[Feature] Add Redis-backed TensorDict (RedisTensorDict)

[vmoens]

Stack from ghstack (oldest at bottom):

• [Performance] Use Lua scripts for scattered index operations #1571
• [Benchmark] Add Redis benchmarks, optimize reads with covering-range strategy #1570
• [Performance] Optimize RedisTensorDict indexed ops with GETRANGE/SETRANGE #1569
• [Feature] Add from_tensordict and from_redis classmethods to RedisTensorDict #1568
• -> [Feature] Add Redis-backed TensorDict (RedisTensorDict) #1567

Adds RedisTensorDict, a TensorDictBase subclass that stores tensors in a
Redis instance for out-of-core storage. This enables datasets that exceed
local RAM to be accessed with the familiar TensorDict interface.

MVP features:

• Async redis.asyncio client on a background event loop with sync wrappers
• Zero-copy serialization via torch.frombuffer / untyped_storage bytes
• Metadata (shape, dtype) stored in Redis Hashes for fast introspection
• Redis Cluster-compatible key schema using hash tags
• Pipelined batch I/O for all get/set operations
• Nested TensorDict support via lightweight prefix-based views
• Pickle support for multi-process usage
• from_dict() factory, to_local() / to_tensordict() materialization
• CI: install redis package and start redis-server before tests

[github-actions]

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

Total Benchmarks: 243. Improved: $\large\color{#35bf28}12$. Worsened: $\large\color{#d91a1a}7$.

