	13.8115 KOps/s	$\color{#35bf28}+0.14\%$
test_keys_nested_leaf	74.9310μs	57.2151μs	17.4779 KOps/s	17.4365 KOps/s	$\color{#35bf28}+0.24\%$
test_keys_stack_nested	0.2620ms	67.3925μs	14.8385 KOps/s	14.9438 KOps/s	$\color{#d91a1a}-0.70\%$
test_keys_stack_nested_leaf	0.2537ms	57.6332μs	17.3511 KOps/s	17.1302 KOps/s	$\color{#35bf28}+1.29\%$
test_keys_stack_nested_locked	0.2674ms	72.7481μs	13.7461 KOps/s	13.6921 KOps/s	$\color{#35bf28}+0.39\%$
test_values	63.9513μs	1.7771μs	562.7289 KOps/s	561.0806 KOps/s	$\color{#35bf28}+0.29\%$
test_values_nested	0.2169ms	33.7256μs	29.6510 KOps/s	29.4816 KOps/s	$\color{#35bf28}+0.57\%$
test_values_nested_locked	0.2204ms	35.6307μs	28.0657 KOps/s	27.6198 KOps/s	$\color{#35bf28}+1.61\%$
test_values_nested_leaf	52.7610μs	30.0000μs	33.3333 KOps/s	32.7862 KOps/s	$\color{#35bf28}+1.67\%$
test_values_stack_nested	0.2274ms	34.4075μs	29.0635 KOps/s	29.1783 KOps/s	$\color{#d91a1a}-0.39\%$
test_values_stack_nested_leaf	0.2188ms	30.3418μs	32.9579 KOps/s	32.5252 KOps/s	$\color{#35bf28}+1.33\%$
test_values_stack_nested_locked	0.2231ms	36.2816μs	27.5622 KOps/s	27.6465 KOps/s	$\color{#d91a1a}-0.30\%$
test_membership	1.2836μs	0.5389μs	1.8557 MOps/s	1.7751 MOps/s	$\color{#35bf28}+4.54\%$
test_membership_nested	0.1039ms	1.9581μs	510.6983 KOps/s	511.4504 KOps/s	$\color{#d91a1a}-0.15\%$
test_membership_nested_leaf	99.9565μs	1.9543μs	511.6888 KOps/s	504.1702 KOps/s	$\color{#35bf28}+1.49\%$
test_membership_stacked_nested	26.9600μs	2.0045μs	498.8844 KOps/s	504.0078 KOps/s	$\color{#d91a1a}-1.02\%$
test_membership_stacked_nested_leaf	17.6310μs	2.0085μs	497.8840 KOps/s	497.3662 KOps/s	$\color{#35bf28}+0.10\%$
test_membership_nested_last	0.1941ms	2.9204μs	342.4173 KOps/s	340.4623 KOps/s	$\color{#35bf28}+0.57\%$
test_membership_nested_leaf_last	16.1000μs	2.9445μs	339.6205 KOps/s	340.0258 KOps/s	$\color{#d91a1a}-0.12\%$
test_membership_stacked_nested_last	27.6310μs	2.9637μs	337.4202 KOps/s	343.4002 KOps/s	$\color{#d91a1a}-1.74\%$
test_membership_stacked_nested_leaf_last	0.2000ms	2.9322μs	341.0416 KOps/s	341.7601 KOps/s	$\color{#d91a1a}-0.21\%$
test_nested_getleaf	0.1986ms	7.9156μs	126.3327 KOps/s	126.4926 KOps/s	$\color{#d91a1a}-0.13\%$
test_nested_get	22.5800μs	7.4479μs	134.2667 KOps/s	133.5551 KOps/s	$\color{#35bf28}+0.53\%$
test_stacked_getleaf	0.2035ms	7.9687μs	125.4914 KOps/s	125.6733 KOps/s	$\color{#d91a1a}-0.14\%$
test_stacked_get	21.5700μs	7.4852μs	133.5974 KOps/s	135.2651 KOps/s	$\color{#d91a1a}-1.23\%$
test_nested_getitemleaf	0.1976ms	8.1207μs	123.1421 KOps/s	121.8433 KOps/s	$\color{#35bf28}+1.07\%$
test_nested_getitem	21.2100μs	7.7050μs	129.7862 KOps/s	130.1232 KOps/s	$\color{#d91a1a}-0.26\%$
test_stacked_getitemleaf	30.7000μs	8.2027μs	121.9114 KOps/s	121.9459 KOps/s	$\color{#d91a1a}-0.03\%$
