[BugFix] Better handling of struct arrays

[vmoens]

Stack from ghstack (oldest at bottom):

• -> [BugFix] Better handling of struct arrays #1193

[github-actions]

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

Total Benchmarks: 229. Improved: $\large\color{#35bf28}43$. Worsened: $\large\color{#d91a1a}8$.

Expand to view detailed results

Name	Max	Mean	Ops	Ops on Repo HEAD	Change
test_plain_set_nested	31.5800μs	11.2227μs	89.1051 KOps/s	76.0996 KOps/s	$\textbf{\color{#35bf28}+17.09\%}$
test_plain_set_stack_nested	38.7510μs	11.3568μs	88.0534 KOps/s	75.6218 KOps/s	$\textbf{\color{#35bf28}+16.44\%}$
test_plain_set_nested_inplace	39.2010μs	12.3004μs	81.2979 KOps/s	69.8058 KOps/s	$\textbf{\color{#35bf28}+16.46\%}$
test_plain_set_stack_nested_inplace	37.3110μs	12.4365μs	80.4086 KOps/s	70.3083 KOps/s	$\textbf{\color{#35bf28}+14.37\%}$
test_items	22.8100μs	2.8826μs	346.9034 KOps/s	338.1740 KOps/s	$\color{#35bf28}+2.58\%$
test_items_nested	0.4107ms	0.3625ms	2.7585 KOps/s	2.7210 KOps/s	$\color{#35bf28}+1.38\%$
test_items_nested_locked	0.4911ms	0.3660ms	2.7323 KOps/s	2.7247 KOps/s	$\color{#35bf28}+0.28\%$
test_items_nested_leaf	88.4200μs	58.9903μs	16.9519 KOps/s	16.8323 KOps/s	$\color{#35bf28}+0.71\%$
test_items_stack_nested	0.4076ms	0.3664ms	2.7292 KOps/s	2.7259 KOps/s	$\color{#35bf28}+0.12\%$
test_items_stack_nested_leaf	0.1093ms	60.0373μs	16.6563 KOps/s	16.6976 KOps/s	$\color{#d91a1a}-0.25\%$
test_items_stack_nested_locked	0.5371ms	0.3641ms	2.7466 KOps/s	2.7284 KOps/s	$\color{#35bf28}+0.66\%$
test_keys	26.3000μs	3.5071μs	285.1350 KOps/s	285.0917 KOps/s	$\color{#35bf28}+0.02\%$
test_keys_nested	0.1297ms	89.4943μs	11.1739 KOps/s	11.0868 KOps/s	$\color{#35bf28}+0.79\%$
test_keys_nested_locked	0.6351ms	94.7836μs	10.5503 KOps/s	10.4885 KOps/s	$\color{#35bf28}+0.59\%$
test_keys_nested_leaf	0.1408ms	80.4822μs	12.4251 KOps/s	12.3376 KOps/s	$\color{#35bf28}+0.71\%$
test_keys_stack_nested	0.1274ms	90.6699μs	11.0290 KOps/s	10.9788 KOps/s	$\color{#35bf28}+0.46\%$
test_keys_stack_nested_leaf	0.1176ms	81.6282μs	12.2507 KOps/s	12.2828 KOps/s	$\color{#d91a1a}-0.26\%$
test_keys_stack_nested_locked	0.1364ms	96.4540μs	10.3676 KOps/s	10.3949 KOps/s	$\color{#d91a1a}-0.26\%$
test_values	6.8135μs	0.8723μs	1.1464 MOps/s	1.1487 MOps/s	$\color{#d91a1a}-0.20\%$
test_values_nested	67.8400μs	37.3031μs	26.8074 KOps/s	26.5036 KOps/s	$\color{#35bf28}+1.15\%$
test_values_nested_locked	65.5910μs	38.9987μs	25.6419 KOps/s	25.6454 KOps/s	$\color{#d91a1a}-0.01\%$
test_values_nested_leaf	69.3600μs	41.8278μs	23.9075 KOps/s	23.8193 KOps/s	$\color{#35bf28}+0.37\%$
test_values_stack_nested	68.3710μs	37.8887μs	26.3931 KOps/s	26.1537 KOps/s	$\color{#35bf28}+0.92\%$
test_values_stack_nested_leaf	66.9010μs	42.3182μs	23.6305 KOps/s	23.3683 KOps/s	$\color{#35bf28}+1.12\%$
test_values_stack_nested_locked	69.0200μs	39.7557μs	25.1536 KOps/s	25.2644 KOps/s	$\color{#d91a1a}-0.44\%$
test_membership	1.7135μs	0.5061μs	1.9759 MOps/s	1.9551 MOps/s	$\color{#35bf28}+1.07\%$
test_membership_nested	19.4100μs	2.0266μs	493.4333 KOps/s	486.5749 KOps/s	$\color{#35bf28}+1.41\%$
