[Performance] Faster params and buffer registration in TensorDictParams

[vmoens]

No description provided.

[github-actions]

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

Total Benchmarks: 113. Improved: $\large\color{#35bf28}1$. Worsened: $\large\color{#d91a1a}16$.

Expand to view detailed results

Name	Max	Mean	Ops	Ops on Repo HEAD	Change
test_plain_set_nested	50.1130μs	15.8994μs	62.8955 KOps/s	64.7885 KOps/s	$\color{#d91a1a}-2.92\%$
test_plain_set_stack_nested	0.2062ms	0.1470ms	6.8048 KOps/s	7.0605 KOps/s	$\color{#d91a1a}-3.62\%$
test_plain_set_nested_inplace	42.8200μs	18.9600μs	52.7427 KOps/s	54.3493 KOps/s	$\color{#d91a1a}-2.96\%$
test_plain_set_stack_nested_inplace	0.2586ms	0.1758ms	5.6886 KOps/s	5.9193 KOps/s	$\color{#d91a1a}-3.90\%$
test_items	23.8350μs	2.3424μs	426.9195 KOps/s	414.5356 KOps/s	$\color{#35bf28}+2.99\%$
test_items_nested	0.4182ms	0.2671ms	3.7432 KOps/s	3.8469 KOps/s	$\color{#d91a1a}-2.69\%$
test_items_nested_locked	0.9954ms	0.2697ms	3.7074 KOps/s	3.8182 KOps/s	$\color{#d91a1a}-2.90\%$
test_items_nested_leaf	0.5435ms	0.1646ms	6.0735 KOps/s	6.2890 KOps/s	$\color{#d91a1a}-3.43\%$
test_items_stack_nested	1.5859ms	1.4785ms	676.3412 Ops/s	702.4100 Ops/s	$\color{#d91a1a}-3.71\%$
test_items_stack_nested_leaf	1.4354ms	1.3469ms	742.4647 Ops/s	769.4412 Ops/s	$\color{#d91a1a}-3.51\%$
test_items_stack_nested_locked	0.8632ms	0.7644ms	1.3081 KOps/s	1.3412 KOps/s	$\color{#d91a1a}-2.46\%$
test_keys	26.4890μs	3.9648μs	252.2187 KOps/s	267.7743 KOps/s	$\textbf{\color{#d91a1a}-5.81\%}$
test_keys_nested	1.3604ms	0.1416ms	7.0619 KOps/s	7.0025 KOps/s	$\color{#35bf28}+0.85\%$
test_keys_nested_locked	0.3255ms	0.1400ms	7.1431 KOps/s	7.2131 KOps/s	$\color{#d91a1a}-0.97\%$
test_keys_nested_leaf	0.2969ms	0.1406ms	7.1108 KOps/s	7.0771 KOps/s	$\color{#35bf28}+0.48\%$
test_keys_stack_nested	1.8425ms	1.4086ms	709.9382 Ops/s	727.3636 Ops/s	$\color{#d91a1a}-2.40\%$
test_keys_stack_nested_leaf	1.6858ms	1.4035ms	712.4884 Ops/s	740.0463 Ops/s	$\color{#d91a1a}-3.72\%$
test_keys_stack_nested_locked	0.7542ms	0.6711ms	1.4901 KOps/s	1.5573 KOps/s	$\color{#d91a1a}-4.32\%$
test_values	6.8605μs	1.1720μs	853.2122 KOps/s	862.3639 KOps/s	$\color{#d91a1a}-1.06\%$
test_values_nested	0.1165ms	49.6535μs	20.1396 KOps/s	20.2775 KOps/s	$\color{#d91a1a}-0.68\%$
test_values_nested_locked	78.5360μs	49.5049μs	20.2000 KOps/s	20.1332 KOps/s	$\color{#35bf28}+0.33\%$
test_values_nested_leaf	62.2960μs	44.5641μs	22.4396 KOps/s	22.5771 KOps/s	$\color{#d91a1a}-0.61\%$
test_values_stack_nested	1.5657ms	1.2142ms	823.5757 Ops/s	863.8438 Ops/s	$\color{#d91a1a}-4.66\%$