test_stacked_getitem	0.2018ms	7.7070μs	129.7521 KOps/s	130.0029 KOps/s	$\color{#d91a1a}-0.19\%$
test_lock_nested	1.2952ms	0.4824ms	2.0731 KOps/s	2.0534 KOps/s	$\color{#35bf28}+0.96\%$
test_lock_stack_nested	0.5010ms	0.4440ms	2.2525 KOps/s	2.2398 KOps/s	$\color{#35bf28}+0.57\%$
test_unlock_nested	0.9116ms	0.4025ms	2.4842 KOps/s	2.4594 KOps/s	$\color{#35bf28}+1.01\%$
test_unlock_stack_nested	0.5014ms	0.3642ms	2.7458 KOps/s	2.7322 KOps/s	$\color{#35bf28}+0.50\%$
test_flatten_speed	0.3118ms	0.1068ms	9.3661 KOps/s	9.5279 KOps/s	$\color{#d91a1a}-1.70\%$
test_unflatten_speed	0.5123ms	0.3190ms	3.1352 KOps/s	3.1369 KOps/s	$\color{#d91a1a}-0.06\%$
test_common_ops	1.7399ms	1.4389ms	694.9654 Ops/s	691.7279 Ops/s	$\color{#35bf28}+0.47\%$
test_creation	19.7110μs	1.6578μs	603.2191 KOps/s	592.2792 KOps/s	$\color{#35bf28}+1.85\%$
test_creation_empty	43.3610μs	16.3219μs	61.2673 KOps/s	51.9830 KOps/s	$\textbf{\color{#35bf28}+17.86\%}$
test_creation_nested_1	35.7010μs	18.6457μs	53.6315 KOps/s	47.5933 KOps/s	$\textbf{\color{#35bf28}+12.69\%}$
test_creation_nested_2	0.2233ms	21.5330μs	46.4403 KOps/s	40.2115 KOps/s	$\textbf{\color{#35bf28}+15.49\%}$
test_clone	54.0510μs	32.8600μs	30.4322 KOps/s	30.1221 KOps/s	$\color{#35bf28}+1.03\%$
test_getitem[int]	1.0586ms	18.2880μs	54.6807 KOps/s	53.0161 KOps/s	$\color{#35bf28}+3.14\%$
test_getitem[slice_int]	0.2430ms	30.3968μs	32.8982 KOps/s	32.1034 KOps/s	$\color{#35bf28}+2.48\%$
test_getitem[range]	0.2904ms	0.1184ms	8.4469 KOps/s	8.5596 KOps/s	$\color{#d91a1a}-1.32\%$
test_getitem[tuple]	0.1444ms	26.6580μs	37.5122 KOps/s	36.2181 KOps/s	$\color{#35bf28}+3.57\%$
test_getitem[list]	0.4035ms	0.1074ms	9.3119 KOps/s	8.7844 KOps/s	$\textbf{\color{#35bf28}+6.01\%}$
test_setitem_dim[int]	78.2420μs	55.8493μs	17.9053 KOps/s	15.7706 KOps/s	$\textbf{\color{#35bf28}+13.54\%}$
test_setitem_dim[slice_int]	0.1021ms	79.6506μs	12.5548 KOps/s	11.6263 KOps/s	$\textbf{\color{#35bf28}+7.99\%}$
test_setitem_dim[range]	0.1671ms	0.1432ms	6.9857 KOps/s	6.4248 KOps/s	$\textbf{\color{#35bf28}+8.73\%}$
test_setitem_dim[tuple]	0.2227ms	73.6526μs	13.5773 KOps/s	12.2162 KOps/s	$\textbf{\color{#35bf28}+11.14\%}$
test_setitem	0.1914ms	45.4787μs	21.9883 KOps/s	20.5316 KOps/s	$\textbf{\color{#35bf28}+7.10\%}$
test_set	0.2389ms	44.8307μs	22.3062 KOps/s	19.8443 KOps/s	$\textbf{\color{#35bf28}+12.41\%}$
test_set_shared	0.3922ms	56.5816μs	17.6736 KOps/s	17.2435 KOps/s	$\color{#35bf28}+2.49\%$
test_update	0.2032ms	53.3731μs	18.7360 KOps/s	17.4172 KOps/s	$\textbf{\color{#35bf28}+7.57\%}$
test_update_nested	0.2590ms	61.3139μs	16.3095 KOps/s	14.4105 KOps/s	$\textbf{\color{#35bf28}+13.18\%}$
test_update__nested	0.2675ms	65.6733μs	15.2269 KOps/s	14.2046 KOps/s	$\textbf{\color{#35bf28}+7.20\%}$