test_membership_nested_leaf	19.0650μs	2.0379μs	490.7122 KOps/s	488.9835 KOps/s	$\color{#35bf28}+0.35\%$
test_membership_stacked_nested	28.5100μs	2.1089μs	474.1831 KOps/s	480.1358 KOps/s	$\color{#d91a1a}-1.24\%$
test_membership_stacked_nested_leaf	28.5500μs	2.0902μs	478.4222 KOps/s	479.4081 KOps/s	$\color{#d91a1a}-0.21\%$
test_membership_nested_last	35.6600μs	3.0922μs	323.3991 KOps/s	320.6956 KOps/s	$\color{#35bf28}+0.84\%$
test_membership_nested_leaf_last	39.5200μs	3.1284μs	319.6477 KOps/s	322.2644 KOps/s	$\color{#d91a1a}-0.81\%$
test_membership_stacked_nested_last	0.9717ms	3.0502μs	327.8474 KOps/s	322.0950 KOps/s	$\color{#35bf28}+1.79\%$
test_membership_stacked_nested_leaf_last	37.2600μs	3.0883μs	323.8066 KOps/s	323.2093 KOps/s	$\color{#35bf28}+0.18\%$
test_nested_getleaf	28.5500μs	6.1499μs	162.6039 KOps/s	161.4464 KOps/s	$\color{#35bf28}+0.72\%$
test_nested_get	26.5410μs	5.8423μs	171.1649 KOps/s	170.9464 KOps/s	$\color{#35bf28}+0.13\%$
test_stacked_getleaf	31.3300μs	6.2023μs	161.2303 KOps/s	162.2969 KOps/s	$\color{#d91a1a}-0.66\%$
test_stacked_get	33.4910μs	5.8628μs	170.5677 KOps/s	170.4133 KOps/s	$\color{#35bf28}+0.09\%$
test_nested_getitemleaf	31.2410μs	6.4300μs	155.5207 KOps/s	154.5669 KOps/s	$\color{#35bf28}+0.62\%$
test_nested_getitem	24.9410μs	6.1659μs	162.1827 KOps/s	161.0554 KOps/s	$\color{#35bf28}+0.70\%$
test_stacked_getitemleaf	32.8310μs	6.4505μs	155.0263 KOps/s	154.0386 KOps/s	$\color{#35bf28}+0.64\%$
test_stacked_getitem	32.3300μs	6.1438μs	162.7669 KOps/s	163.5470 KOps/s	$\color{#d91a1a}-0.48\%$
test_lock_nested	0.4293ms	0.3367ms	2.9702 KOps/s	2.8541 KOps/s	$\color{#35bf28}+4.07\%$
test_lock_stack_nested	0.5045ms	0.3437ms	2.9094 KOps/s	2.9069 KOps/s	$\color{#35bf28}+0.09\%$
test_unlock_nested	0.3514ms	0.2795ms	3.5776 KOps/s	3.5622 KOps/s	$\color{#35bf28}+0.43\%$
test_unlock_stack_nested	0.3161ms	0.2784ms	3.5922 KOps/s	3.5436 KOps/s	$\color{#35bf28}+1.37\%$
test_flatten_speed	0.1245ms	76.2664μs	13.1119 KOps/s	12.8706 KOps/s	$\color{#35bf28}+1.88\%$
test_unflatten_speed	0.3754ms	0.3237ms	3.0894 KOps/s	3.0648 KOps/s	$\color{#35bf28}+0.80\%$
test_common_ops	0.7337ms	0.5669ms	1.7640 KOps/s	1.5509 KOps/s	$\textbf{\color{#35bf28}+13.74\%}$
test_creation	0.1244ms	1.7773μs	562.6574 KOps/s	563.4787 KOps/s	$\color{#d91a1a}-0.15\%$
test_creation_empty	29.4600μs	6.5890μs	151.7673 KOps/s	98.7118 KOps/s	$\textbf{\color{#35bf28}+53.75\%}$
test_creation_nested_1	35.3610μs	8.1897μs	122.1039 KOps/s	84.5267 KOps/s	$\textbf{\color{#35bf28}+44.46\%}$
test_creation_nested_2	42.2800μs	10.9560μs	91.2739 KOps/s	68.3253 KOps/s	$\textbf{\color{#35bf28}+33.59\%}$
test_clone	57.4400μs	9.9403μs	100.6004 KOps/s	98.8259 KOps/s	$\color{#35bf28}+1.80\%$
test_getitem[int]	1.3087ms	10.7694μs	92.8559 KOps/s	92.7630 KOps/s	$\color{#35bf28}+0.10\%$
test_getitem[slice_int]	0.1110ms	21.0142μs	47.5870 KOps/s	47.2514 KOps/s	$\color{#35bf28}+0.71\%$
test_getitem[range]	0.1270ms	37.7573μs	26.4849 KOps/s	26.9210 KOps/s	$\color{#d91a1a}-1.62\%$