Expand to view detailed results

Name	Max	Mean	Ops	Ops on Repo HEAD	Change
test_plain_set_nested	38.8100μs	14.8761μs	67.2220 KOps/s	66.7814 KOps/s	$\color{#35bf28}+0.66\%$
test_plain_set_stack_nested	52.4900μs	15.1528μs	65.9944 KOps/s	66.0276 KOps/s	$\color{#d91a1a}-0.05\%$
test_plain_set_nested_inplace	45.5210μs	16.5238μs	60.5186 KOps/s	59.7986 KOps/s	$\color{#35bf28}+1.20\%$
test_plain_set_stack_nested_inplace	48.6410μs	16.4485μs	60.7960 KOps/s	60.0789 KOps/s	$\color{#35bf28}+1.19\%$
test_items	37.3200μs	5.7434μs	174.1140 KOps/s	173.1432 KOps/s	$\color{#35bf28}+0.56\%$
test_items_nested	0.5996ms	0.5333ms	1.8752 KOps/s	1.8502 KOps/s	$\color{#35bf28}+1.35\%$
test_items_nested_locked	0.5971ms	0.5373ms	1.8613 KOps/s	1.8400 KOps/s	$\color{#35bf28}+1.16\%$
test_items_nested_leaf	0.1378ms	96.6860μs	10.3428 KOps/s	10.5075 KOps/s	$\color{#d91a1a}-1.57\%$
test_items_stack_nested	0.5766ms	0.5304ms	1.8853 KOps/s	1.8528 KOps/s	$\color{#35bf28}+1.76\%$
test_items_stack_nested_leaf	0.1372ms	96.8933μs	10.3206 KOps/s	10.4732 KOps/s	$\color{#d91a1a}-1.46\%$
test_items_stack_nested_locked	0.5981ms	0.5399ms	1.8521 KOps/s	1.8368 KOps/s	$\color{#35bf28}+0.83\%$
test_keys	41.7300μs	4.3934μs	227.6158 KOps/s	235.9342 KOps/s	$\color{#d91a1a}-3.53\%$
test_keys_nested	0.1513ms	0.1204ms	8.3029 KOps/s	8.2911 KOps/s	$\color{#35bf28}+0.14\%$
test_keys_nested_locked	88.5357ms	0.1412ms	7.0831 KOps/s	7.7570 KOps/s	$\textbf{\color{#d91a1a}-8.69\%}$
test_keys_nested_leaf	0.1950ms	0.1113ms	8.9826 KOps/s	9.0781 KOps/s	$\color{#d91a1a}-1.05\%$
test_keys_stack_nested	0.1621ms	0.1210ms	8.2613 KOps/s	8.3147 KOps/s	$\color{#d91a1a}-0.64\%$
test_keys_stack_nested_leaf	0.1493ms	0.1113ms	8.9837 KOps/s	9.0314 KOps/s	$\color{#d91a1a}-0.53\%$
test_keys_stack_nested_locked	0.1724ms	0.1295ms	7.7197 KOps/s	7.7985 KOps/s	$\color{#d91a1a}-1.01\%$
test_values	6.5562μs	1.0195μs	980.8678 KOps/s	972.5758 KOps/s	$\color{#35bf28}+0.85\%$
test_values_nested	76.1810μs	47.8318μs	20.9066 KOps/s	20.9505 KOps/s	$\color{#d91a1a}-0.21\%$
test_values_nested_locked	81.0710μs	50.6063μs	19.7604 KOps/s	19.5009 KOps/s	$\color{#35bf28}+1.33\%$
test_values_nested_leaf	81.6810μs	54.2028μs	18.4492 KOps/s	18.4555 KOps/s	$\color{#d91a1a}-0.03\%$
test_values_stack_nested	94.2510μs	47.5883μs	21.0136 KOps/s	20.9630 KOps/s	$\color{#35bf28}+0.24\%$
test_values_stack_nested_leaf	93.2620μs	54.3416μs	18.4021 KOps/s	18.4109 KOps/s	$\color{#d91a1a}-0.05\%$
test_values_stack_nested_locked	0.1325ms	50.1099μs	19.9561 KOps/s	19.7002 KOps/s	$\color{#35bf28}+1.30\%$
test_membership	11.8168μs	0.8633μs	1.1583 MOps/s	1.1848 MOps/s	$\color{#d91a1a}-2.23\%$
test_membership_nested	31.8200μs	3.1531μs	317.1526 KOps/s	314.6553 KOps/s	$\color{#35bf28}+0.79\%$
test_membership_nested_leaf	26.3600μs	3.1701μs	315.4464 KOps/s	314.9292 KOps/s	$\color{#35bf28}+0.16\%$
test_membership_stacked_nested	41.7610μs	3.1961μs	312.8817 KOps/s	311.8411 KOps/s	$\color{#35bf28}+0.33\%$
test_membership_stacked_nested_leaf	32.4410μs	3.2024μs	312.2706 KOps/s	310.9717 KOps/s	$\color{#35bf28}+0.42\%$
test_membership_nested_last	33.5700μs	4.6278μs	216.0867 KOps/s	212.2262 KOps/s	$\color{#35bf28}+1.82\%$
test_membership_nested_leaf_last	33.5000μs	4.6339μs	215.8022 KOps/s	212.8174 KOps/s	$\color{#35bf28}+1.40\%$
test_membership_stacked_nested_last	46.8510μs	4.6284μs	216.0559 KOps/s	213.0321 KOps/s	$\color{#35bf28}+1.42\%$
test_membership_stacked_nested_leaf_last	42.1610μs	4.6303μs	215.9678 KOps/s	210.7766 KOps/s	$\color{#35bf28}+2.46\%$
test_nested_getleaf	52.2410μs	21.7358μs	46.0071 KOps/s	45.7895 KOps/s	$\color{#35bf28}+0.48\%$
test_nested_get	48.1810μs	20.1031μs	49.7436 KOps/s	48.6377 KOps/s	$\color{#35bf28}+2.27\%$
test_stacked_getleaf	51.3310μs	21.5575μs	46.3875 KOps/s	46.1158 KOps/s	$\color{#35bf28}+0.59\%$
test_stacked_get	49.3510μs	20.6307μs	48.4715 KOps/s	47.4049 KOps/s	$\color{#35bf28}+2.25\%$
test_nested_getitemleaf	89.7820μs	21.5987μs	46.2990 KOps/s	44.4674 KOps/s	$\color{#35bf28}+4.12\%$
test_nested_getitem	60.7710μs	20.9220μs	47.7966 KOps/s	47.3372 KOps/s	$\color{#35bf28}+0.97\%$
test_stacked_getitemleaf	54.2010μs	21.7717μs	45.9311 KOps/s	44.8127 KOps/s	$\color{#35bf28}+2.50\%$
test_stacked_getitem	61.5410μs	20.8582μs	47.9428 KOps/s	47.0166 KOps/s	$\color{#35bf28}+1.97\%$
test_lock_nested	7.7418ms	0.4781ms	2.0917 KOps/s	2.0997 KOps/s	$\color{#d91a1a}-0.38\%$
test_lock_stack_nested	0.5782ms	0.4777ms	2.0933 KOps/s	2.0591 KOps/s	$\color{#35bf28}+1.66\%$
test_unlock_nested	0.4861ms	0.3790ms	2.6386 KOps/s	2.6068 KOps/s	$\color{#35bf28}+1.22\%$
test_unlock_stack_nested	0.4923ms	0.3824ms	2.6149 KOps/s	2.5587 KOps/s	$\color{#35bf28}+2.20\%$
test_flatten_speed	0.1987ms	0.1221ms	8.1887 KOps/s	8.0733 KOps/s	$\color{#35bf28}+1.43\%$
test_unflatten_speed	0.6312ms	0.5927ms	1.6873 KOps/s	1.6475 KOps/s	$\color{#35bf28}+2.41\%$
test_common_ops	0.8029ms	0.6782ms	1.4745 KOps/s	1.4520 KOps/s	$\color{#35bf28}+1.55\%$
test_creation	0.2025ms	2.9452μs	339.5368 KOps/s	343.7897 KOps/s	$\color{#d91a1a}-1.24\%$
test_creation_empty	35.8410μs	6.1466μs	162.6908 KOps/s	162.0687 KOps/s	$\color{#35bf28}+0.38\%$
test_creation_nested_1	35.7810μs	10.7669μs	92.8776 KOps/s	91.6629 KOps/s	$\color{#35bf28}+1.33\%$
test_creation_nested_2	73.5910μs	11.8233μs	84.5789 KOps/s	84.3577 KOps/s	$\color{#35bf28}+0.26\%$
test_creation_many_keys[10]	45.6410μs	18.2740μs	54.7227 KOps/s	54.7909 KOps/s	$\color{#d91a1a}-0.12\%$
test_creation_many_keys[50]	0.1187ms	77.9019μs	12.8367 KOps/s	12.6842 KOps/s	$\color{#35bf28}+1.20\%$
test_creation_many_keys[100]	0.1822ms	0.1522ms	6.5683 KOps/s	6.4817 KOps/s	$\color{#35bf28}+1.34\%$
test_creation_nested_many_keys[10]	75.4020μs	39.2902μs	25.4517 KOps/s	25.4173 KOps/s	$\color{#35bf28}+0.14\%$
test_creation_nested_many_keys[50]	0.1964ms	0.1594ms	6.2747 KOps/s	6.2383 KOps/s	$\color{#35bf28}+0.58\%$
test_clone	44.3600μs	13.3148μs	75.1043 KOps/s	72.3906 KOps/s	$\color{#35bf28}+3.75\%$
test_getitem[int]	1.6267ms	14.5622μs	68.6709 KOps/s	59.9894 KOps/s	$\textbf{\color{#35bf28}+14.47\%}$
test_getitem[slice_int]	0.1414ms	25.4012μs	39.3682 KOps/s	39.0950 KOps/s	$\color{#35bf28}+0.70\%$
test_getitem[range]	0.1800ms	61.6472μs	16.2213 KOps/s	15.9601 KOps/s	$\color{#35bf28}+1.64\%$
test_getitem[tuple]	0.1428ms	24.7496μs	40.4047 KOps/s	40.6766 KOps/s	$\color{#d91a1a}-0.67\%$
test_getitem[list]	0.1921ms	57.0904μs	17.5161 KOps/s	17.3126 KOps/s	$\color{#35bf28}+1.18\%$
test_setitem_dim[int]	48.3600μs	26.2116μs	38.1511 KOps/s	37.4252 KOps/s	$\color{#35bf28}+1.94\%$
test_setitem_dim[slice_int]	67.2510μs	44.8000μs	22.3214 KOps/s	22.0219 KOps/s	$\color{#35bf28}+1.36\%$
test_setitem_dim[range]	0.1325ms	93.9189μs	10.6475 KOps/s	10.6252 KOps/s	$\color{#35bf28}+0.21\%$
test_setitem_dim[tuple]	62.4610μs	41.7655μs	23.9432 KOps/s	23.5620 KOps/s	$\color{#35bf28}+1.62\%$
test_setitem	56.0110μs	18.1592μs	55.0685 KOps/s	53.3905 KOps/s	$\color{#35bf28}+3.14\%$
test_set	61.7810μs	17.0090μs	58.7925 KOps/s	55.8208 KOps/s	$\textbf{\color{#35bf28}+5.32\%}$
test_set_shared	0.4988ms	0.2076ms	4.8162 KOps/s	4.8221 KOps/s	$\color{#d91a1a}-0.12\%$
test_update	0.3734ms	22.5558μs	44.3345 KOps/s	43.8610 KOps/s	$\color{#35bf28}+1.08\%$
test_update_nested	75.0410μs	34.7768μs	28.7548 KOps/s	28.0326 KOps/s	$\color{#35bf28}+2.58\%$
test_update__nested	0.4623ms	34.8749μs	28.6739 KOps/s	27.5375 KOps/s	$\color{#35bf28}+4.13\%$
test_set_nested	53.2310μs	19.1006μs	52.3543 KOps/s	50.5732 KOps/s	$\color{#35bf28}+3.52\%$
test_set_nested_new	62.7710μs	24.3212μs	41.1164 KOps/s	39.6876 KOps/s	$\color{#35bf28}+3.60\%$
test_select	77.6410μs	42.7179μs	23.4094 KOps/s	23.1380 KOps/s	$\color{#35bf28}+1.17\%$
test_select_nested	0.1239ms	75.5640μs	13.2338 KOps/s	13.2843 KOps/s	$\color{#d91a1a}-0.38\%$
test_exclude_nested	0.1416ms	97.5130μs	10.2550 KOps/s	10.1010 KOps/s	$\color{#35bf28}+1.53\%$
test_empty[True]	0.5086ms	0.4375ms	2.2855 KOps/s	2.2607 KOps/s	$\color{#35bf28}+1.10\%$
test_empty[False]	7.1900μs	1.3258μs	754.2791 KOps/s	757.3441 KOps/s	$\color{#d91a1a}-0.40\%$
test_to	0.1017ms	71.1624μs	14.0524 KOps/s	13.1335 KOps/s	$\textbf{\color{#35bf28}+7.00\%}$
test_to_nonblocking	0.1148ms	64.4632μs	15.5127 KOps/s	14.9826 KOps/s	$\color{#35bf28}+3.54\%$
test_unbind_speed	0.3563ms	0.3280ms	3.0490 KOps/s	3.0276 KOps/s	$\color{#35bf28}+0.71\%$
test_unbind_speed_stack0	0.3985ms	0.3241ms	3.0855 KOps/s	3.0524 KOps/s	$\color{#35bf28}+1.08\%$