test_set_nested	0.2478ms	47.8081μs	20.9170 KOps/s	19.5774 KOps/s	$\textbf{\color{#35bf28}+6.84\%}$
test_set_nested_new	0.2033ms	51.8052μs	19.3031 KOps/s	17.5298 KOps/s	$\textbf{\color{#35bf28}+10.12\%}$
test_select	0.2682ms	66.2801μs	15.0875 KOps/s	13.7352 KOps/s	$\textbf{\color{#35bf28}+9.85\%}$
test_select_nested	74.4010μs	51.1689μs	19.5431 KOps/s	19.2942 KOps/s	$\color{#35bf28}+1.29\%$
test_exclude_nested	0.2569ms	69.9652μs	14.2928 KOps/s	14.3221 KOps/s	$\color{#d91a1a}-0.20\%$
test_empty[True]	0.4803ms	0.2838ms	3.5238 KOps/s	3.4937 KOps/s	$\color{#35bf28}+0.86\%$
test_empty[False]	19.2453μs	0.8658μs	1.1550 MOps/s	1.1052 MOps/s	$\color{#35bf28}+4.51\%$
test_to	49.1610μs	28.4711μs	35.1233 KOps/s	35.4237 KOps/s	$\color{#d91a1a}-0.85\%$
test_to_nonblocking	0.2264ms	28.3789μs	35.2375 KOps/s	37.3586 KOps/s	$\textbf{\color{#d91a1a}-5.68\%}$
test_unbind_speed	1.1815ms	0.3165ms	3.1591 KOps/s	3.1614 KOps/s	$\color{#d91a1a}-0.07\%$
test_unbind_speed_stack0	0.5105ms	0.3169ms	3.1553 KOps/s	3.1718 KOps/s	$\color{#d91a1a}-0.52\%$
test_unbind_speed_stack1	89.6224ms	0.7871ms	1.2704 KOps/s	1.3758 KOps/s	$\textbf{\color{#d91a1a}-7.66\%}$
test_split	92.4675ms	2.4759ms	403.9010 Ops/s	403.7185 Ops/s	$\color{#35bf28}+0.05\%$
test_chunk	92.7088ms	2.4837ms	402.6261 Ops/s	404.6127 Ops/s	$\color{#d91a1a}-0.49\%$
test_creation[device0]	0.2493ms	0.1073ms	9.3208 KOps/s	9.3143 KOps/s	$\color{#35bf28}+0.07\%$
test_creation_from_tensor	0.3158ms	0.1047ms	9.5501 KOps/s	9.2820 KOps/s	$\color{#35bf28}+2.89\%$
test_add_one[memmap_tensor0]	0.1288ms	10.0660μs	99.3444 KOps/s	99.1214 KOps/s	$\color{#35bf28}+0.22\%$
test_contiguous[memmap_tensor0]	0.1904ms	2.3505μs	425.4451 KOps/s	421.8040 KOps/s	$\color{#35bf28}+0.86\%$
test_stack[memmap_tensor0]	33.6510μs	7.2983μs	137.0184 KOps/s	130.6959 KOps/s	$\color{#35bf28}+4.84\%$
test_memmaptd_index	1.2428ms	0.4770ms	2.0966 KOps/s	2.0861 KOps/s	$\color{#35bf28}+0.51\%$
test_memmaptd_index_astensor	0.7904ms	0.5431ms	1.8412 KOps/s	1.8512 KOps/s	$\color{#d91a1a}-0.54\%$
test_memmaptd_index_op	1.5526ms	1.1397ms	877.4285 Ops/s	839.8908 Ops/s	$\color{#35bf28}+4.47\%$
test_serialize_model	93.7375ms	89.2760ms	11.2012 Ops/s	10.8196 Ops/s	$\color{#35bf28}+3.53\%$
test_serialize_model_pickle	1.3491s	1.2362s	0.8089 Ops/s	0.8083 Ops/s	$\color{#35bf28}+0.07\%$
test_serialize_weights	0.1826s	96.6039ms	10.3515 Ops/s	9.7019 Ops/s	$\textbf{\color{#35bf28}+6.70\%}$
test_serialize_weights_returnearly	0.2732s	67.8955ms	14.7285 Ops/s	14.9354 Ops/s	$\color{#d91a1a}-1.39\%$
test_serialize_weights_pickle	1.3550s	1.2370s	0.8084 Ops/s	0.8029 Ops/s	$\color{#35bf28}+0.69\%$
test_reshape_pytree	0.1197ms	39.8769μs	25.0772 KOps/s	24.4544 KOps/s	$\color{#35bf28}+2.55\%$
test_reshape_td	0.1482ms	50.0326μs	19.9870 KOps/s	20.9994 KOps/s	$\color{#d91a1a}-4.82\%$