test_getitem[tuple]	0.1045ms	18.0994μs	55.2503 KOps/s	53.3228 KOps/s	$\color{#35bf28}+3.61\%$
test_getitem[list]	0.1729ms	32.2538μs	31.0041 KOps/s	30.5433 KOps/s	$\color{#35bf28}+1.51\%$
test_setitem_dim[int]	44.9010μs	18.3053μs	54.6289 KOps/s	52.8129 KOps/s	$\color{#35bf28}+3.44\%$
test_setitem_dim[slice_int]	62.2810μs	38.3449μs	26.0791 KOps/s	26.6257 KOps/s	$\color{#d91a1a}-2.05\%$
test_setitem_dim[range]	74.2010μs	52.5509μs	19.0292 KOps/s	18.8310 KOps/s	$\color{#35bf28}+1.05\%$
test_setitem_dim[tuple]	52.5010μs	31.0871μs	32.1677 KOps/s	31.4688 KOps/s	$\color{#35bf28}+2.22\%$
test_setitem	71.8600μs	13.2651μs	75.3859 KOps/s	64.1868 KOps/s	$\textbf{\color{#35bf28}+17.45\%}$
test_set	62.3410μs	13.1477μs	76.0589 KOps/s	65.3940 KOps/s	$\textbf{\color{#35bf28}+16.31\%}$
test_set_shared	0.5078ms	0.1593ms	6.2777 KOps/s	6.2352 KOps/s	$\color{#35bf28}+0.68\%$
test_update	0.3976ms	14.8236μs	67.4600 KOps/s	52.8273 KOps/s	$\textbf{\color{#35bf28}+27.70\%}$
test_update_nested	87.4410μs	20.1101μs	49.7264 KOps/s	40.6527 KOps/s	$\textbf{\color{#35bf28}+22.32\%}$
test_update__nested	0.5331ms	24.7064μs	40.4754 KOps/s	40.0788 KOps/s	$\color{#35bf28}+0.99\%$
test_set_nested	75.7400μs	14.2084μs	70.3810 KOps/s	60.6238 KOps/s	$\textbf{\color{#35bf28}+16.09\%}$
test_set_nested_new	0.1114ms	16.3515μs	61.1565 KOps/s	53.2189 KOps/s	$\textbf{\color{#35bf28}+14.92\%}$
test_select	93.4410μs	27.3962μs	36.5015 KOps/s	32.2179 KOps/s	$\textbf{\color{#35bf28}+13.30\%}$
test_select_nested	76.5700μs	44.0685μs	22.6919 KOps/s	22.5566 KOps/s	$\color{#35bf28}+0.60\%$
test_exclude_nested	97.7410μs	62.9466μs	15.8865 KOps/s	15.6145 KOps/s	$\color{#35bf28}+1.74\%$
test_empty[True]	0.3461ms	0.2956ms	3.3825 KOps/s	3.3394 KOps/s	$\color{#35bf28}+1.29\%$
test_empty[False]	3.5850μs	0.8296μs	1.2053 MOps/s	1.1948 MOps/s	$\color{#35bf28}+0.89\%$
test_to	87.8710μs	56.7949μs	17.6072 KOps/s	17.3287 KOps/s	$\color{#35bf28}+1.61\%$
test_to_nonblocking	88.8000μs	49.0856μs	20.3726 KOps/s	21.3146 KOps/s	$\color{#d91a1a}-4.42\%$
test_unbind_speed	0.2959ms	0.2429ms	4.1175 KOps/s	4.1542 KOps/s	$\color{#d91a1a}-0.88\%$
test_unbind_speed_stack0	0.2957ms	0.2398ms	4.1696 KOps/s	4.1333 KOps/s	$\color{#35bf28}+0.88\%$
test_unbind_speed_stack1	93.5189ms	0.7354ms	1.3598 KOps/s	1.2282 KOps/s	$\textbf{\color{#35bf28}+10.71\%}$
test_split	95.4502ms	1.5993ms	625.2800 Ops/s	627.6837 Ops/s	$\color{#d91a1a}-0.38\%$
test_chunk	95.2710ms	1.5969ms	626.2009 Ops/s	626.6563 Ops/s	$\color{#d91a1a}-0.07\%$
test_consolidate[False-None]	3.5966ms	2.7151ms	368.3078 Ops/s	364.5578 Ops/s	$\color{#35bf28}+1.03\%$
test_consolidate[default-None]	2.1643ms	1.7163ms	582.6445 Ops/s	585.4959 Ops/s	$\color{#d91a1a}-0.49\%$
test_consolidate[reduce-overhead-None]	1.8352ms	1.7360ms	576.0458 Ops/s	577.7619 Ops/s	$\color{#d91a1a}-0.30\%$
test_consolidate_njt[False-None]	6.7483ms	6.4509ms	155.0173 Ops/s	155.5221 Ops/s	$\color{#d91a1a}-0.32\%$
test_to[False-False-None]	1.7968ms	1.6892ms	591.9905 Ops/s	606.3551 Ops/s	$\color{#d91a1a}-2.37\%$
test_to[True-False-None]	1.5396ms	1.3508ms	740.2953 Ops/s	762.9142 Ops/s	$\color{#d91a1a}-2.96\%$