test_unbind_speed_stack1	0.1033s	0.9158ms	1.0920 KOps/s	1.1797 KOps/s	$\textbf{\color{#d91a1a}-7.44\%}$
test_split	1.2225ms	1.1389ms	878.0329 Ops/s	883.4957 Ops/s	$\color{#d91a1a}-0.62\%$
test_chunk	0.1030s	1.2127ms	824.5811 Ops/s	912.9976 Ops/s	$\textbf{\color{#d91a1a}-9.68\%}$
test_to_cpu_blocking	28.4204ms	28.3334ms	35.2940 Ops/s	35.4372 Ops/s	$\color{#d91a1a}-0.40\%$
test_to_cpu_global_sync	11.3438ms	11.2063ms	89.2354 Ops/s	87.3708 Ops/s	$\color{#35bf28}+2.13\%$
test_to_cpu_event_sync	12.4547ms	12.2027ms	81.9490 Ops/s	80.6291 Ops/s	$\color{#35bf28}+1.64\%$
test_to_cpu_default	12.4839ms	12.2623ms	81.5509 Ops/s	80.5745 Ops/s	$\color{#35bf28}+1.21\%$
test_consolidate[False-None]	4.2464ms	4.0808ms	245.0526 Ops/s	243.3805 Ops/s	$\color{#35bf28}+0.69\%$
test_consolidate[default-None]	2.1288ms	2.0192ms	495.2481 Ops/s	482.2397 Ops/s	$\color{#35bf28}+2.70\%$
test_consolidate[reduce-overhead-None]	2.0573ms	1.9494ms	512.9901 Ops/s	504.3587 Ops/s	$\color{#35bf28}+1.71\%$
test_consolidate_njt[False-None]	9.0339ms	8.6180ms	116.0359 Ops/s	117.4618 Ops/s	$\color{#d91a1a}-1.21\%$
test_to[False-False-None]	2.2485ms	2.0750ms	481.9345 Ops/s	472.0930 Ops/s	$\color{#35bf28}+2.08\%$
test_to[True-False-None]	2.1706ms	1.9171ms	521.6186 Ops/s	522.3086 Ops/s	$\color{#d91a1a}-0.13\%$
test_to[within-False-None]	6.2734ms	6.0884ms	164.2465 Ops/s	164.1972 Ops/s	$\color{#35bf28}+0.03\%$
test_to[True-default-None]	8.0055ms	7.7475ms	129.0734 Ops/s	131.3863 Ops/s	$\color{#d91a1a}-1.76\%$
test_to_njt[False-False-None]	8.9005ms	8.5668ms	116.7292 Ops/s	115.1994 Ops/s	$\color{#35bf28}+1.33\%$
test_to_njt[True-False-None]	7.1006ms	6.9267ms	144.3690 Ops/s	142.9813 Ops/s	$\color{#35bf28}+0.97\%$
test_to_njt[within-False-None]	16.1117ms	15.4796ms	64.6012 Ops/s	63.4711 Ops/s	$\color{#35bf28}+1.78\%$
test_creation[device0]	0.3869ms	0.1169ms	8.5513 KOps/s	8.5303 KOps/s	$\color{#35bf28}+0.25\%$
test_creation_from_tensor	0.4203ms	0.1175ms	8.5142 KOps/s	8.6616 KOps/s	$\color{#d91a1a}-1.70\%$
test_add_one[memmap_tensor0]	0.2097ms	6.5079μs	153.6596 KOps/s	145.8026 KOps/s	$\textbf{\color{#35bf28}+5.39\%}$
test_contiguous[memmap_tensor0]	26.2800μs	0.6924μs	1.4443 MOps/s	2.1773 MOps/s	$\textbf{\color{#d91a1a}-33.66\%}$
test_stack[memmap_tensor0]	35.1900μs	4.6728μs	214.0033 KOps/s	214.3699 KOps/s	$\color{#d91a1a}-0.17\%$
test_memmaptd_index	1.0587ms	0.2623ms	3.8120 KOps/s	3.8442 KOps/s	$\color{#d91a1a}-0.84\%$
test_memmaptd_index_astensor	0.5166ms	0.3584ms	2.7904 KOps/s	2.8338 KOps/s	$\color{#d91a1a}-1.53\%$
test_memmaptd_index_op	0.7479ms	0.5996ms	1.6677 KOps/s	1.6390 KOps/s	$\color{#35bf28}+1.75\%$
test_serialize_model	0.1403s	0.1366s	7.3221 Ops/s	7.2873 Ops/s	$\color{#35bf28}+0.48\%$
test_serialize_model_pickle	1.3501s	1.2151s	0.8230 Ops/s	0.8357 Ops/s	$\color{#d91a1a}-1.53\%$
test_serialize_weights	0.1375s	0.1350s	7.4064 Ops/s	7.3401 Ops/s	$\color{#35bf28}+0.90\%$
test_serialize_weights_returnearly	0.3801s	94.2204ms	10.6134 Ops/s	14.2458 Ops/s	$\textbf{\color{#d91a1a}-25.50\%}$
test_serialize_weights_pickle	1.3635s	1.2126s	0.8247 Ops/s	0.8232 Ops/s	$\color{#35bf28}+0.18\%$
test_reshape_pytree	0.2312ms	33.4882μs	29.8613 KOps/s	30.0470 KOps/s	$\color{#d91a1a}-0.62\%$
test_reshape_td	78.4520μs	44.2527μs	22.5975 KOps/s	22.4181 KOps/s	$\color{#35bf28}+0.80\%$
test_view_pytree	0.2210ms	32.7771μs	30.5091 KOps/s	30.1352 KOps/s	$\color{#35bf28}+1.24\%$
test_view_td	88.0710μs	52.1178μs	19.1873 KOps/s	18.8571 KOps/s	$\color{#35bf28}+1.75\%$
test_unbind_pytree	0.2327ms	37.0252μs	27.0086 KOps/s	26.9573 KOps/s	$\color{#35bf28}+0.19\%$
test_unbind_td	0.1373ms	48.1297μs	20.7772 KOps/s	20.1772 KOps/s	$\color{#35bf28}+2.97\%$
test_split_pytree	0.2539ms	43.4082μs	23.0371 KOps/s	23.2688 KOps/s	$\color{#d91a1a}-1.00\%$
test_split_td	0.1637ms	65.5897μs	15.2463 KOps/s	15.2849 KOps/s	$\color{#d91a1a}-0.25\%$
test_add_pytree	0.1983ms	42.2851μs	23.6490 KOps/s	23.1334 KOps/s	$\color{#35bf28}+2.23\%$
test_add_td	0.1594ms	52.9230μs	18.8954 KOps/s	18.4168 KOps/s	$\color{#35bf28}+2.60\%$
test_compile_add_one_nested[tensordict-compile]	0.1901ms	0.1404ms	7.1200 KOps/s	6.6027 KOps/s	$\textbf{\color{#35bf28}+7.84\%}$
test_compile_add_one_nested[tensordict-eager]	0.2735ms	0.1946ms	5.1380 KOps/s	5.1179 KOps/s	$\color{#35bf28}+0.39\%$
test_compile_add_one_nested[pytree-compile]	0.1597ms	0.1094ms	9.1393 KOps/s	9.0529 KOps/s	$\color{#35bf28}+0.95\%$
test_compile_add_one_nested[pytree-eager]	0.4386ms	0.1852ms	5.3988 KOps/s	5.4644 KOps/s	$\color{#d91a1a}-1.20\%$
test_compile_copy_nested[tensordict-compile]	69.4820μs	31.4807μs	31.7655 KOps/s	27.3936 KOps/s	$\textbf{\color{#35bf28}+15.96\%}$
test_compile_copy_nested[tensordict-eager]	98.6820μs	52.1287μs	19.1833 KOps/s	19.5848 KOps/s	$\color{#d91a1a}-2.05\%$
test_compile_copy_nested[pytree-compile]	51.8800μs	9.9452μs	100.5514 KOps/s	102.2964 KOps/s	$\color{#d91a1a}-1.71\%$
test_compile_copy_nested[pytree-eager]	0.4708ms	70.0124μs	14.2832 KOps/s	14.7194 KOps/s	$\color{#d91a1a}-2.96\%$
test_compile_add_one_flat[tensordict-compile]	0.2485ms	0.1793ms	5.5759 KOps/s	5.2510 KOps/s	$\textbf{\color{#35bf28}+6.19\%}$
test_compile_add_one_flat[tensordict-eager]	0.3808ms	0.2566ms	3.8976 KOps/s	3.7538 KOps/s	$\color{#35bf28}+3.83\%$
test_compile_add_one_flat[tensorclass-compile]	0.1940ms	0.1192ms	8.3906 KOps/s	8.1248 KOps/s	$\color{#35bf28}+3.27\%$
test_compile_add_one_flat[tensorclass-eager]	0.1261ms	69.5568μs	14.3767 KOps/s	14.0912 KOps/s	$\color{#35bf28}+2.03\%$
test_compile_add_one_flat[pytree-compile]	0.2212ms	0.1600ms	6.2495 KOps/s	6.0679 KOps/s	$\color{#35bf28}+2.99\%$
test_compile_add_one_flat[pytree-eager]	0.8920ms	0.5290ms	1.8904 KOps/s	1.8295 KOps/s	$\color{#35bf28}+3.32\%$
test_compile_add_self_flat[tensordict-eager]	0.4589ms	0.3093ms	3.2328 KOps/s	3.0870 KOps/s	$\color{#35bf28}+4.73\%$
test_compile_add_self_flat[tensordict-compile]	0.2403ms	0.1822ms	5.4877 KOps/s	5.2425 KOps/s	$\color{#35bf28}+4.68\%$
test_compile_add_self_flat[tensorclass-eager]	0.1467ms	85.7217μs	11.6657 KOps/s	11.3902 KOps/s	$\color{#35bf28}+2.42\%$
test_compile_add_self_flat[tensorclass-compile]	0.1823ms	0.1213ms	8.2468 KOps/s	8.0336 KOps/s	$\color{#35bf28}+2.65\%$
test_compile_add_self_flat[pytree-eager]	0.7024ms	0.4375ms	2.2859 KOps/s	1.8322 KOps/s	$\textbf{\color{#35bf28}+24.77\%}$
test_compile_add_self_flat[pytree-compile]	0.2230ms	0.1610ms	6.2100 KOps/s	6.0648 KOps/s	$\color{#35bf28}+2.40\%$
test_compile_copy_flat[tensordict-compile]	0.1455ms	23.5729μs	42.4216 KOps/s	38.2469 KOps/s	$\textbf{\color{#35bf28}+10.91\%}$
test_compile_copy_flat[tensordict-eager]	0.1064ms	41.6068μs	24.0345 KOps/s	23.7493 KOps/s	$\color{#35bf28}+1.20\%$
test_compile_copy_flat[pytree-compile]	63.9210μs	11.0320μs	90.6450 KOps/s	93.4306 KOps/s	$\color{#d91a1a}-2.98\%$
test_compile_copy_flat[pytree-eager]	0.4056ms	51.9494μs	19.2495 KOps/s	19.0931 KOps/s	$\color{#35bf28}+0.82\%$
test_compile_assign_and_add[tensordict-compile]	2.0656ms	0.1835ms	5.4484 KOps/s	5.3656 KOps/s	$\color{#35bf28}+1.54\%$
test_compile_assign_and_add[tensordict-eager]	3.4961ms	3.3131ms	301.8281 Ops/s	296.8796 Ops/s	$\color{#35bf28}+1.67\%$
test_compile_assign_and_add[pytree-compile]	1.9831ms	0.1626ms	6.1507 KOps/s	5.9677 KOps/s	$\color{#35bf28}+3.07\%$
test_compile_assign_and_add[pytree-eager]	2.9606ms	2.8013ms	356.9826 Ops/s	348.4601 Ops/s	$\color{#35bf28}+2.45\%$
test_compile_indexing[tensor-tensordict-compile]	0.1707ms	0.1124ms	8.8966 KOps/s	8.6883 KOps/s	$\color{#35bf28}+2.40\%$
test_compile_indexing[tensor-tensordict-eager]	0.3198ms	73.3438μs	13.6344 KOps/s	13.0560 KOps/s	$\color{#35bf28}+4.43\%$
test_compile_indexing[tensor-tensorclass-compile]	0.2390ms	0.1016ms	9.8465 KOps/s	10.1507 KOps/s	$\color{#d91a1a}-3.00\%$
test_compile_indexing[tensor-tensorclass-eager]	0.2640ms	45.4480μs	22.0032 KOps/s	21.4678 KOps/s	$\color{#35bf28}+2.49\%$
test_compile_indexing[tensor-pytree-compile]	0.1554ms	0.1036ms	9.6570 KOps/s	10.0384 KOps/s	$\color{#d91a1a}-3.80\%$
test_compile_indexing[tensor-pytree-eager]	0.2548ms	45.4154μs	22.0190 KOps/s	22.0784 KOps/s	$\color{#d91a1a}-0.27\%$
test_compile_indexing[slice-tensordict-compile]	0.1055ms	56.5880μs	17.6716 KOps/s	17.2491 KOps/s	$\color{#35bf28}+2.45\%$
test_compile_indexing[slice-tensordict-eager]	0.2186ms	28.4199μs	35.1866 KOps/s	35.6141 KOps/s	$\color{#d91a1a}-1.20\%$
test_compile_indexing[slice-tensorclass-compile]	0.1074ms	45.7301μs	21.8674 KOps/s	21.9837 KOps/s	$\color{#d91a1a}-0.53\%$