test_values_stack_nested_leaf	1.2936ms	1.2000ms	833.3159 Ops/s	884.7348 Ops/s	$\textbf{\color{#d91a1a}-5.81\%}$
test_values_stack_nested_locked	0.8244ms	0.5128ms	1.9502 KOps/s	1.9844 KOps/s	$\color{#d91a1a}-1.72\%$
test_membership	11.9530μs	1.3576μs	736.6166 KOps/s	748.7768 KOps/s	$\color{#d91a1a}-1.62\%$
test_membership_nested	26.5600μs	2.7992μs	357.2402 KOps/s	358.6144 KOps/s	$\color{#d91a1a}-0.38\%$
test_membership_nested_leaf	28.1520μs	2.8096μs	355.9271 KOps/s	374.9869 KOps/s	$\textbf{\color{#d91a1a}-5.08\%}$
test_membership_stacked_nested	51.4760μs	11.8029μs	84.7249 KOps/s	87.8429 KOps/s	$\color{#d91a1a}-3.55\%$
test_membership_stacked_nested_leaf	34.9150μs	11.7732μs	84.9390 KOps/s	88.7420 KOps/s	$\color{#d91a1a}-4.29\%$
test_membership_nested_last	33.5720μs	5.8564μs	170.7547 KOps/s	176.3483 KOps/s	$\color{#d91a1a}-3.17\%$
test_membership_nested_leaf_last	19.4160μs	5.9428μs	168.2698 KOps/s	176.4126 KOps/s	$\color{#d91a1a}-4.62\%$
test_membership_stacked_nested_last	0.3385ms	0.1698ms	5.8878 KOps/s	6.2331 KOps/s	$\textbf{\color{#d91a1a}-5.54\%}$
test_membership_stacked_nested_leaf_last	50.5840μs	13.9243μs	71.8169 KOps/s	73.7655 KOps/s	$\color{#d91a1a}-2.64\%$
test_nested_getleaf	33.2220μs	10.8501μs	92.1648 KOps/s	96.1868 KOps/s	$\color{#d91a1a}-4.18\%$
test_nested_get	37.2290μs	10.2687μs	97.3835 KOps/s	99.8696 KOps/s	$\color{#d91a1a}-2.49\%$
test_stacked_getleaf	1.1769ms	0.6493ms	1.5401 KOps/s	1.6313 KOps/s	$\textbf{\color{#d91a1a}-5.59\%}$
test_stacked_get	1.2903ms	0.6117ms	1.6347 KOps/s	1.7031 KOps/s	$\color{#d91a1a}-4.02\%$
test_nested_getitemleaf	38.3010μs	10.7315μs	93.1834 KOps/s	94.7444 KOps/s	$\color{#d91a1a}-1.65\%$
test_nested_getitem	40.2350μs	10.2817μs	97.2603 KOps/s	102.3342 KOps/s	$\color{#d91a1a}-4.96\%$
test_stacked_getitemleaf	1.2384ms	0.6345ms	1.5760 KOps/s	1.6345 KOps/s	$\color{#d91a1a}-3.58\%$
test_stacked_getitem	0.8348ms	0.6008ms	1.6643 KOps/s	1.6937 KOps/s	$\color{#d91a1a}-1.73\%$
test_lock_nested	53.5346ms	0.5370ms	1.8623 KOps/s	2.0887 KOps/s	$\textbf{\color{#d91a1a}-10.84\%}$
test_lock_stack_nested	69.9764ms	7.8091ms	128.0562 Ops/s	134.8617 Ops/s	$\textbf{\color{#d91a1a}-5.05\%}$
test_unlock_nested	58.2626ms	0.5094ms	1.9631 KOps/s	2.0236 KOps/s	$\color{#d91a1a}-2.99\%$
test_unlock_stack_nested	62.8419ms	7.5608ms	132.2617 Ops/s	206.5750 Ops/s	$\textbf{\color{#d91a1a}-35.97\%}$
test_flatten_speed	0.5367ms	0.2668ms	3.7487 KOps/s	3.8225 KOps/s	$\color{#d91a1a}-1.93\%$
test_unflatten_speed	0.8683ms	0.4690ms	2.1324 KOps/s	2.2428 KOps/s	$\color{#d91a1a}-4.92\%$
test_common_ops	2.8302ms	0.6848ms	1.4603 KOps/s	1.4982 KOps/s	$\color{#d91a1a}-2.53\%$