test_to[within-False-None]	4.4454ms	4.1174ms	242.8702 Ops/s	239.7858 Ops/s	$\color{#35bf28}+1.29\%$
test_to[True-default-None]	5.4427ms	5.1987ms	192.3543 Ops/s	193.2269 Ops/s	$\color{#d91a1a}-0.45\%$
test_to_njt[False-False-None]	7.0005ms	6.8783ms	145.3855 Ops/s	146.1084 Ops/s	$\color{#d91a1a}-0.49\%$
test_to_njt[True-False-None]	5.6766ms	5.4454ms	183.6425 Ops/s	184.2486 Ops/s	$\color{#d91a1a}-0.33\%$
test_to_njt[within-False-None]	12.4822ms	12.0902ms	82.7118 Ops/s	83.5055 Ops/s	$\color{#d91a1a}-0.95\%$
test_creation[device0]	0.3607ms	81.9136μs	12.2080 KOps/s	12.3579 KOps/s	$\color{#d91a1a}-1.21\%$
test_creation_from_tensor	0.6071ms	84.9544μs	11.7710 KOps/s	11.8649 KOps/s	$\color{#d91a1a}-0.79\%$
test_add_one[memmap_tensor0]	0.4147ms	6.2198μs	160.7764 KOps/s	161.4665 KOps/s	$\color{#d91a1a}-0.43\%$
test_contiguous[memmap_tensor0]	5.2901μs	0.4205μs	2.3781 MOps/s	2.3310 MOps/s	$\color{#35bf28}+2.02\%$
test_stack[memmap_tensor0]	98.7810μs	4.4609μs	224.1683 KOps/s	233.0354 KOps/s	$\color{#d91a1a}-3.81\%$
test_memmaptd_index	0.4494ms	0.2372ms	4.2162 KOps/s	4.1958 KOps/s	$\color{#35bf28}+0.49\%$
test_memmaptd_index_astensor	0.4304ms	0.2978ms	3.3582 KOps/s	3.3206 KOps/s	$\color{#35bf28}+1.13\%$
test_memmaptd_index_op	0.7793ms	0.5292ms	1.8895 KOps/s	1.7217 KOps/s	$\textbf{\color{#35bf28}+9.75\%}$
test_serialize_model	0.1322s	0.1303s	7.6734 Ops/s	7.6257 Ops/s	$\color{#35bf28}+0.63\%$
test_serialize_model_pickle	1.3515s	1.2136s	0.8240 Ops/s	0.8232 Ops/s	$\color{#35bf28}+0.10\%$
test_serialize_weights	0.4069s	0.1695s	5.8998 Ops/s	5.5325 Ops/s	$\textbf{\color{#35bf28}+6.64\%}$
test_serialize_weights_returnearly	0.3319s	54.7898ms	18.2516 Ops/s	23.1837 Ops/s	$\textbf{\color{#d91a1a}-21.27\%}$
test_serialize_weights_pickle	1.3774s	1.2184s	0.8208 Ops/s	0.8191 Ops/s	$\color{#35bf28}+0.20\%$
test_reshape_pytree	0.1041ms	21.8203μs	45.8290 KOps/s	44.7269 KOps/s	$\color{#35bf28}+2.46\%$
test_reshape_td	0.4240ms	26.3344μs	37.9731 KOps/s	36.6777 KOps/s	$\color{#35bf28}+3.53\%$
test_view_pytree	58.9200μs	21.7455μs	45.9866 KOps/s	45.4532 KOps/s	$\color{#35bf28}+1.17\%$
test_view_td	0.1133ms	29.1749μs	34.2760 KOps/s	30.3140 KOps/s	$\textbf{\color{#35bf28}+13.07\%}$
test_unbind_pytree	65.7410μs	27.4083μs	36.4854 KOps/s	36.3689 KOps/s	$\color{#35bf28}+0.32\%$
test_unbind_td	0.7493ms	35.8784μs	27.8720 KOps/s	27.6181 KOps/s	$\color{#35bf28}+0.92\%$
test_split_pytree	90.7810μs	29.9023μs	33.4422 KOps/s	32.9532 KOps/s	$\color{#35bf28}+1.48\%$
test_split_td	0.9218ms	38.3120μs	26.1015 KOps/s	25.3836 KOps/s	$\color{#35bf28}+2.83\%$
test_add_pytree	78.3610μs	32.2374μs	31.0198 KOps/s	30.9932 KOps/s	$\color{#35bf28}+0.09\%$
test_add_td	0.1354ms	43.2456μs	23.1237 KOps/s	20.5867 KOps/s	$\textbf{\color{#35bf28}+12.32\%}$
test_compile_add_one_nested[tensordict-compile]	0.1798ms	0.1253ms	7.9820 KOps/s	7.8026 KOps/s	$\color{#35bf28}+2.30\%$
test_compile_add_one_nested[tensordict-eager]	0.2317ms	0.1336ms	7.4857 KOps/s	7.5877 KOps/s	$\color{#d91a1a}-1.34\%$