test_compile_indexing[slice-tensorclass-eager]	0.2602ms	22.5128μs	44.4191 KOps/s	44.4752 KOps/s	$\color{#d91a1a}-0.13\%$
test_compile_indexing[slice-pytree-compile]	83.4110μs	45.9758μs	21.7506 KOps/s	21.5551 KOps/s	$\color{#35bf28}+0.91\%$
test_compile_indexing[slice-pytree-eager]	0.2715ms	22.5451μs	44.3555 KOps/s	44.3825 KOps/s	$\color{#d91a1a}-0.06\%$
test_compile_indexing[int-tensordict-compile]	0.1014ms	57.6009μs	17.3608 KOps/s	17.0898 KOps/s	$\color{#35bf28}+1.59\%$
test_compile_indexing[int-tensordict-eager]	0.2354ms	28.3772μs	35.2396 KOps/s	35.4238 KOps/s	$\color{#d91a1a}-0.52\%$
test_compile_indexing[int-tensorclass-compile]	95.9610μs	46.4561μs	21.5257 KOps/s	21.5092 KOps/s	$\color{#35bf28}+0.08\%$
test_compile_indexing[int-tensorclass-eager]	0.2602ms	22.5402μs	44.3652 KOps/s	44.2483 KOps/s	$\color{#35bf28}+0.26\%$
test_compile_indexing[int-pytree-compile]	82.6820μs	46.0722μs	21.7051 KOps/s	20.8411 KOps/s	$\color{#35bf28}+4.15\%$
test_compile_indexing[int-pytree-eager]	0.2560ms	22.7018μs	44.0495 KOps/s	44.1107 KOps/s	$\color{#d91a1a}-0.14\%$
test_mod_add[eager]	95.8220μs	51.3667μs	19.4679 KOps/s	18.1954 KOps/s	$\textbf{\color{#35bf28}+6.99\%}$
test_mod_add[compile]	0.1828ms	0.1049ms	9.5337 KOps/s	9.2834 KOps/s	$\color{#35bf28}+2.70\%$
test_mod_add[compile-overhead]	0.2571ms	0.1480ms	6.7569 KOps/s	6.5492 KOps/s	$\color{#35bf28}+3.17\%$
test_mod_wrap[eager]	0.3706ms	0.2953ms	3.3863 KOps/s	3.2830 KOps/s	$\color{#35bf28}+3.15\%$
test_mod_wrap[compile]	0.4368ms	0.3627ms	2.7569 KOps/s	2.6855 KOps/s	$\color{#35bf28}+2.66\%$
test_mod_wrap[compile-overhead]	7.2599ms	4.0084ms	249.4737 Ops/s	251.6971 Ops/s	$\color{#d91a1a}-0.88\%$
test_mod_wrap_and_backward[eager]	1.7709ms	1.6331ms	612.3472 Ops/s	611.4631 Ops/s	$\color{#35bf28}+0.14\%$
test_mod_wrap_and_backward[compile]	1.6393ms	1.5481ms	645.9625 Ops/s	674.6945 Ops/s	$\color{#d91a1a}-4.26\%$
test_mod_wrap_and_backward[compile-overhead]	1.2934ms	0.9846ms	1.0156 KOps/s	1.1020 KOps/s	$\textbf{\color{#d91a1a}-7.84\%}$
test_seq_add[eager]	0.2708ms	0.1580ms	6.3278 KOps/s	6.2103 KOps/s	$\color{#35bf28}+1.89\%$
test_seq_add[compile]	0.1729ms	0.1158ms	8.6379 KOps/s	8.3691 KOps/s	$\color{#35bf28}+3.21\%$
test_seq_add[compile-overhead]	0.2313ms	0.1550ms	6.4529 KOps/s	6.2774 KOps/s	$\color{#35bf28}+2.80\%$
test_seq_wrap[eager]	0.6024ms	0.5292ms	1.8895 KOps/s	1.8660 KOps/s	$\color{#35bf28}+1.26\%$
test_seq_wrap[compile]	0.4657ms	0.3816ms	2.6205 KOps/s	2.6672 KOps/s	$\color{#d91a1a}-1.75\%$
test_seq_wrap[compile-overhead]	0.3306ms	0.2655ms	3.7658 KOps/s	3.7278 KOps/s	$\color{#35bf28}+1.02\%$
test_func_call_runtime[False-eager]	0.9488ms	0.8491ms	1.1777 KOps/s	1.1687 KOps/s	$\color{#35bf28}+0.77\%$
test_func_call_runtime[False-compile]	1.1123ms	0.9131ms	1.0951 KOps/s	1.0762 KOps/s	$\color{#35bf28}+1.76\%$
test_func_call_runtime[False-compile-overhead]	0.6054ms	0.4601ms	2.1737 KOps/s	2.1401 KOps/s	$\color{#35bf28}+1.57\%$
test_func_call_runtime[True-eager]	1.2796ms	1.0896ms	917.7396 Ops/s	905.5765 Ops/s	$\color{#35bf28}+1.34\%$
test_func_call_runtime[True-compile]	1.1100ms	0.9261ms	1.0798 KOps/s	1.0602 KOps/s	$\color{#35bf28}+1.84\%$
test_func_call_runtime[True-compile-overhead]	0.6271ms	0.4759ms	2.1013 KOps/s	2.0595 KOps/s	$\color{#35bf28}+2.03\%$
test_func_call_cm_runtime[False-eager]	0.9796ms	0.8530ms	1.1724 KOps/s	1.1721 KOps/s	$\color{#35bf28}+0.02\%$
test_func_call_cm_runtime[False-compile]	1.1656ms	0.9161ms	1.0916 KOps/s	1.0704 KOps/s	$\color{#35bf28}+1.98\%$
test_func_call_cm_runtime[False-compile-overhead]	0.5575ms	0.4659ms	2.1462 KOps/s	2.1246 KOps/s	$\color{#35bf28}+1.02\%$
test_func_call_cm_runtime[True-eager]	1.3689ms	1.2329ms	811.1066 Ops/s	796.7297 Ops/s	$\color{#35bf28}+1.80\%$
test_func_call_cm_runtime[True-compile]	1.2319ms	0.9601ms	1.0415 KOps/s	1.0244 KOps/s	$\color{#35bf28}+1.67\%$
test_func_call_cm_runtime[True-compile-overhead]	0.5772ms	0.5085ms	1.9665 KOps/s	1.9288 KOps/s	$\color{#35bf28}+1.95\%$
test_vmap_func_call_cm_runtime[eager]	2.8771ms	2.3754ms	420.9759 Ops/s	415.5628 Ops/s	$\color{#35bf28}+1.30\%$
test_vmap_func_call_cm_runtime[compile]	1.4423ms	0.9765ms	1.0240 KOps/s	1.0036 KOps/s	$\color{#35bf28}+2.03\%$
test_vmap_func_call_cm_runtime[compile-overhead]	0.6033ms	0.5179ms	1.9309 KOps/s	1.9071 KOps/s	$\color{#35bf28}+1.25\%$
test_distributed	0.5562ms	0.1559ms	6.4151 KOps/s	6.0267 KOps/s	$\textbf{\color{#35bf28}+6.44\%}$
test_tdmodule	56.9310μs	28.5873μs	34.9805 KOps/s	33.4947 KOps/s	$\color{#35bf28}+4.44\%$
test_tdmodule_dispatch	76.8510μs	46.5002μs	21.5053 KOps/s	21.0842 KOps/s	$\color{#35bf28}+2.00\%$
test_tdseq	49.5710μs	27.4797μs	36.3905 KOps/s	36.0686 KOps/s	$\color{#35bf28}+0.89\%$
test_tdseq_dispatch	84.4710μs	48.4127μs	20.6557 KOps/s	20.3030 KOps/s	$\color{#35bf28}+1.74\%$
test_instantiation_functorch	2.1572ms	2.0547ms	486.6906 Ops/s	479.1383 Ops/s	$\color{#35bf28}+1.58\%$
test_exec_functorch	0.2484ms	0.1798ms	5.5609 KOps/s	5.4946 KOps/s	$\color{#35bf28}+1.21\%$
test_exec_functional_call	0.2234ms	0.1629ms	6.1380 KOps/s	6.2068 KOps/s	$\color{#d91a1a}-1.11\%$
test_exec_td_decorator	0.4444ms	0.2352ms	4.2524 KOps/s	4.1743 KOps/s	$\color{#35bf28}+1.87\%$
test_vmap_mlp_speed_decorator[True-True]	1.0182ms	0.8242ms	1.2134 KOps/s	1.1933 KOps/s	$\color{#35bf28}+1.68\%$
test_vmap_mlp_speed_decorator[True-False]	1.0022ms	0.8230ms	1.2150 KOps/s	1.1955 KOps/s	$\color{#35bf28}+1.63\%$
test_vmap_mlp_speed_decorator[False-True]	0.9262ms	0.7124ms	1.4038 KOps/s	1.3828 KOps/s	$\color{#35bf28}+1.51\%$
test_vmap_mlp_speed_decorator[False-False]	0.9131ms	0.7168ms	1.3950 KOps/s	1.3798 KOps/s	$\color{#35bf28}+1.10\%$
test_vmap_transformer_speed_decorator[True-True]	21.4903ms	20.6609ms	48.4006 Ops/s	48.1414 Ops/s	$\color{#35bf28}+0.54\%$
test_vmap_transformer_speed_decorator[True-False]	21.2953ms	20.6013ms	48.5406 Ops/s	48.1103 Ops/s	$\color{#35bf28}+0.89\%$
test_vmap_transformer_speed_decorator[False-True]	21.0924ms	20.3711ms	49.0892 Ops/s	48.6067 Ops/s	$\color{#35bf28}+0.99\%$
test_vmap_transformer_speed_decorator[False-False]	20.9951ms	20.3883ms	49.0478 Ops/s	48.5230 Ops/s	$\color{#35bf28}+1.08\%$
test_to_module_speed[True]	2.0448ms	1.4747ms	678.0883 Ops/s	669.6816 Ops/s	$\color{#35bf28}+1.26\%$
test_to_module_speed[False]	1.9644ms	1.4464ms	691.3759 Ops/s	686.8315 Ops/s	$\color{#35bf28}+0.66\%$
test_tc_init	0.1130ms	46.3185μs	21.5896 KOps/s	21.3095 KOps/s	$\color{#35bf28}+1.31\%$
test_tc_init_tensor_only	0.4200ms	9.8377μs	101.6498 KOps/s	100.3485 KOps/s	$\color{#35bf28}+1.30\%$
test_tc_init_nested	0.1386ms	92.0386μs	10.8650 KOps/s	10.7582 KOps/s	$\color{#35bf28}+0.99\%$
test_tc_init_many_fields	68.3610μs	16.5271μs	60.5065 KOps/s	59.2717 KOps/s	$\color{#35bf28}+2.08\%$
test_tc_first_layer_tensor	27.1810μs	1.8496μs	540.6550 KOps/s	531.7546 KOps/s	$\color{#35bf28}+1.67\%$
test_tc_first_layer_tensor_only	8.5301μs	0.7569μs	1.3212 MOps/s	1.3097 MOps/s	$\color{#35bf28}+0.87\%$
test_tc_first_layer_tensor_set	26.0300μs	4.2018μs	237.9954 KOps/s	229.9180 KOps/s	$\color{#35bf28}+3.51\%$
test_tc_first_layer_tensor_only_set	48.6600μs	3.2022μs	312.2824 KOps/s	312.0821 KOps/s	$\color{#35bf28}+0.06\%$
test_tc_first_layer_nontensor	43.6500μs	6.2149μs	160.9037 KOps/s	158.0249 KOps/s	$\color{#35bf28}+1.82\%$
test_tc_second_layer_tensor	65.1710μs	4.5138μs	221.5408 KOps/s	222.4747 KOps/s	$\color{#d91a1a}-0.42\%$
test_tc_second_layer_nontensor	41.0400μs	8.8609μs	112.8558 KOps/s	111.1884 KOps/s	$\color{#35bf28}+1.50\%$
test_unbind	0.2597s	16.0434ms	62.3309 Ops/s	57.0841 Ops/s	$\textbf{\color{#35bf28}+9.19\%}$
test_full_like	4.8676ms	4.3731ms	228.6730 Ops/s	227.8673 Ops/s	$\color{#35bf28}+0.35\%$
test_zeros_like	4.8571ms	4.3647ms	229.1087 Ops/s	229.3249 Ops/s	$\color{#d91a1a}-0.09\%$
test_ones_like	4.7639ms	4.3706ms	228.8004 Ops/s	228.7921 Ops/s	$+0.00\%$
test_clone	6.6302ms	6.4414ms	155.2457 Ops/s	155.3133 Ops/s	$\color{#d91a1a}-0.04\%$
test_squeeze	69.8310μs	15.7133μs	63.6403 KOps/s	70.0136 KOps/s	$\textbf{\color{#d91a1a}-9.10\%}$
test_unsqueeze	0.5526ms	0.1132ms	8.8336 KOps/s	9.1136 KOps/s	$\color{#d91a1a}-3.07\%$
test_split	0.3666ms	0.1910ms	5.2346 KOps/s	5.4595 KOps/s	$\color{#d91a1a}-4.12\%$
test_permute	0.7832ms	0.2111ms	4.7377 KOps/s	4.9190 KOps/s	$\color{#d91a1a}-3.68\%$
test_stack	51.6454ms	51.2433ms	19.5147 Ops/s	19.5604 Ops/s	$\color{#d91a1a}-0.23\%$
test_cat	51.4113ms	51.2002ms	19.5312 Ops/s	19.5760 Ops/s	$\color{#d91a1a}-0.23\%$