test_creation	49.8230μs	2.4307μs	411.4026 KOps/s	421.1741 KOps/s	$\color{#d91a1a}-2.32\%$
test_creation_empty	39.8950μs	8.1685μs	122.4214 KOps/s	122.8991 KOps/s	$\color{#d91a1a}-0.39\%$
test_creation_nested_1	45.7950μs	11.5286μs	86.7406 KOps/s	89.4600 KOps/s	$\color{#d91a1a}-3.04\%$
test_creation_nested_2	34.4640μs	15.1832μs	65.8624 KOps/s	67.6852 KOps/s	$\color{#d91a1a}-2.69\%$
test_clone	90.0780μs	13.5365μs	73.8743 KOps/s	76.4908 KOps/s	$\color{#d91a1a}-3.42\%$
test_getitem[int]	55.9040μs	13.1572μs	76.0042 KOps/s	76.8686 KOps/s	$\color{#d91a1a}-1.12\%$
test_getitem[slice_int]	64.0690μs	25.1341μs	39.7866 KOps/s	39.4468 KOps/s	$\color{#35bf28}+0.86\%$
test_getitem[range]	0.2022ms	42.2052μs	23.6938 KOps/s	22.3820 KOps/s	$\textbf{\color{#35bf28}+5.86\%}$
test_getitem[tuple]	63.8790μs	20.8067μs	48.0615 KOps/s	48.5792 KOps/s	$\color{#d91a1a}-1.07\%$
test_getitem[list]	0.2558ms	38.3661μs	26.0647 KOps/s	25.5200 KOps/s	$\color{#35bf28}+2.13\%$
test_setitem_dim[int]	53.6600μs	28.1646μs	35.5056 KOps/s	36.9804 KOps/s	$\color{#d91a1a}-3.99\%$
test_setitem_dim[slice_int]	0.1204ms	53.8913μs	18.5559 KOps/s	20.3673 KOps/s	$\textbf{\color{#d91a1a}-8.89\%}$
test_setitem_dim[range]	0.1206ms	71.3533μs	14.0148 KOps/s	13.7529 KOps/s	$\color{#35bf28}+1.90\%$
test_setitem_dim[tuple]	68.9880μs	41.0921μs	24.3356 KOps/s	25.1132 KOps/s	$\color{#d91a1a}-3.10\%$
test_setitem	82.5240μs	18.5597μs	53.8803 KOps/s	55.2216 KOps/s	$\color{#d91a1a}-2.43\%$
test_set	78.8670μs	18.0650μs	55.3555 KOps/s	58.1057 KOps/s	$\color{#d91a1a}-4.73\%$
test_set_shared	1.8502ms	0.1379ms	7.2528 KOps/s	7.3408 KOps/s	$\color{#d91a1a}-1.20\%$
test_update	95.6380μs	23.5938μs	42.3841 KOps/s	43.3697 KOps/s	$\color{#d91a1a}-2.27\%$
test_update_nested	0.1012ms	34.8542μs	28.6909 KOps/s	30.0434 KOps/s	$\color{#d91a1a}-4.50\%$
test_set_nested	79.7190μs	19.8445μs	50.3917 KOps/s	52.6650 KOps/s	$\color{#d91a1a}-4.32\%$
test_set_nested_new	97.7320μs	26.3600μs	37.9363 KOps/s	41.2183 KOps/s	$\textbf{\color{#d91a1a}-7.96\%}$
test_select	0.1084ms	51.3400μs	19.4780 KOps/s	20.7869 KOps/s	$\textbf{\color{#d91a1a}-6.30\%}$
test_unbind_speed	0.6965ms	0.3794ms	2.6354 KOps/s	2.6837 KOps/s	$\color{#d91a1a}-1.80\%$
test_unbind_speed_stack0	66.7466ms	5.3629ms	186.4669 Ops/s	250.2493 Ops/s	$\textbf{\color{#d91a1a}-25.49\%}$
test_unbind_speed_stack1	2.4145μs	0.6338μs	1.5779 MOps/s	1.6030 MOps/s	$\color{#d91a1a}-1.57\%$
test_split	55.9680ms	1.7799ms	561.8276 Ops/s	556.5728 Ops/s	$\color{#35bf28}+0.94\%$
test_chunk	50.4416ms	1.7431ms	573.6765 Ops/s	568.6560 Ops/s	$\color{#35bf28}+0.88\%$
test_creation[device0]	0.3886ms	0.2946ms	3.3943 KOps/s	3.4662 KOps/s	$\color{#d91a1a}-2.07\%$