test_compile_add_one_nested[pytree-compile]	0.2098ms	96.8371μs	10.3266 KOps/s	10.3178 KOps/s	$\color{#35bf28}+0.09\%$
test_compile_add_one_nested[pytree-eager]	0.2519ms	0.1469ms	6.8089 KOps/s	6.9024 KOps/s	$\color{#d91a1a}-1.35\%$
test_compile_copy_nested[tensordict-compile]	0.1266ms	22.6140μs	44.2203 KOps/s	44.4719 KOps/s	$\color{#d91a1a}-0.57\%$
test_compile_copy_nested[tensordict-eager]	0.1224ms	29.7789μs	33.5808 KOps/s	33.1889 KOps/s	$\color{#35bf28}+1.18\%$
test_compile_copy_nested[pytree-compile]	0.3256ms	63.4717μs	15.7550 KOps/s	15.3592 KOps/s	$\color{#35bf28}+2.58\%$
test_compile_copy_nested[pytree-eager]	0.1268ms	48.8193μs	20.4837 KOps/s	19.9684 KOps/s	$\color{#35bf28}+2.58\%$
test_compile_add_one_flat[tensordict-compile]	0.1836ms	0.1433ms	6.9764 KOps/s	7.0085 KOps/s	$\color{#d91a1a}-0.46\%$
test_compile_add_one_flat[tensordict-eager]	0.3243ms	0.2184ms	4.5779 KOps/s	4.6800 KOps/s	$\color{#d91a1a}-2.18\%$
test_compile_add_one_flat[tensorclass-compile]	0.2138ms	99.0959μs	10.0912 KOps/s	10.1351 KOps/s	$\color{#d91a1a}-0.43\%$
test_compile_add_one_flat[tensorclass-eager]	0.1530ms	56.1303μs	17.8157 KOps/s	18.3797 KOps/s	$\color{#d91a1a}-3.07\%$
test_compile_add_one_flat[pytree-compile]	0.2458ms	0.1360ms	7.3528 KOps/s	7.4282 KOps/s	$\color{#d91a1a}-1.01\%$
test_compile_add_one_flat[pytree-eager]	0.6212ms	0.4661ms	2.1455 KOps/s	2.1413 KOps/s	$\color{#35bf28}+0.20\%$
test_compile_add_self_flat[tensordict-eager]	0.6643ms	0.2617ms	3.8211 KOps/s	3.8680 KOps/s	$\color{#d91a1a}-1.21\%$
test_compile_add_self_flat[tensordict-compile]	0.5739ms	0.1458ms	6.8568 KOps/s	6.9974 KOps/s	$\color{#d91a1a}-2.01\%$
test_compile_add_self_flat[tensorclass-eager]	0.4877ms	69.7741μs	14.3320 KOps/s	14.8473 KOps/s	$\color{#d91a1a}-3.47\%$
test_compile_add_self_flat[tensorclass-compile]	0.1512ms	0.1012ms	9.8825 KOps/s	10.1156 KOps/s	$\color{#d91a1a}-2.30\%$
test_compile_add_self_flat[pytree-eager]	0.5357ms	0.4038ms	2.4762 KOps/s	2.4920 KOps/s	$\color{#d91a1a}-0.63\%$
test_compile_add_self_flat[pytree-compile]	0.1999ms	0.1363ms	7.3375 KOps/s	7.5064 KOps/s	$\color{#d91a1a}-2.25\%$
test_compile_copy_flat[tensordict-compile]	92.2410μs	17.4488μs	57.3107 KOps/s	57.8757 KOps/s	$\color{#d91a1a}-0.98\%$
test_compile_copy_flat[tensordict-eager]	0.1301ms	33.0320μs	30.2737 KOps/s	30.4628 KOps/s	$\color{#d91a1a}-0.62\%$
test_compile_copy_flat[pytree-compile]	0.2044ms	70.4585μs	14.1928 KOps/s	14.2220 KOps/s	$\color{#d91a1a}-0.21\%$
test_compile_copy_flat[pytree-eager]	0.1336ms	52.4463μs	19.0671 KOps/s	19.4143 KOps/s	$\color{#d91a1a}-1.79\%$
test_compile_assign_and_add[tensordict-compile]	1.6376ms	0.3934ms	2.5417 KOps/s	2.1398 KOps/s	$\textbf{\color{#35bf28}+18.78\%}$
test_compile_assign_and_add[tensordict-eager]	2.9236ms	2.6497ms	377.3974 Ops/s	388.0128 Ops/s	$\color{#d91a1a}-2.74\%$
test_compile_assign_and_add[pytree-compile]	1.6026ms	0.4342ms	2.3030 KOps/s	2.1970 KOps/s	$\color{#35bf28}+4.83\%$
test_compile_assign_and_add[pytree-eager]	2.9309ms	2.8062ms	356.3568 Ops/s	357.6045 Ops/s	$\color{#d91a1a}-0.35\%$