[github-actions]

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

Total Benchmarks: 243. Improved: $\large\color{#35bf28}9$. Worsened: $\large\color{#d91a1a}5$.

Expand to view detailed results

Name	Max	Mean	Ops	Ops on Repo HEAD	Change
test_plain_set_nested	41.1020μs	14.8039μs	67.5499 KOps/s	66.3922 KOps/s	$\color{#35bf28}+1.74\%$
test_plain_set_stack_nested	39.9420μs	15.1255μs	66.1135 KOps/s	65.4059 KOps/s	$\color{#35bf28}+1.08\%$
test_plain_set_nested_inplace	41.2020μs	16.5429μs	60.4491 KOps/s	59.1621 KOps/s	$\color{#35bf28}+2.18\%$
test_plain_set_stack_nested_inplace	41.2020μs	16.4431μs	60.8157 KOps/s	59.2763 KOps/s	$\color{#35bf28}+2.60\%$
test_items	29.1520μs	5.7094μs	175.1488 KOps/s	172.2860 KOps/s	$\color{#35bf28}+1.66\%$
test_items_nested	0.6139ms	0.5323ms	1.8787 KOps/s	1.8620 KOps/s	$\color{#35bf28}+0.90\%$
test_items_nested_locked	0.5969ms	0.5368ms	1.8628 KOps/s	1.8523 KOps/s	$\color{#35bf28}+0.56\%$
test_items_nested_leaf	0.1502ms	95.6673μs	10.4529 KOps/s	10.4474 KOps/s	$\color{#35bf28}+0.05\%$
test_items_stack_nested	0.6205ms	0.5368ms	1.8629 KOps/s	1.8653 KOps/s	$\color{#d91a1a}-0.13\%$
test_items_stack_nested_leaf	0.1529ms	95.7700μs	10.4417 KOps/s	10.0847 KOps/s	$\color{#35bf28}+3.54\%$
test_items_stack_nested_locked	0.5925ms	0.5401ms	1.8514 KOps/s	1.8267 KOps/s	$\color{#35bf28}+1.35\%$
test_keys	30.5520μs	4.2016μs	238.0046 KOps/s	238.7535 KOps/s	$\color{#d91a1a}-0.31\%$
test_keys_nested	0.1642ms	0.1190ms	8.4022 KOps/s	8.2091 KOps/s	$\color{#35bf28}+2.35\%$
test_keys_nested_locked	89.0343ms	0.1393ms	7.1799 KOps/s	7.5516 KOps/s	$\color{#d91a1a}-4.92\%$
test_keys_nested_leaf	0.1511ms	0.1097ms	9.1161 KOps/s	8.8741 KOps/s	$\color{#35bf28}+2.73\%$
test_keys_stack_nested	0.1647ms	0.1193ms	8.3823 KOps/s	8.2176 KOps/s	$\color{#35bf28}+2.00\%$
test_keys_stack_nested_leaf	0.1402ms	0.1099ms	9.1032 KOps/s	8.9258 KOps/s	$\color{#35bf28}+1.99\%$
test_keys_stack_nested_locked	0.1603ms	0.1272ms	7.8635 KOps/s	7.6686 KOps/s	$\color{#35bf28}+2.54\%$
test_values	7.0924μs	1.0109μs	989.2086 KOps/s	976.6508 KOps/s	$\color{#35bf28}+1.29\%$
test_values_nested	75.9940μs	47.6666μs	20.9791 KOps/s	20.7713 KOps/s	$\color{#35bf28}+1.00\%$
test_values_nested_locked	79.3940μs	50.2585μs	19.8971 KOps/s	19.4220 KOps/s	$\color{#35bf28}+2.45\%$
test_values_nested_leaf	85.2940μs	53.9334μs	18.5414 KOps/s	18.2193 KOps/s	$\color{#35bf28}+1.77\%$
test_values_stack_nested	1.0662ms	47.3207μs	21.1324 KOps/s	20.8430 KOps/s	$\color{#35bf28}+1.39\%$
test_values_stack_nested_leaf	80.8040μs	53.9359μs	18.5405 KOps/s	18.3605 KOps/s	$\color{#35bf28}+0.98\%$
test_values_stack_nested_locked	87.5840μs	50.6828μs	19.7305 KOps/s	19.6024 KOps/s	$\color{#35bf28}+0.65\%$
test_membership	5.4953μs	0.8499μs	1.1766 MOps/s	1.1984 MOps/s	$\color{#d91a1a}-1.81\%$
test_membership_nested	32.2520μs	3.1503μs	317.4257 KOps/s	317.9781 KOps/s	$\color{#d91a1a}-0.17\%$
test_membership_nested_leaf	41.4120μs	3.1555μs	316.9096 KOps/s	317.5069 KOps/s	$\color{#d91a1a}-0.19\%$
test_membership_stacked_nested	28.1010μs	3.2004μs	312.4568 KOps/s	315.6988 KOps/s	$\color{#d91a1a}-1.03\%$
test_membership_stacked_nested_leaf	34.3610μs	3.1833μs	314.1415 KOps/s	315.2285 KOps/s	$\color{#d91a1a}-0.34\%$
test_membership_nested_last	33.3920μs	4.5922μs	217.7585 KOps/s	216.4191 KOps/s	$\color{#35bf28}+0.62\%$
test_membership_nested_leaf_last	32.4720μs	4.6158μs	216.6486 KOps/s	220.3237 KOps/s	$\color{#d91a1a}-1.67\%$
test_membership_stacked_nested_last	35.9120μs	4.6508μs	215.0173 KOps/s	219.0620 KOps/s	$\color{#d91a1a}-1.85\%$
test_membership_stacked_nested_leaf_last	35.5620μs	4.5575μs	219.4168 KOps/s	218.7080 KOps/s	$\color{#35bf28}+0.32\%$
test_nested_getleaf	54.5530μs	21.7188μs	46.0431 KOps/s	45.8114 KOps/s	$\color{#35bf28}+0.51\%$
test_nested_get	49.4930μs	20.2324μs	49.4258 KOps/s	48.8635 KOps/s	$\color{#35bf28}+1.15\%$
test_stacked_getleaf	55.7530μs	21.3837μs	46.7647 KOps/s	46.4175 KOps/s	$\color{#35bf28}+0.75\%$
test_stacked_get	43.1720μs	20.6730μs	48.3722 KOps/s	48.7358 KOps/s	$\color{#d91a1a}-0.75\%$
test_nested_getitemleaf	50.9230μs	21.8250μs	45.8191 KOps/s	44.8081 KOps/s	$\color{#35bf28}+2.26\%$
test_nested_getitem	49.2620μs	21.0831μs	47.4313 KOps/s	47.9748 KOps/s	$\color{#d91a1a}-1.13\%$
test_stacked_getitemleaf	46.1520μs	22.2885μs	44.8661 KOps/s	44.6298 KOps/s	$\color{#35bf28}+0.53\%$
test_stacked_getitem	50.9520μs	21.2884μs	46.9740 KOps/s	47.2890 KOps/s	$\color{#d91a1a}-0.67\%$
test_lock_nested	7.5117ms	0.4830ms	2.0705 KOps/s	2.1019 KOps/s	$\color{#d91a1a}-1.49\%$
test_lock_stack_nested	0.5230ms	0.4818ms	2.0755 KOps/s	2.0548 KOps/s	$\color{#35bf28}+1.01\%$
test_unlock_nested	0.4534ms	0.3855ms	2.5938 KOps/s	2.5932 KOps/s	$\color{#35bf28}+0.02\%$
test_unlock_stack_nested	0.4251ms	0.3870ms	2.5839 KOps/s	2.5585 KOps/s	$\color{#35bf28}+1.00\%$
test_flatten_speed	0.1595ms	0.1219ms	8.2051 KOps/s	8.1081 KOps/s	$\color{#35bf28}+1.20\%$
test_unflatten_speed	0.6509ms	0.5939ms	1.6837 KOps/s	1.6719 KOps/s	$\color{#35bf28}+0.71\%$
test_common_ops	0.8080ms	0.6812ms	1.4680 KOps/s	1.4651 KOps/s	$\color{#35bf28}+0.20\%$
test_creation	72.3640μs	2.9212μs	342.3281 KOps/s	343.7786 KOps/s	$\color{#d91a1a}-0.42\%$
test_creation_empty	28.9420μs	6.1357μs	162.9814 KOps/s	161.8800 KOps/s	$\color{#35bf28}+0.68\%$
test_creation_nested_1	36.6620μs	10.8540μs	92.1318 KOps/s	91.9630 KOps/s	$\color{#35bf28}+0.18\%$
test_creation_nested_2	50.4030μs	11.9434μs	83.7285 KOps/s	83.8562 KOps/s	$\color{#d91a1a}-0.15\%$
test_creation_many_keys[10]	47.3520μs	17.9759μs	55.6302 KOps/s	54.8379 KOps/s	$\color{#35bf28}+1.44\%$
test_creation_many_keys[50]	0.1300ms	77.5863μs	12.8889 KOps/s	12.8414 KOps/s	$\color{#35bf28}+0.37\%$
test_creation_many_keys[100]	0.2314ms	0.1534ms	6.5184 KOps/s	6.5556 KOps/s	$\color{#d91a1a}-0.57\%$
test_creation_nested_many_keys[10]	72.7640μs	39.2096μs	25.5039 KOps/s	25.7431 KOps/s	$\color{#d91a1a}-0.93\%$
test_creation_nested_many_keys[50]	0.2009ms	0.1593ms	6.2758 KOps/s	6.2821 KOps/s	$\color{#d91a1a}-0.10\%$
test_clone	40.2820μs	13.3023μs	75.1749 KOps/s	76.0248 KOps/s	$\color{#d91a1a}-1.12\%$
test_getitem[int]	1.6991ms	14.7760μs	67.6771 KOps/s	60.7592 KOps/s	$\textbf{\color{#35bf28}+11.39\%}$
test_getitem[slice_int]	0.1439ms	25.2344μs	39.6284 KOps/s	38.0171 KOps/s	$\color{#35bf28}+4.24\%$
test_getitem[range]	0.1727ms	61.3349μs	16.3039 KOps/s	15.4026 KOps/s	$\textbf{\color{#35bf28}+5.85\%}$