test_view_pytree	85.0420μs	41.0625μs	24.3531 KOps/s	24.7627 KOps/s	$\color{#d91a1a}-1.65\%$
test_view_td	0.1963ms	51.9561μs	19.2470 KOps/s	18.7668 KOps/s	$\color{#35bf28}+2.56\%$
test_unbind_pytree	0.1782ms	39.4786μs	25.3302 KOps/s	25.0898 KOps/s	$\color{#35bf28}+0.96\%$
test_unbind_td	0.3942ms	48.4505μs	20.6396 KOps/s	20.5387 KOps/s	$\color{#35bf28}+0.49\%$
test_split_pytree	0.2301ms	55.1450μs	18.1340 KOps/s	18.6660 KOps/s	$\color{#d91a1a}-2.85\%$
test_split_td	0.5602ms	64.4068μs	15.5263 KOps/s	15.7340 KOps/s	$\color{#d91a1a}-1.32\%$
test_add_pytree	0.2356ms	66.7445μs	14.9825 KOps/s	15.9179 KOps/s	$\textbf{\color{#d91a1a}-5.88\%}$
test_add_td	0.2743ms	98.7287μs	10.1288 KOps/s	9.3595 KOps/s	$\textbf{\color{#35bf28}+8.22\%}$
test_compile_add_one_nested[tensordict-compile]	0.4234ms	0.2238ms	4.4681 KOps/s	4.4062 KOps/s	$\color{#35bf28}+1.40\%$
test_compile_add_one_nested[tensordict-eager]	0.3220ms	0.1788ms	5.5918 KOps/s	5.5808 KOps/s	$\color{#35bf28}+0.20\%$
test_compile_add_one_nested[pytree-compile]	0.3373ms	0.1554ms	6.4332 KOps/s	6.4438 KOps/s	$\color{#d91a1a}-0.16\%$
test_compile_add_one_nested[pytree-eager]	0.3841ms	0.2025ms	4.9371 KOps/s	4.7727 KOps/s	$\color{#35bf28}+3.45\%$
test_compile_copy_nested[tensordict-compile]	0.2203ms	23.3914μs	42.7508 KOps/s	41.8699 KOps/s	$\color{#35bf28}+2.10\%$
test_compile_copy_nested[tensordict-eager]	87.3710μs	50.0557μs	19.9778 KOps/s	19.8509 KOps/s	$\color{#35bf28}+0.64\%$
test_compile_copy_nested[pytree-compile]	0.2791ms	74.5599μs	13.4120 KOps/s	13.4291 KOps/s	$\color{#d91a1a}-0.13\%$
test_compile_copy_nested[pytree-eager]	85.9410μs	59.8062μs	16.7207 KOps/s	16.7698 KOps/s	$\color{#d91a1a}-0.29\%$
test_compile_add_one_flat[tensordict-compile]	0.4744ms	0.3438ms	2.9091 KOps/s	2.8936 KOps/s	$\color{#35bf28}+0.53\%$
test_compile_add_one_flat[tensordict-eager]	0.3701ms	0.2257ms	4.4316 KOps/s	4.4517 KOps/s	$\color{#d91a1a}-0.45\%$
test_compile_add_one_flat[tensorclass-compile]	0.2899ms	0.1386ms	7.2143 KOps/s	6.9034 KOps/s	$\color{#35bf28}+4.50\%$
test_compile_add_one_flat[tensorclass-eager]	0.2058ms	64.7664μs	15.4401 KOps/s	15.2622 KOps/s	$\color{#35bf28}+1.17\%$
test_compile_add_one_flat[pytree-compile]	0.4503ms	0.3428ms	2.9174 KOps/s	2.9022 KOps/s	$\color{#35bf28}+0.52\%$
test_compile_add_one_flat[pytree-eager]	0.8195ms	0.6725ms	1.4869 KOps/s	1.4398 KOps/s	$\color{#35bf28}+3.27\%$
test_compile_add_self_flat[tensordict-eager]	0.4203ms	0.2740ms	3.6492 KOps/s	3.6039 KOps/s	$\color{#35bf28}+1.26\%$
test_compile_add_self_flat[tensordict-compile]	0.4694ms	0.3481ms	2.8724 KOps/s	2.8771 KOps/s	$\color{#d91a1a}-0.16\%$
test_compile_add_self_flat[tensorclass-eager]	0.2306ms	76.0074μs	13.1566 KOps/s	13.0137 KOps/s	$\color{#35bf28}+1.10\%$