test_compile_indexing[tensor-tensordict-compile]	0.7532ms	0.1173ms	8.5276 KOps/s	8.3660 KOps/s	$\color{#35bf28}+1.93\%$
test_compile_indexing[tensor-tensordict-eager]	0.5982ms	80.2971μs	12.4537 KOps/s	11.4366 KOps/s	$\textbf{\color{#35bf28}+8.89\%}$
test_compile_indexing[tensor-tensorclass-compile]	0.7154ms	0.1092ms	9.1585 KOps/s	9.0614 KOps/s	$\color{#35bf28}+1.07\%$
test_compile_indexing[tensor-tensorclass-eager]	0.1134ms	67.3689μs	14.8436 KOps/s	14.3588 KOps/s	$\color{#35bf28}+3.38\%$
test_compile_indexing[tensor-pytree-compile]	0.1763ms	0.1126ms	8.8772 KOps/s	9.4793 KOps/s	$\textbf{\color{#d91a1a}-6.35\%}$
test_compile_indexing[tensor-pytree-eager]	0.2109ms	71.7854μs	13.9304 KOps/s	13.5326 KOps/s	$\color{#35bf28}+2.94\%$
test_compile_indexing[slice-tensordict-compile]	0.1488ms	0.1019ms	9.8089 KOps/s	9.1398 KOps/s	$\textbf{\color{#35bf28}+7.32\%}$
test_compile_indexing[slice-tensordict-eager]	0.1439ms	17.5984μs	56.8235 KOps/s	49.8601 KOps/s	$\textbf{\color{#35bf28}+13.97\%}$
test_compile_indexing[slice-tensorclass-compile]	0.1732ms	97.4555μs	10.2611 KOps/s	9.5728 KOps/s	$\textbf{\color{#35bf28}+7.19\%}$
test_compile_indexing[slice-tensorclass-eager]	47.8900μs	16.0068μs	62.4736 KOps/s	63.3868 KOps/s	$\color{#d91a1a}-1.44\%$
test_compile_indexing[slice-pytree-compile]	0.1518ms	0.1006ms	9.9433 KOps/s	10.1310 KOps/s	$\color{#d91a1a}-1.85\%$
test_compile_indexing[slice-pytree-eager]	51.5910μs	15.5241μs	64.4161 KOps/s	62.5365 KOps/s	$\color{#35bf28}+3.01\%$
test_compile_indexing[int-tensordict-compile]	0.2103ms	0.1072ms	9.3244 KOps/s	8.9962 KOps/s	$\color{#35bf28}+3.65\%$
test_compile_indexing[int-tensordict-eager]	0.5718ms	17.0666μs	58.5939 KOps/s	49.2407 KOps/s	$\textbf{\color{#35bf28}+18.99\%}$
test_compile_indexing[int-tensorclass-compile]	0.1622ms	0.1014ms	9.8583 KOps/s	9.3432 KOps/s	$\textbf{\color{#35bf28}+5.51\%}$
test_compile_indexing[int-tensorclass-eager]	54.8900μs	15.5868μs	64.1569 KOps/s	57.9995 KOps/s	$\textbf{\color{#35bf28}+10.62\%}$
test_compile_indexing[int-pytree-compile]	0.1510ms	97.6531μs	10.2403 KOps/s	9.3809 KOps/s	$\textbf{\color{#35bf28}+9.16\%}$
test_compile_indexing[int-pytree-eager]	66.6010μs	20.7239μs	48.2534 KOps/s	54.1669 KOps/s	$\textbf{\color{#d91a1a}-10.92\%}$
test_mod_add[eager]	80.2610μs	36.4818μs	27.4110 KOps/s	24.3688 KOps/s	$\textbf{\color{#35bf28}+12.48\%}$
test_mod_add[compile]	0.5366ms	81.0890μs	12.3321 KOps/s	12.2260 KOps/s	$\color{#35bf28}+0.87\%$
test_mod_add[compile-overhead]	0.3347ms	0.1733ms	5.7706 KOps/s	5.5342 KOps/s	$\color{#35bf28}+4.27\%$
test_mod_wrap[eager]	0.3268ms	0.2435ms	4.1069 KOps/s	3.9021 KOps/s	$\textbf{\color{#35bf28}+5.25\%}$
test_mod_wrap[compile]	0.3437ms	0.2857ms	3.5002 KOps/s	3.4788 KOps/s	$\color{#35bf28}+0.61\%$
test_mod_wrap[compile-overhead]	7.1900ms	3.7257ms	268.4083 Ops/s	274.3644 Ops/s	$\color{#d91a1a}-2.17\%$
test_mod_wrap_and_backward[eager]	1.5758ms	1.4302ms	699.1787 Ops/s	692.8986 Ops/s	$\color{#35bf28}+0.91\%$
test_mod_wrap_and_backward[compile]	1.4411ms	1.3364ms	748.2597 Ops/s	730.7943 Ops/s	$\color{#35bf28}+2.39\%$