test_getitem[tuple]	0.1502ms	24.5988μs	40.6525 KOps/s	39.1156 KOps/s	$\color{#35bf28}+3.93\%$
test_getitem[list]	0.1825ms	56.7717μs	17.6144 KOps/s	16.9334 KOps/s	$\color{#35bf28}+4.02\%$
test_setitem_dim[int]	53.7230μs	26.3676μs	37.9254 KOps/s	36.9508 KOps/s	$\color{#35bf28}+2.64\%$
test_setitem_dim[slice_int]	66.4440μs	43.8949μs	22.7817 KOps/s	21.2756 KOps/s	$\textbf{\color{#35bf28}+7.08\%}$
test_setitem_dim[range]	0.1226ms	93.8977μs	10.6499 KOps/s	10.7048 KOps/s	$\color{#d91a1a}-0.51\%$
test_setitem_dim[tuple]	68.4130μs	41.0424μs	24.3650 KOps/s	24.1905 KOps/s	$\color{#35bf28}+0.72\%$
test_setitem	46.9920μs	17.6954μs	56.5119 KOps/s	55.7334 KOps/s	$\color{#35bf28}+1.40\%$
test_set	52.1630μs	17.1859μs	58.1871 KOps/s	58.1212 KOps/s	$\color{#35bf28}+0.11\%$
test_set_shared	0.5418ms	0.2042ms	4.8980 KOps/s	4.7682 KOps/s	$\color{#35bf28}+2.72\%$
test_update	0.3871ms	22.4336μs	44.5761 KOps/s	45.2067 KOps/s	$\color{#d91a1a}-1.40\%$
test_update_nested	71.0330μs	34.6431μs	28.8658 KOps/s	28.8527 KOps/s	$\color{#35bf28}+0.05\%$
test_update__nested	0.4741ms	34.3347μs	29.1251 KOps/s	28.7162 KOps/s	$\color{#35bf28}+1.42\%$
test_set_nested	56.3830μs	19.0132μs	52.5949 KOps/s	52.2120 KOps/s	$\color{#35bf28}+0.73\%$
test_set_nested_new	68.9630μs	23.9491μs	41.7552 KOps/s	41.5631 KOps/s	$\color{#35bf28}+0.46\%$
test_select	73.1830μs	42.0124μs	23.8025 KOps/s	23.8242 KOps/s	$\color{#d91a1a}-0.09\%$
test_select_nested	0.1171ms	75.5933μs	13.2287 KOps/s	13.4230 KOps/s	$\color{#d91a1a}-1.45\%$
test_exclude_nested	0.1282ms	97.2427μs	10.2835 KOps/s	10.2995 KOps/s	$\color{#d91a1a}-0.16\%$
test_empty[True]	0.4801ms	0.4383ms	2.2815 KOps/s	2.2811 KOps/s	$\color{#35bf28}+0.02\%$
test_empty[False]	8.1205μs	1.3211μs	756.9307 KOps/s	755.5254 KOps/s	$\color{#35bf28}+0.19\%$
test_to	0.1028ms	72.5689μs	13.7800 KOps/s	13.7452 KOps/s	$\color{#35bf28}+0.25\%$
test_to_nonblocking	0.1168ms	64.1186μs	15.5961 KOps/s	15.4219 KOps/s	$\color{#35bf28}+1.13\%$
test_unbind_speed	0.3829ms	0.3307ms	3.0242 KOps/s	3.0912 KOps/s	$\color{#d91a1a}-2.17\%$
test_unbind_speed_stack0	0.4087ms	0.3297ms	3.0329 KOps/s	3.0853 KOps/s	$\color{#d91a1a}-1.70\%$
test_unbind_speed_stack1	0.1036s	0.8423ms	1.1872 KOps/s	1.1822 KOps/s	$\color{#35bf28}+0.42\%$
test_split	0.1030s	1.2798ms	781.3985 Ops/s	789.0487 Ops/s	$\color{#d91a1a}-0.97\%$
test_chunk	0.1028s	1.2174ms	821.4535 Ops/s	923.9799 Ops/s	$\textbf{\color{#d91a1a}-11.10\%}$
test_to_cpu_blocking	19.4532ms	19.3358ms	51.7174 Ops/s	47.2084 Ops/s	$\textbf{\color{#35bf28}+9.55\%}$
test_to_cpu_global_sync	11.3406ms	11.2148ms	89.1678 Ops/s	90.4132 Ops/s	$\color{#d91a1a}-1.38\%$
test_to_cpu_event_sync	12.3965ms	12.1800ms	82.1017 Ops/s	82.9855 Ops/s	$\color{#d91a1a}-1.07\%$
test_to_cpu_default	0.1145s	13.4246ms	74.4901 Ops/s	83.1621 Ops/s	$\textbf{\color{#d91a1a}-10.43\%}$
test_consolidate[False-None]	4.3190ms	4.1138ms	243.0860 Ops/s	220.3182 Ops/s	$\textbf{\color{#35bf28}+10.33\%}$
test_consolidate[default-None]	2.1310ms	2.0325ms	492.0016 Ops/s	484.9602 Ops/s	$\color{#35bf28}+1.45\%$
test_consolidate[reduce-overhead-None]	2.0454ms	1.9552ms	511.4439 Ops/s	506.2937 Ops/s	$\color{#35bf28}+1.02\%$
test_consolidate_njt[False-None]	8.6283ms	8.4527ms	118.3055 Ops/s	118.5736 Ops/s	$\color{#d91a1a}-0.23\%$
test_to[False-False-None]	2.1667ms	2.0652ms	484.2058 Ops/s	486.7050 Ops/s	$\color{#d91a1a}-0.51\%$
test_to[True-False-None]	2.1575ms	1.9097ms	523.6349 Ops/s	511.7328 Ops/s	$\color{#35bf28}+2.33\%$
test_to[within-False-None]	6.3219ms	6.0998ms	163.9401 Ops/s	164.1838 Ops/s	$\color{#d91a1a}-0.15\%$
test_to[True-default-None]	0.1769s	8.8406ms	113.1139 Ops/s	125.0978 Ops/s	$\textbf{\color{#d91a1a}-9.58\%}$
test_to_njt[False-False-None]	8.8518ms	8.5440ms	117.0413 Ops/s	116.2744 Ops/s	$\color{#35bf28}+0.66\%$
test_to_njt[True-False-None]	7.2936ms	6.9321ms	144.2560 Ops/s	141.9425 Ops/s	$\color{#35bf28}+1.63\%$
test_to_njt[within-False-None]	16.3893ms	15.4682ms	64.6486 Ops/s	63.6290 Ops/s	$\color{#35bf28}+1.60\%$
test_creation[device0]	0.2550ms	0.1161ms	8.6099 KOps/s	8.2013 KOps/s	$\color{#35bf28}+4.98\%$
test_creation_from_tensor	0.3603ms	0.1162ms	8.6043 KOps/s	8.5467 KOps/s	$\color{#35bf28}+0.67\%$
test_add_one[memmap_tensor0]	0.2791ms	6.5394μs	152.9196 KOps/s	152.1972 KOps/s	$\color{#35bf28}+0.47\%$
test_contiguous[memmap_tensor0]	15.8500μs	0.6535μs	1.5303 MOps/s	2.1397 MOps/s	$\textbf{\color{#d91a1a}-28.48\%}$
test_stack[memmap_tensor0]	60.5340μs	4.6228μs	216.3198 KOps/s	221.9371 KOps/s	$\color{#d91a1a}-2.53\%$
test_memmaptd_index	1.0250ms	0.2637ms	3.7924 KOps/s	3.7694 KOps/s	$\color{#35bf28}+0.61\%$
test_memmaptd_index_astensor	0.5156ms	0.3610ms	2.7704 KOps/s	2.8022 KOps/s	$\color{#d91a1a}-1.14\%$
test_memmaptd_index_op	0.9048ms	0.6048ms	1.6534 KOps/s	1.6526 KOps/s	$\color{#35bf28}+0.05\%$
test_serialize_model	0.1404s	0.1358s	7.3625 Ops/s	7.3157 Ops/s	$\color{#35bf28}+0.64\%$
test_serialize_model_pickle	1.3484s	1.2098s	0.8266 Ops/s	0.8260 Ops/s	$\color{#35bf28}+0.07\%$
test_serialize_weights	0.1386s	0.1355s	7.3793 Ops/s	7.3625 Ops/s	$\color{#35bf28}+0.23\%$
test_serialize_weights_returnearly	0.4051s	91.0123ms	10.9875 Ops/s	10.3534 Ops/s	$\textbf{\color{#35bf28}+6.13\%}$
test_serialize_weights_pickle	1.3584s	1.2145s	0.8234 Ops/s	0.8210 Ops/s	$\color{#35bf28}+0.29\%$
test_reshape_pytree	0.2068ms	33.6047μs	29.7578 KOps/s	30.0986 KOps/s	$\color{#d91a1a}-1.13\%$
test_reshape_td	89.6550μs	45.1162μs	22.1650 KOps/s	22.6971 KOps/s	$\color{#d91a1a}-2.34\%$
test_view_pytree	0.2398ms	33.2681μs	30.0588 KOps/s	30.3642 KOps/s	$\color{#d91a1a}-1.01\%$
test_view_td	92.7950μs	52.8714μs	18.9138 KOps/s	19.0357 KOps/s	$\color{#d91a1a}-0.64\%$
test_unbind_pytree	0.2360ms	37.1907μs	26.8885 KOps/s	27.0006 KOps/s	$\color{#d91a1a}-0.42\%$
test_unbind_td	0.1309ms	49.2682μs	20.2971 KOps/s	20.3649 KOps/s	$\color{#d91a1a}-0.33\%$
test_split_pytree	0.2592ms	43.0572μs	23.2249 KOps/s	23.3264 KOps/s	$\color{#d91a1a}-0.43\%$
test_split_td	0.2223ms	66.7872μs	14.9729 KOps/s	15.3116 KOps/s	$\color{#d91a1a}-2.21\%$
test_add_pytree	0.2366ms	42.7689μs	23.3815 KOps/s	23.9051 KOps/s	$\color{#d91a1a}-2.19\%$
test_add_td	99.3050μs	53.7840μs	18.5929 KOps/s	18.5222 KOps/s	$\color{#35bf28}+0.38\%$
test_compile_add_one_nested[tensordict-compile]	0.1917ms	0.1388ms	7.2046 KOps/s	6.9835 KOps/s	$\color{#35bf28}+3.17\%$