test_compile_add_self_flat[tensorclass-compile]	0.2878ms	0.1385ms	7.2203 KOps/s	6.8739 KOps/s	$\textbf{\color{#35bf28}+5.04\%}$
test_compile_add_self_flat[pytree-eager]	0.7200ms	0.5741ms	1.7418 KOps/s	1.6861 KOps/s	$\color{#35bf28}+3.30\%$
test_compile_add_self_flat[pytree-compile]	0.4519ms	0.3432ms	2.9140 KOps/s	2.9016 KOps/s	$\color{#35bf28}+0.43\%$
test_compile_copy_flat[tensordict-compile]	0.1613ms	20.3449μs	49.1523 KOps/s	48.7364 KOps/s	$\color{#35bf28}+0.85\%$
test_compile_copy_flat[tensordict-eager]	64.4810μs	32.5812μs	30.6925 KOps/s	30.4628 KOps/s	$\color{#35bf28}+0.75\%$
test_compile_copy_flat[pytree-compile]	0.1842ms	77.0855μs	12.9726 KOps/s	12.8907 KOps/s	$\color{#35bf28}+0.64\%$
test_compile_copy_flat[pytree-eager]	93.1320μs	60.0547μs	16.6515 KOps/s	16.4629 KOps/s	$\color{#35bf28}+1.15\%$
test_compile_assign_and_add[tensordict-compile]	2.5072ms	0.8737ms	1.1446 KOps/s	1.0632 KOps/s	$\textbf{\color{#35bf28}+7.66\%}$
test_compile_assign_and_add[tensordict-eager]	3.7284ms	3.5239ms	283.7751 Ops/s	281.1193 Ops/s	$\color{#35bf28}+0.94\%$
test_compile_assign_and_add[pytree-compile]	2.5108ms	0.8612ms	1.1612 KOps/s	1.0682 KOps/s	$\textbf{\color{#35bf28}+8.71\%}$
test_compile_assign_and_add[pytree-eager]	3.9175ms	3.6066ms	277.2728 Ops/s	276.1444 Ops/s	$\color{#35bf28}+0.41\%$
test_compile_indexing[tensor-tensordict-compile]	0.2632ms	0.1195ms	8.3662 KOps/s	8.0101 KOps/s	$\color{#35bf28}+4.45\%$
test_compile_indexing[tensor-tensordict-eager]	0.3107ms	66.8198μs	14.9656 KOps/s	14.3023 KOps/s	$\color{#35bf28}+4.64\%$
test_compile_indexing[tensor-tensorclass-compile]	0.2723ms	0.1159ms	8.6295 KOps/s	8.6640 KOps/s	$\color{#d91a1a}-0.40\%$
test_compile_indexing[tensor-tensorclass-eager]	0.2352ms	51.4709μs	19.4285 KOps/s	20.5001 KOps/s	$\textbf{\color{#d91a1a}-5.23\%}$
test_compile_indexing[tensor-pytree-compile]	0.2801ms	0.1155ms	8.6605 KOps/s	8.8064 KOps/s	$\color{#d91a1a}-1.66\%$
test_compile_indexing[tensor-pytree-eager]	0.2409ms	51.2303μs	19.5197 KOps/s	20.2853 KOps/s	$\color{#d91a1a}-3.77\%$
test_compile_indexing[slice-tensordict-compile]	0.3695ms	0.1524ms	6.5599 KOps/s	6.5622 KOps/s	$\color{#d91a1a}-0.03\%$
test_compile_indexing[slice-tensordict-eager]	0.2396ms	29.3606μs	34.0592 KOps/s	35.0555 KOps/s	$\color{#d91a1a}-2.84\%$
test_compile_indexing[slice-tensorclass-compile]	0.2891ms	0.1407ms	7.1095 KOps/s	6.8271 KOps/s	$\color{#35bf28}+4.14\%$
test_compile_indexing[slice-tensorclass-eager]	0.1586ms	23.6932μs	42.2062 KOps/s	40.7193 KOps/s	$\color{#35bf28}+3.65\%$
test_compile_indexing[slice-pytree-compile]	0.3015ms	0.1405ms	7.1167 KOps/s	6.8680 KOps/s	$\color{#35bf28}+3.62\%$
test_compile_indexing[slice-pytree-eager]	0.2254ms	23.8495μs	41.9296 KOps/s	41.5179 KOps/s	$\color{#35bf28}+0.99\%$