test_mod_wrap_and_backward[compile-overhead]	1.4185ms	0.9329ms	1.0720 KOps/s	968.5536 Ops/s	$\textbf{\color{#35bf28}+10.68\%}$
test_seq_add[eager]	0.6074ms	0.1187ms	8.4226 KOps/s	8.2317 KOps/s	$\color{#35bf28}+2.32\%$
test_seq_add[compile]	0.1941ms	92.2211μs	10.8435 KOps/s	10.6258 KOps/s	$\color{#35bf28}+2.05\%$
test_seq_add[compile-overhead]	0.1826ms	0.1384ms	7.2240 KOps/s	7.6031 KOps/s	$\color{#d91a1a}-4.99\%$
test_seq_wrap[eager]	0.4954ms	0.4243ms	2.3566 KOps/s	2.2030 KOps/s	$\textbf{\color{#35bf28}+6.97\%}$
test_seq_wrap[compile]	0.3693ms	0.3009ms	3.3231 KOps/s	3.3013 KOps/s	$\color{#35bf28}+0.66\%$
test_seq_wrap[compile-overhead]	0.2899ms	0.2277ms	4.3918 KOps/s	4.3570 KOps/s	$\color{#35bf28}+0.80\%$
test_func_call_runtime[False-eager]	0.7732ms	0.7059ms	1.4166 KOps/s	1.2612 KOps/s	$\textbf{\color{#35bf28}+12.32\%}$
test_func_call_runtime[False-compile]	0.8040ms	0.7313ms	1.3675 KOps/s	1.3186 KOps/s	$\color{#35bf28}+3.71\%$
test_func_call_runtime[False-compile-overhead]	0.4187ms	0.3686ms	2.7128 KOps/s	2.6888 KOps/s	$\color{#35bf28}+0.89\%$
test_func_call_runtime[True-eager]	0.9776ms	0.8788ms	1.1379 KOps/s	1.1149 KOps/s	$\color{#35bf28}+2.06\%$
test_func_call_runtime[True-compile]	0.8148ms	0.7550ms	1.3245 KOps/s	1.3034 KOps/s	$\color{#35bf28}+1.61\%$
test_func_call_runtime[True-compile-overhead]	0.4562ms	0.3907ms	2.5594 KOps/s	2.5576 KOps/s	$\color{#35bf28}+0.07\%$
test_func_call_cm_runtime[False-eager]	0.7970ms	0.7016ms	1.4253 KOps/s	1.2571 KOps/s	$\textbf{\color{#35bf28}+13.38\%}$
test_func_call_cm_runtime[False-compile]	0.8477ms	0.7336ms	1.3631 KOps/s	1.3246 KOps/s	$\color{#35bf28}+2.91\%$
test_func_call_cm_runtime[False-compile-overhead]	0.4224ms	0.3703ms	2.7006 KOps/s	2.6809 KOps/s	$\color{#35bf28}+0.74\%$
test_func_call_cm_runtime[True-eager]	1.0741ms	0.9739ms	1.0268 KOps/s	936.3896 Ops/s	$\textbf{\color{#35bf28}+9.66\%}$
test_func_call_cm_runtime[True-compile]	1.0817ms	0.9617ms	1.0398 KOps/s	1.0211 KOps/s	$\color{#35bf28}+1.83\%$
test_func_call_cm_runtime[True-compile-overhead]	1.0411ms	0.9621ms	1.0394 KOps/s	1.0161 KOps/s	$\color{#35bf28}+2.29\%$
test_vmap_func_call_cm_runtime[eager]	2.4322ms	2.0226ms	494.4221 Ops/s	484.9704 Ops/s	$\color{#35bf28}+1.95\%$
test_vmap_func_call_cm_runtime[compile]	0.8591ms	0.7976ms	1.2537 KOps/s	1.2078 KOps/s	$\color{#35bf28}+3.80\%$
test_vmap_func_call_cm_runtime[compile-overhead]	0.4706ms	0.4230ms	2.3639 KOps/s	2.3456 KOps/s	$\color{#35bf28}+0.78\%$
test_distributed	3.9814ms	0.1940ms	5.1545 KOps/s	8.5213 KOps/s	$\textbf{\color{#d91a1a}-39.51\%}$
test_tdmodule	0.4462ms	19.3383μs	51.7107 KOps/s	49.0122 KOps/s	$\textbf{\color{#35bf28}+5.51\%}$
test_tdmodule_dispatch	51.8400μs	32.5937μs	30.6808 KOps/s	24.6353 KOps/s	$\textbf{\color{#35bf28}+24.54\%}$
test_tdseq	37.9900μs	19.6552μs	50.8772 KOps/s	42.5406 KOps/s	$\textbf{\color{#35bf28}+19.60\%}$
test_tdseq_dispatch	70.5000μs	36.6718μs	27.2689 KOps/s	22.7332 KOps/s	$\textbf{\color{#35bf28}+19.95\%}$
test_instantiation_functorch	1.7651ms	1.5805ms	632.7133 Ops/s	635.6256 Ops/s	$\color{#d91a1a}-0.46\%$