test_compile_add_one_nested[tensordict-eager]	0.6047ms	0.1930ms	5.1811 KOps/s	5.1779 KOps/s	$\color{#35bf28}+0.06\%$
test_compile_add_one_nested[pytree-compile]	0.1721ms	0.1087ms	9.1966 KOps/s	9.0752 KOps/s	$\color{#35bf28}+1.34\%$
test_compile_add_one_nested[pytree-eager]	0.6122ms	0.1828ms	5.4695 KOps/s	5.5508 KOps/s	$\color{#d91a1a}-1.46\%$
test_compile_copy_nested[tensordict-compile]	68.3140μs	30.0555μs	33.2718 KOps/s	32.8376 KOps/s	$\color{#35bf28}+1.32\%$
test_compile_copy_nested[tensordict-eager]	0.1268ms	52.3918μs	19.0870 KOps/s	18.6722 KOps/s	$\color{#35bf28}+2.22\%$
test_compile_copy_nested[pytree-compile]	41.8320μs	9.7931μs	102.1126 KOps/s	105.0910 KOps/s	$\color{#d91a1a}-2.83\%$
test_compile_copy_nested[pytree-eager]	0.4550ms	70.2780μs	14.2292 KOps/s	14.4065 KOps/s	$\color{#d91a1a}-1.23\%$
test_compile_add_one_flat[tensordict-compile]	0.2763ms	0.1760ms	5.6818 KOps/s	5.4825 KOps/s	$\color{#35bf28}+3.64\%$
test_compile_add_one_flat[tensordict-eager]	0.3201ms	0.2528ms	3.9550 KOps/s	3.9086 KOps/s	$\color{#35bf28}+1.19\%$
test_compile_add_one_flat[tensorclass-compile]	0.2501ms	0.1158ms	8.6392 KOps/s	8.3385 KOps/s	$\color{#35bf28}+3.61\%$
test_compile_add_one_flat[tensorclass-eager]	0.1277ms	68.5560μs	14.5866 KOps/s	14.2765 KOps/s	$\color{#35bf28}+2.17\%$
test_compile_add_one_flat[pytree-compile]	0.2034ms	0.1588ms	6.2973 KOps/s	6.1230 KOps/s	$\color{#35bf28}+2.85\%$
test_compile_add_one_flat[pytree-eager]	0.8311ms	0.5428ms	1.8423 KOps/s	1.8783 KOps/s	$\color{#d91a1a}-1.92\%$
test_compile_add_self_flat[tensordict-eager]	0.3588ms	0.3059ms	3.2687 KOps/s	3.1835 KOps/s	$\color{#35bf28}+2.68\%$
test_compile_add_self_flat[tensordict-compile]	0.2332ms	0.1797ms	5.5650 KOps/s	5.1982 KOps/s	$\textbf{\color{#35bf28}+7.06\%}$
test_compile_add_self_flat[tensorclass-eager]	0.1293ms	84.9662μs	11.7694 KOps/s	11.5265 KOps/s	$\color{#35bf28}+2.11\%$
test_compile_add_self_flat[tensorclass-compile]	0.1691ms	0.1181ms	8.4644 KOps/s	7.8869 KOps/s	$\textbf{\color{#35bf28}+7.32\%}$
test_compile_add_self_flat[pytree-eager]	0.6600ms	0.4435ms	2.2549 KOps/s	2.3019 KOps/s	$\color{#d91a1a}-2.04\%$
test_compile_add_self_flat[pytree-compile]	0.1999ms	0.1596ms	6.2672 KOps/s	6.1310 KOps/s	$\color{#35bf28}+2.22\%$
test_compile_copy_flat[tensordict-compile]	53.8430μs	24.3683μs	41.0369 KOps/s	41.2146 KOps/s	$\color{#d91a1a}-0.43\%$
test_compile_copy_flat[tensordict-eager]	0.1393ms	41.5300μs	24.0790 KOps/s	24.2078 KOps/s	$\color{#d91a1a}-0.53\%$
test_compile_copy_flat[pytree-compile]	36.7120μs	10.9684μs	91.1711 KOps/s	90.0701 KOps/s	$\color{#35bf28}+1.22\%$
test_compile_copy_flat[pytree-eager]	0.4169ms	52.9486μs	18.8862 KOps/s	19.2205 KOps/s	$\color{#d91a1a}-1.74\%$
test_compile_assign_and_add[tensordict-compile]	1.9986ms	0.1744ms	5.7348 KOps/s	5.5009 KOps/s	$\color{#35bf28}+4.25\%$
test_compile_assign_and_add[tensordict-eager]	3.4202ms	3.3156ms	301.6059 Ops/s	301.9230 Ops/s	$\color{#d91a1a}-0.11\%$
test_compile_assign_and_add[pytree-compile]	1.9817ms	0.1623ms	6.1608 KOps/s	6.1002 KOps/s	$\color{#35bf28}+0.99\%$
test_compile_assign_and_add[pytree-eager]	2.9829ms	2.8232ms	354.2037 Ops/s	359.0176 Ops/s	$\color{#d91a1a}-1.34\%$
test_compile_indexing[tensor-tensordict-compile]	0.1726ms	0.1078ms	9.2728 KOps/s	8.9912 KOps/s	$\color{#35bf28}+3.13\%$
test_compile_indexing[tensor-tensordict-eager]	0.3160ms	73.0188μs	13.6951 KOps/s	13.6530 KOps/s	$\color{#35bf28}+0.31\%$
test_compile_indexing[tensor-tensorclass-compile]	0.2220ms	95.6873μs	10.4507 KOps/s	10.2449 KOps/s	$\color{#35bf28}+2.01\%$
test_compile_indexing[tensor-tensorclass-eager]	0.2570ms	45.4354μs	22.0093 KOps/s	22.3016 KOps/s	$\color{#d91a1a}-1.31\%$
test_compile_indexing[tensor-pytree-compile]	0.1415ms	96.7035μs	10.3409 KOps/s	10.1937 KOps/s	$\color{#35bf28}+1.44\%$
test_compile_indexing[tensor-pytree-eager]	0.2560ms	45.0586μs	22.1933 KOps/s	22.2557 KOps/s	$\color{#d91a1a}-0.28\%$
test_compile_indexing[slice-tensordict-compile]	94.5450μs	55.3631μs	18.0626 KOps/s	17.5316 KOps/s	$\color{#35bf28}+3.03\%$
test_compile_indexing[slice-tensordict-eager]	0.2266ms	28.3421μs	35.2832 KOps/s	35.7393 KOps/s	$\color{#d91a1a}-1.28\%$
test_compile_indexing[slice-tensorclass-compile]	83.9650μs	45.2949μs	22.0775 KOps/s	21.9064 KOps/s	$\color{#35bf28}+0.78\%$
test_compile_indexing[slice-tensorclass-eager]	0.2646ms	22.8489μs	43.7657 KOps/s	43.2980 KOps/s	$\color{#35bf28}+1.08\%$
test_compile_indexing[slice-pytree-compile]	78.8240μs	45.9017μs	21.7857 KOps/s	21.7501 KOps/s	$\color{#35bf28}+0.16\%$
test_compile_indexing[slice-pytree-eager]	0.2761ms	22.8375μs	43.7875 KOps/s	43.1896 KOps/s	$\color{#35bf28}+1.38\%$
test_compile_indexing[int-tensordict-compile]	96.0650μs	57.2522μs	17.4666 KOps/s	17.4749 KOps/s	$\color{#d91a1a}-0.05\%$
test_compile_indexing[int-tensordict-eager]	0.2817ms	28.4499μs	35.1495 KOps/s	35.4883 KOps/s	$\color{#d91a1a}-0.95\%$
test_compile_indexing[int-tensorclass-compile]	85.5640μs	45.8102μs	21.8292 KOps/s	21.5160 KOps/s	$\color{#35bf28}+1.46\%$
test_compile_indexing[int-tensorclass-eager]	0.2602ms	22.8504μs	43.7629 KOps/s	43.2590 KOps/s	$\color{#35bf28}+1.16\%$
test_compile_indexing[int-pytree-compile]	95.4650μs	45.6022μs	21.9288 KOps/s	21.6005 KOps/s	$\color{#35bf28}+1.52\%$
test_compile_indexing[int-pytree-eager]	0.2752ms	23.0035μs	43.4716 KOps/s	43.2933 KOps/s	$\color{#35bf28}+0.41\%$
test_mod_add[eager]	97.8850μs	51.5056μs	19.4154 KOps/s	19.3960 KOps/s	$\color{#35bf28}+0.10\%$
test_mod_add[compile]	0.1946ms	0.1047ms	9.5536 KOps/s	9.3685 KOps/s	$\color{#35bf28}+1.98\%$
test_mod_add[compile-overhead]	0.2314ms	0.1461ms	6.8439 KOps/s	6.5885 KOps/s	$\color{#35bf28}+3.88\%$
test_mod_wrap[eager]	0.3727ms	0.2930ms	3.4126 KOps/s	3.3929 KOps/s	$\color{#35bf28}+0.58\%$
test_mod_wrap[compile]	0.4313ms	0.3552ms	2.8152 KOps/s	2.8396 KOps/s	$\color{#d91a1a}-0.86\%$
test_mod_wrap[compile-overhead]	7.3632ms	4.0581ms	246.4193 Ops/s	247.5345 Ops/s	$\color{#d91a1a}-0.45\%$
test_mod_wrap_and_backward[eager]	1.6182ms	1.4965ms	668.2378 Ops/s	649.7632 Ops/s	$\color{#35bf28}+2.84\%$
test_mod_wrap_and_backward[compile]	1.6172ms	1.4531ms	688.1667 Ops/s	686.9249 Ops/s	$\color{#35bf28}+0.18\%$
test_mod_wrap_and_backward[compile-overhead]	1.7191ms	0.8828ms	1.1328 KOps/s	1.1084 KOps/s	$\color{#35bf28}+2.20\%$
test_seq_add[eager]	0.2233ms	0.1638ms	6.1055 KOps/s	5.9676 KOps/s	$\color{#35bf28}+2.31\%$