test_compile_indexing[int-tensordict-compile]	0.4086ms	0.1514ms	6.6032 KOps/s	6.4556 KOps/s	$\color{#35bf28}+2.29\%$
test_compile_indexing[int-tensordict-eager]	0.5075ms	27.5010μs	36.3623 KOps/s	35.4150 KOps/s	$\color{#35bf28}+2.67\%$
test_compile_indexing[int-tensorclass-compile]	0.2957ms	0.1407ms	7.1069 KOps/s	7.0473 KOps/s	$\color{#35bf28}+0.85\%$
test_compile_indexing[int-tensorclass-eager]	0.1350ms	25.1764μs	39.7197 KOps/s	40.8442 KOps/s	$\color{#d91a1a}-2.75\%$
test_compile_indexing[int-pytree-compile]	0.3200ms	0.1465ms	6.8245 KOps/s	7.1011 KOps/s	$\color{#d91a1a}-3.89\%$
test_compile_indexing[int-pytree-eager]	53.8500μs	25.3930μs	39.3809 KOps/s	41.6349 KOps/s	$\textbf{\color{#d91a1a}-5.41\%}$
test_mod_add[eager]	0.2093ms	35.4128μs	28.2384 KOps/s	27.8625 KOps/s	$\color{#35bf28}+1.35\%$
test_mod_add[compile]	0.2072ms	74.9678μs	13.3391 KOps/s	13.1028 KOps/s	$\color{#35bf28}+1.80\%$
test_mod_add[compile-overhead]	0.2707ms	0.1425ms	7.0183 KOps/s	6.4285 KOps/s	$\textbf{\color{#35bf28}+9.17\%}$
test_mod_wrap[eager]	0.4012ms	0.2489ms	4.0180 KOps/s	3.8881 KOps/s	$\color{#35bf28}+3.34\%$
test_mod_wrap[compile]	1.2326ms	0.3070ms	3.2578 KOps/s	3.1664 KOps/s	$\color{#35bf28}+2.89\%$
test_mod_wrap[compile-overhead]	8.4132ms	4.3819ms	228.2117 Ops/s	231.6199 Ops/s	$\color{#d91a1a}-1.47\%$
test_mod_wrap_and_backward[eager]	1.5631ms	1.3902ms	719.3127 Ops/s	704.6603 Ops/s	$\color{#35bf28}+2.08\%$
test_mod_wrap_and_backward[compile]	1.5439ms	1.3892ms	719.8564 Ops/s	662.5210 Ops/s	$\textbf{\color{#35bf28}+8.65\%}$
test_mod_wrap_and_backward[compile-overhead]	1.3375ms	0.9270ms	1.0787 KOps/s	966.9465 Ops/s	$\textbf{\color{#35bf28}+11.56\%}$
test_seq_add[eager]	0.2529ms	0.1043ms	9.5914 KOps/s	9.2046 KOps/s	$\color{#35bf28}+4.20\%$
test_seq_add[compile]	0.2347ms	87.3096μs	11.4535 KOps/s	11.0183 KOps/s	$\color{#35bf28}+3.95\%$
test_seq_add[compile-overhead]	0.2659ms	0.1246ms	8.0261 KOps/s	7.9537 KOps/s	$\color{#35bf28}+0.91\%$
test_seq_wrap[eager]	0.5615ms	0.4048ms	2.4705 KOps/s	2.4129 KOps/s	$\color{#35bf28}+2.39\%$
test_seq_wrap[compile]	0.5138ms	0.3349ms	2.9860 KOps/s	2.9863 KOps/s	$-0.01\%$
test_seq_wrap[compile-overhead]	0.3822ms	0.2422ms	4.1280 KOps/s	4.1108 KOps/s	$\color{#35bf28}+0.42\%$
test_func_call_runtime[False-eager]	0.9049ms	0.7605ms	1.3149 KOps/s	1.2982 KOps/s	$\color{#35bf28}+1.29\%$
test_func_call_runtime[False-compile]	1.0351ms	0.8683ms	1.1516 KOps/s	1.1880 KOps/s	$\color{#d91a1a}-3.06\%$
test_func_call_runtime[False-compile-overhead]	0.4722ms	0.3850ms	2.5972 KOps/s	2.5764 KOps/s	$\color{#35bf28}+0.81\%$
test_func_call_runtime[True-eager]	1.1765ms	0.9584ms	1.0434 KOps/s	1.0407 KOps/s	$\color{#35bf28}+0.25\%$
test_func_call_runtime[True-compile]	1.0188ms	0.8782ms	1.1387 KOps/s	1.1280 KOps/s	$\color{#35bf28}+0.95\%$