test_exec_functorch	0.1933ms	0.1435ms	6.9679 KOps/s	6.6986 KOps/s	$\color{#35bf28}+4.02\%$
test_exec_functional_call	0.1992ms	0.1362ms	7.3448 KOps/s	7.0965 KOps/s	$\color{#35bf28}+3.50\%$
test_exec_td_decorator	0.3689ms	0.1836ms	5.4467 KOps/s	5.1444 KOps/s	$\textbf{\color{#35bf28}+5.88\%}$
test_vmap_mlp_speed_decorator[True-True]	0.8284ms	0.6815ms	1.4673 KOps/s	1.4577 KOps/s	$\color{#35bf28}+0.66\%$
test_vmap_mlp_speed_decorator[True-False]	0.7956ms	0.6662ms	1.5009 KOps/s	1.4817 KOps/s	$\color{#35bf28}+1.30\%$
test_vmap_mlp_speed_decorator[False-True]	0.7246ms	0.6092ms	1.6415 KOps/s	1.7316 KOps/s	$\textbf{\color{#d91a1a}-5.21\%}$
test_vmap_mlp_speed_decorator[False-False]	0.7184ms	0.6088ms	1.6426 KOps/s	1.7303 KOps/s	$\textbf{\color{#d91a1a}-5.07\%}$
test_vmap_transformer_speed_decorator[True-True]	19.7670ms	18.7875ms	53.2269 Ops/s	53.9603 Ops/s	$\color{#d91a1a}-1.36\%$
test_vmap_transformer_speed_decorator[True-False]	19.3546ms	18.7027ms	53.4683 Ops/s	54.0645 Ops/s	$\color{#d91a1a}-1.10\%$
test_vmap_transformer_speed_decorator[False-True]	18.9089ms	18.5708ms	53.8480 Ops/s	54.5691 Ops/s	$\color{#d91a1a}-1.32\%$
test_vmap_transformer_speed_decorator[False-False]	18.7033ms	18.5763ms	53.8320 Ops/s	54.5187 Ops/s	$\color{#d91a1a}-1.26\%$
test_to_module_speed[True]	1.0435ms	0.9712ms	1.0296 KOps/s	1.0276 KOps/s	$\color{#35bf28}+0.20\%$
test_to_module_speed[False]	1.0377ms	0.9617ms	1.0398 KOps/s	1.0397 KOps/s	$+0.01\%$
test_tc_init	76.1800μs	34.4252μs	29.0485 KOps/s	26.4293 KOps/s	$\textbf{\color{#35bf28}+9.91\%}$
test_tc_init_nested	0.1122ms	70.4584μs	14.1928 KOps/s	13.3059 KOps/s	$\textbf{\color{#35bf28}+6.67\%}$
test_tc_first_layer_tensor	5.3971μs	0.6971μs	1.4345 MOps/s	1.4110 MOps/s	$\color{#35bf28}+1.67\%$
test_tc_first_layer_nontensor	27.7800μs	2.2738μs	439.7862 KOps/s	442.0247 KOps/s	$\color{#d91a1a}-0.51\%$
test_tc_second_layer_tensor	9.0400μs	1.4155μs	706.4657 KOps/s	705.7274 KOps/s	$\color{#35bf28}+0.10\%$
test_tc_second_layer_nontensor	33.0500μs	3.0033μs	332.9698 KOps/s	331.4314 KOps/s	$\color{#35bf28}+0.46\%$
test_unbind	7.3057ms	7.0006ms	142.8441 Ops/s	142.5800 Ops/s	$\color{#35bf28}+0.19\%$
test_full_like	10.7435ms	9.2219ms	108.4376 Ops/s	106.6513 Ops/s	$\color{#35bf28}+1.67\%$
test_zeros_like	4.9619ms	4.3343ms	230.7191 Ops/s	230.7546 Ops/s	$\color{#d91a1a}-0.02\%$
test_ones_like	4.4559ms	4.3390ms	230.4679 Ops/s	230.5699 Ops/s	$\color{#d91a1a}-0.04\%$
test_clone	11.9361ms	9.2295ms	108.3482 Ops/s	156.6431 Ops/s	$\textbf{\color{#d91a1a}-30.83\%}$
test_squeeze	85.1810μs	10.0462μs	99.5399 KOps/s	104.0000 KOps/s	$\color{#d91a1a}-4.29\%$
test_unsqueeze	0.1273ms	75.7221μs	13.2062 KOps/s	13.9445 KOps/s	$\textbf{\color{#d91a1a}-5.29\%}$
test_split	0.2962ms	0.1602ms	6.2424 KOps/s	6.3516 KOps/s	$\color{#d91a1a}-1.72\%$
test_permute	0.7284ms	0.1812ms	5.5199 KOps/s	5.7894 KOps/s	$\color{#d91a1a}-4.65\%$
test_stack	50.9250ms	50.6804ms	19.7315 Ops/s	19.8739 Ops/s	$\color{#d91a1a}-0.72\%$
test_cat	50.9058ms	50.6363ms	19.7487 Ops/s	19.8963 Ops/s	$\color{#d91a1a}-0.74\%$