test_seq_add[compile]	0.1760ms	0.1204ms	8.3065 KOps/s	8.3047 KOps/s	$\color{#35bf28}+0.02\%$
test_seq_add[compile-overhead]	0.2127ms	0.1590ms	6.2902 KOps/s	6.2783 KOps/s	$\color{#35bf28}+0.19\%$
test_seq_wrap[eager]	0.6301ms	0.5572ms	1.7948 KOps/s	1.8262 KOps/s	$\color{#d91a1a}-1.72\%$
test_seq_wrap[compile]	0.4701ms	0.3811ms	2.6243 KOps/s	2.6609 KOps/s	$\color{#d91a1a}-1.38\%$
test_seq_wrap[compile-overhead]	0.3362ms	0.2729ms	3.6637 KOps/s	3.7405 KOps/s	$\color{#d91a1a}-2.05\%$
test_func_call_runtime[False-eager]	0.9755ms	0.8786ms	1.1382 KOps/s	1.1807 KOps/s	$\color{#d91a1a}-3.59\%$
test_func_call_runtime[False-compile]	1.0436ms	0.9406ms	1.0632 KOps/s	1.0879 KOps/s	$\color{#d91a1a}-2.27\%$
test_func_call_runtime[False-compile-overhead]	0.5135ms	0.4560ms	2.1929 KOps/s	2.1493 KOps/s	$\color{#35bf28}+2.03\%$
test_func_call_runtime[True-eager]	1.2320ms	1.0991ms	909.8338 Ops/s	914.0972 Ops/s	$\color{#d91a1a}-0.47\%$
test_func_call_runtime[True-compile]	1.0388ms	0.9481ms	1.0548 KOps/s	1.0707 KOps/s	$\color{#d91a1a}-1.49\%$
test_func_call_runtime[True-compile-overhead]	0.5193ms	0.4709ms	2.1235 KOps/s	2.0761 KOps/s	$\color{#35bf28}+2.28\%$
test_func_call_cm_runtime[False-eager]	0.9095ms	0.8385ms	1.1926 KOps/s	1.1422 KOps/s	$\color{#35bf28}+4.41\%$
test_func_call_cm_runtime[False-compile]	1.0750ms	0.9161ms	1.0916 KOps/s	1.0736 KOps/s	$\color{#35bf28}+1.68\%$
test_func_call_cm_runtime[False-compile-overhead]	0.5341ms	0.4601ms	2.1735 KOps/s	2.1339 KOps/s	$\color{#35bf28}+1.86\%$
test_func_call_cm_runtime[True-eager]	1.3197ms	1.2224ms	818.0837 Ops/s	797.9728 Ops/s	$\color{#35bf28}+2.52\%$
test_func_call_cm_runtime[True-compile]	1.0197ms	0.9559ms	1.0462 KOps/s	1.0293 KOps/s	$\color{#35bf28}+1.64\%$
test_func_call_cm_runtime[True-compile-overhead]	0.5582ms	0.5035ms	1.9863 KOps/s	1.9518 KOps/s	$\color{#35bf28}+1.76\%$
test_vmap_func_call_cm_runtime[eager]	2.8425ms	2.3473ms	426.0246 Ops/s	417.1863 Ops/s	$\color{#35bf28}+2.12\%$
test_vmap_func_call_cm_runtime[compile]	1.0810ms	0.9743ms	1.0263 KOps/s	1.0211 KOps/s	$\color{#35bf28}+0.51\%$
test_vmap_func_call_cm_runtime[compile-overhead]	0.5529ms	0.5064ms	1.9746 KOps/s	1.9358 KOps/s	$\color{#35bf28}+2.01\%$
test_distributed	0.6858ms	0.1545ms	6.4725 KOps/s	6.4883 KOps/s	$\color{#d91a1a}-0.24\%$
test_tdmodule	0.2680ms	29.6059μs	33.7770 KOps/s	34.9812 KOps/s	$\color{#d91a1a}-3.44\%$
test_tdmodule_dispatch	77.3740μs	46.4540μs	21.5267 KOps/s	21.3803 KOps/s	$\color{#35bf28}+0.68\%$
test_tdseq	48.1320μs	27.9357μs	35.7964 KOps/s	35.3917 KOps/s	$\color{#35bf28}+1.14\%$
test_tdseq_dispatch	69.8130μs	48.8922μs	20.4531 KOps/s	20.1647 KOps/s	$\color{#35bf28}+1.43\%$
test_instantiation_functorch	2.1901ms	2.0566ms	486.2512 Ops/s	484.8135 Ops/s	$\color{#35bf28}+0.30\%$
test_exec_functorch	0.2329ms	0.1790ms	5.5856 KOps/s	5.4687 KOps/s	$\color{#35bf28}+2.14\%$
test_exec_functional_call	0.2111ms	0.1679ms	5.9562 KOps/s	6.1385 KOps/s	$\color{#d91a1a}-2.97\%$
test_exec_td_decorator	0.4642ms	0.2367ms	4.2256 KOps/s	4.2065 KOps/s	$\color{#35bf28}+0.46\%$
test_vmap_mlp_speed_decorator[True-True]	1.0197ms	0.8194ms	1.2204 KOps/s	1.2090 KOps/s	$\color{#35bf28}+0.94\%$
test_vmap_mlp_speed_decorator[True-False]	1.0082ms	0.8430ms	1.1863 KOps/s	1.2110 KOps/s	$\color{#d91a1a}-2.04\%$
test_vmap_mlp_speed_decorator[False-True]	0.9321ms	0.7160ms	1.3966 KOps/s	1.4012 KOps/s	$\color{#d91a1a}-0.33\%$
test_vmap_mlp_speed_decorator[False-False]	0.8714ms	0.7094ms	1.4097 KOps/s	1.4014 KOps/s	$\color{#35bf28}+0.59\%$
test_vmap_transformer_speed_decorator[True-True]	21.1009ms	20.5919ms	48.5627 Ops/s	48.5387 Ops/s	$\color{#35bf28}+0.05\%$
test_vmap_transformer_speed_decorator[True-False]	20.7170ms	20.5539ms	48.6526 Ops/s	48.4994 Ops/s	$\color{#35bf28}+0.32\%$
test_vmap_transformer_speed_decorator[False-True]	20.9515ms	20.3823ms	49.0621 Ops/s	48.9317 Ops/s	$\color{#35bf28}+0.27\%$
test_vmap_transformer_speed_decorator[False-False]	21.0341ms	20.3757ms	49.0782 Ops/s	48.9304 Ops/s	$\color{#35bf28}+0.30\%$
test_to_module_speed[True]	2.0412ms	1.4736ms	678.6007 Ops/s	667.7855 Ops/s	$\color{#35bf28}+1.62\%$
test_to_module_speed[False]	1.9711ms	1.4450ms	692.0253 Ops/s	682.1545 Ops/s	$\color{#35bf28}+1.45\%$
test_tc_init	82.2440μs	46.0791μs	21.7018 KOps/s	21.2611 KOps/s	$\color{#35bf28}+2.07\%$
test_tc_init_tensor_only	35.8720μs	9.9138μs	100.8695 KOps/s	101.0767 KOps/s	$\color{#d91a1a}-0.21\%$
test_tc_init_nested	0.1703ms	92.7423μs	10.7826 KOps/s	10.8651 KOps/s	$\color{#d91a1a}-0.76\%$
test_tc_init_many_fields	44.8530μs	16.5403μs	60.4584 KOps/s	60.6943 KOps/s	$\color{#d91a1a}-0.39\%$
test_tc_first_layer_tensor	15.2810μs	1.8247μs	548.0313 KOps/s	540.2200 KOps/s	$\color{#35bf28}+1.45\%$
test_tc_first_layer_tensor_only	4.0173μs	0.7525μs	1.3289 MOps/s	1.3396 MOps/s	$\color{#d91a1a}-0.79\%$
test_tc_first_layer_tensor_set	24.6720μs	4.2070μs	237.6987 KOps/s	250.1938 KOps/s	$\color{#d91a1a}-4.99\%$
test_tc_first_layer_tensor_only_set	18.4710μs	3.1647μs	315.9831 KOps/s	319.6985 KOps/s	$\color{#d91a1a}-1.16\%$
test_tc_first_layer_nontensor	28.0720μs	6.1860μs	161.6558 KOps/s	160.7817 KOps/s	$\color{#35bf28}+0.54\%$
test_tc_second_layer_tensor	23.9210μs	4.4168μs	226.4081 KOps/s	228.4642 KOps/s	$\color{#d91a1a}-0.90\%$
test_tc_second_layer_nontensor	29.2310μs	8.7549μs	114.2213 KOps/s	115.0136 KOps/s	$\color{#d91a1a}-0.69\%$
test_unbind	0.2673s	16.1285ms	62.0020 Ops/s	56.9528 Ops/s	$\textbf{\color{#35bf28}+8.87\%}$
test_full_like	10.4718ms	4.3975ms	227.4010 Ops/s	227.7124 Ops/s	$\color{#d91a1a}-0.14\%$
test_zeros_like	4.7837ms	4.3488ms	229.9503 Ops/s	228.6936 Ops/s	$\color{#35bf28}+0.55\%$
test_ones_like	4.5044ms	4.3561ms	229.5634 Ops/s	228.6935 Ops/s	$\color{#35bf28}+0.38\%$
test_clone	6.8953ms	6.4059ms	156.1068 Ops/s	155.8422 Ops/s	$\color{#35bf28}+0.17\%$
test_squeeze	0.1873ms	14.4104μs	69.3944 KOps/s	68.2709 KOps/s	$\color{#35bf28}+1.65\%$
test_unsqueeze	0.1596ms	0.1113ms	8.9859 KOps/s	8.7081 KOps/s	$\color{#35bf28}+3.19\%$
test_split	0.2834ms	0.1882ms	5.3132 KOps/s	5.2430 KOps/s	$\color{#35bf28}+1.34\%$
test_permute	0.2590ms	0.2127ms	4.7019 KOps/s	4.6756 KOps/s	$\color{#35bf28}+0.56\%$
test_stack	51.2381ms	50.6467ms	19.7446 Ops/s	19.5978 Ops/s	$\color{#35bf28}+0.75\%$
test_cat	51.2586ms	51.0277ms	19.5972 Ops/s	23.3962 Ops/s	$\textbf{\color{#d91a1a}-16.24\%}$