test_func_call_runtime[True-compile-overhead]	0.6080ms	0.4315ms	2.3176 KOps/s	2.3132 KOps/s	$\color{#35bf28}+0.19\%$
test_func_call_cm_runtime[False-eager]	0.9634ms	0.7907ms	1.2647 KOps/s	1.3024 KOps/s	$\color{#d91a1a}-2.89\%$
test_func_call_cm_runtime[False-compile]	1.0574ms	0.8470ms	1.1806 KOps/s	1.1804 KOps/s	$\color{#35bf28}+0.01\%$
test_func_call_cm_runtime[False-compile-overhead]	0.5384ms	0.3879ms	2.5779 KOps/s	2.5459 KOps/s	$\color{#35bf28}+1.26\%$
test_func_call_cm_runtime[True-eager]	1.2330ms	1.0636ms	940.2349 Ops/s	927.3585 Ops/s	$\color{#35bf28}+1.39\%$
test_func_call_cm_runtime[True-compile]	1.1852ms	1.0387ms	962.7831 Ops/s	947.7602 Ops/s	$\color{#35bf28}+1.59\%$
test_func_call_cm_runtime[True-compile-overhead]	1.1877ms	1.0315ms	969.4998 Ops/s	946.4511 Ops/s	$\color{#35bf28}+2.44\%$
test_distributed	1.2916ms	71.7629μs	13.9348 KOps/s	14.6363 KOps/s	$\color{#d91a1a}-4.79\%$
test_tdmodule	69.8810μs	16.0064μs	62.4751 KOps/s	55.8470 KOps/s	$\textbf{\color{#35bf28}+11.87\%}$
test_tdmodule_dispatch	53.6600μs	31.8687μs	31.3788 KOps/s	27.9016 KOps/s	$\textbf{\color{#35bf28}+12.46\%}$
test_tdseq	33.3310μs	16.5918μs	60.2708 KOps/s	54.0815 KOps/s	$\textbf{\color{#35bf28}+11.44\%}$
test_tdseq_dispatch	50.7410μs	33.1169μs	30.1961 KOps/s	26.5467 KOps/s	$\textbf{\color{#35bf28}+13.75\%}$
test_instantiation_functorch	2.2553ms	2.0908ms	478.2768 Ops/s	478.6613 Ops/s	$\color{#d91a1a}-0.08\%$
test_instantiation_td	2.0537ms	1.3550ms	737.9967 Ops/s	737.9661 Ops/s	$+0.00\%$
test_exec_functorch	0.3749ms	0.2355ms	4.2455 KOps/s	4.2760 KOps/s	$\color{#d91a1a}-0.71\%$
test_exec_functional_call	0.4148ms	0.2393ms	4.1784 KOps/s	4.4427 KOps/s	$\textbf{\color{#d91a1a}-5.95\%}$
test_exec_td	0.4373ms	0.2437ms	4.1032 KOps/s	4.2906 KOps/s	$\color{#d91a1a}-4.37\%$
test_exec_td_decorator	0.6028ms	0.2874ms	3.4790 KOps/s	3.4770 KOps/s	$\color{#35bf28}+0.06\%$
test_vmap_mlp_speed[True-True]	0.8229ms	0.6610ms	1.5129 KOps/s	1.4897 KOps/s	$\color{#35bf28}+1.55\%$
test_vmap_mlp_speed[True-False]	0.8804ms	0.6877ms	1.4542 KOps/s	1.5008 KOps/s	$\color{#d91a1a}-3.10\%$
test_vmap_mlp_speed[False-True]	0.7819ms	0.6091ms	1.6417 KOps/s	1.7231 KOps/s	$\color{#d91a1a}-4.72\%$
test_vmap_mlp_speed[False-False]	0.7695ms	0.6105ms	1.6379 KOps/s	1.7231 KOps/s	$\color{#d91a1a}-4.95\%$
test_vmap_mlp_speed_decorator[True-True]	1.3559ms	0.7125ms	1.4034 KOps/s	1.3838 KOps/s	$\color{#35bf28}+1.41\%$
test_vmap_mlp_speed_decorator[True-False]	0.9632ms	0.7391ms	1.3530 KOps/s	1.3835 KOps/s	$\color{#d91a1a}-2.20\%$
test_vmap_mlp_speed_decorator[False-True]	0.8326ms	0.6491ms	1.5407 KOps/s	1.6051 KOps/s	$\color{#d91a1a}-4.01\%$
test_vmap_mlp_speed_decorator[False-False]	0.7998ms	0.6310ms	1.5847 KOps/s	1.5997 KOps/s	$\color{#d91a1a}-0.94\%$
test_vmap_transformer_speed[True-True]	9.3381ms	8.8883ms	112.5076 Ops/s	111.7936 Ops/s	$\color{#35bf28}+0.64\%$
test_vmap_transformer_speed[True-False]	9.2134ms	8.8588ms	112.8820 Ops/s	112.5895 Ops/s	$\color{#35bf28}+0.26\%$
test_vmap_transformer_speed[False-True]	9.1413ms	8.7763ms	113.9428 Ops/s	113.3041 Ops/s	$\color{#35bf28}+0.56\%$
test_vmap_transformer_speed[False-False]	9.2636ms	8.7850ms	113.8299 Ops/s	113.7433 Ops/s	$\color{#35bf28}+0.08\%$
test_vmap_transformer_speed_decorator[True-True]	22.0769ms	21.0437ms	47.5202 Ops/s	47.8268 Ops/s	$\color{#d91a1a}-0.64\%$
test_vmap_transformer_speed_decorator[True-False]	0.1984s	24.7288ms	40.4387 Ops/s	47.6432 Ops/s	$\textbf{\color{#d91a1a}-15.12\%}$
test_vmap_transformer_speed_decorator[False-True]	21.8742ms	20.8787ms	47.8958 Ops/s	48.0046 Ops/s	$\color{#d91a1a}-0.23\%$
test_vmap_transformer_speed_decorator[False-False]	21.7712ms	20.8365ms	47.9927 Ops/s	47.9871 Ops/s	$\color{#35bf28}+0.01\%$
test_to_module_speed[True]	1.6815ms	1.1513ms	868.5457 Ops/s	866.5182 Ops/s	$\color{#35bf28}+0.23\%$
test_to_module_speed[False]	1.6016ms	1.1216ms	891.5949 Ops/s	889.7193 Ops/s	$\color{#35bf28}+0.21\%$
test_tc_init	0.1351ms	37.7610μs	26.4823 KOps/s	23.7695 KOps/s	$\textbf{\color{#35bf28}+11.41\%}$
test_tc_init_nested	0.2044ms	78.6685μs	12.7116 KOps/s	11.8059 KOps/s	$\textbf{\color{#35bf28}+7.67\%}$
test_tc_first_layer_tensor	3.5502μs	0.7794μs	1.2831 MOps/s	1.2756 MOps/s	$\color{#35bf28}+0.58\%$
test_tc_first_layer_nontensor	16.8100μs	2.5558μs	391.2723 KOps/s	397.0708 KOps/s	$\color{#d91a1a}-1.46\%$
test_tc_second_layer_tensor	25.9900μs	1.7052μs	586.4248 KOps/s	618.7240 KOps/s	$\textbf{\color{#d91a1a}-5.22\%}$
test_tc_second_layer_nontensor	27.6210μs	3.4540μs	289.5202 KOps/s	296.8910 KOps/s	$\color{#d91a1a}-2.48\%$
test_unbind	0.1819s	12.9144ms	77.4328 Ops/s	84.2086 Ops/s	$\textbf{\color{#d91a1a}-8.05\%}$
test_full_like	0.7558ms	0.5782ms	1.7296 KOps/s	1.7346 KOps/s	$\color{#d91a1a}-0.29\%$
test_zeros_like	0.3406ms	0.1979ms	5.0525 KOps/s	5.0591 KOps/s	$\color{#d91a1a}-0.13\%$
test_ones_like	0.3506ms	0.1979ms	5.0524 KOps/s	5.0608 KOps/s	$\color{#d91a1a}-0.17\%$
test_clone	0.5628ms	0.4153ms	2.4080 KOps/s	2.4138 KOps/s	$\color{#d91a1a}-0.24\%$
test_squeeze	0.1457ms	11.2988μs	88.5047 KOps/s	89.2232 KOps/s	$\color{#d91a1a}-0.81\%$
test_unsqueeze	0.2814ms	82.9407μs	12.0568 KOps/s	12.0587 KOps/s	$\color{#d91a1a}-0.02\%$
test_split	0.2948ms	0.1805ms	5.5404 KOps/s	5.5363 KOps/s	$\color{#35bf28}+0.07\%$
test_permute	0.3108ms	0.1972ms	5.0704 KOps/s	5.1269 KOps/s	$\color{#d91a1a}-1.10\%$
test_stack	1.3791ms	0.9141ms	1.0940 KOps/s	1.0905 KOps/s	$\color{#35bf28}+0.32\%$
test_cat	1.2743ms	1.2314ms	812.0684 Ops/s	811.6571 Ops/s	$\color{#35bf28}+0.05\%$