test_creation_from_tensor	3.4794ms	0.3287ms	3.0425 KOps/s	3.0520 KOps/s	$\color{#d91a1a}-0.31\%$
test_add_one[memmap_tensor0]	73.8880μs	25.3452μs	39.4552 KOps/s	40.2115 KOps/s	$\color{#d91a1a}-1.88\%$
test_contiguous[memmap_tensor0]	46.4170μs	5.6985μs	175.4851 KOps/s	174.1955 KOps/s	$\color{#35bf28}+0.74\%$
test_stack[memmap_tensor0]	60.2220μs	19.5464μs	51.1604 KOps/s	51.7114 KOps/s	$\color{#d91a1a}-1.07\%$
test_memmaptd_index	0.2595ms	0.1924ms	5.1974 KOps/s	5.1989 KOps/s	$\color{#d91a1a}-0.03\%$
test_memmaptd_index_astensor	0.4076ms	0.2480ms	4.0329 KOps/s	4.0138 KOps/s	$\color{#35bf28}+0.48\%$
test_memmaptd_index_op	0.6558ms	0.4885ms	2.0469 KOps/s	2.0261 KOps/s	$\color{#35bf28}+1.03\%$
test_reshape_pytree	67.5150μs	23.5601μs	42.4446 KOps/s	44.8024 KOps/s	$\textbf{\color{#d91a1a}-5.26\%}$
test_reshape_td	66.6840μs	31.2562μs	31.9937 KOps/s	32.8544 KOps/s	$\color{#d91a1a}-2.62\%$
test_view_pytree	55.5730μs	23.0130μs	43.4536 KOps/s	44.7210 KOps/s	$\color{#d91a1a}-2.83\%$
test_view_td	29.5550μs	4.8362μs	206.7744 KOps/s	209.4473 KOps/s	$\color{#d91a1a}-1.28\%$
test_unbind_pytree	0.6363ms	26.1998μs	38.1682 KOps/s	38.6905 KOps/s	$\color{#d91a1a}-1.35\%$
test_unbind_td	0.1108ms	60.0244μs	16.6599 KOps/s	16.8776 KOps/s	$\color{#d91a1a}-1.29\%$
test_split_pytree	57.9880μs	26.3791μs	37.9088 KOps/s	39.2936 KOps/s	$\color{#d91a1a}-3.52\%$
test_split_td	90.0470μs	46.0257μs	21.7270 KOps/s	21.5111 KOps/s	$\color{#35bf28}+1.00\%$
test_add_pytree	79.6790μs	32.3212μs	30.9395 KOps/s	31.2846 KOps/s	$\color{#d91a1a}-1.10\%$
test_add_td	95.5980μs	44.3351μs	22.5555 KOps/s	22.7875 KOps/s	$\color{#d91a1a}-1.02\%$
test_distributed	20.1980μs	5.9936μs	166.8446 KOps/s	172.0924 KOps/s	$\color{#d91a1a}-3.05\%$
test_tdmodule	1.5871ms	22.4976μs	44.4493 KOps/s	47.6618 KOps/s	$\textbf{\color{#d91a1a}-6.74\%}$
test_tdmodule_dispatch	0.2051ms	38.6877μs	25.8480 KOps/s	26.1992 KOps/s	$\color{#d91a1a}-1.34\%$
test_tdseq	42.5890μs	23.4047μs	42.7265 KOps/s	42.3866 KOps/s	$\color{#35bf28}+0.80\%$
test_tdseq_dispatch	0.1366ms	42.0017μs	23.8085 KOps/s	23.8229 KOps/s	$\color{#d91a1a}-0.06\%$
test_instantiation_functorch	1.4487ms	1.3375ms	747.6533 Ops/s	791.8239 Ops/s	$\textbf{\color{#d91a1a}-5.58\%}$
test_instantiation_td	69.5943ms	1.1217ms	891.5276 Ops/s	950.4616 Ops/s	$\textbf{\color{#d91a1a}-6.20\%}$
test_exec_functorch	0.2310ms	0.1610ms	6.2098 KOps/s	6.3443 KOps/s	$\color{#d91a1a}-2.12\%$
test_exec_functional_call	0.2278ms	0.1501ms	6.6630 KOps/s	6.7826 KOps/s	$\color{#d91a1a}-1.76\%$
test_exec_td	0.2248ms	0.1465ms	6.8267 KOps/s	6.8537 KOps/s	$\color{#d91a1a}-0.39\%$
test_exec_td_decorator	0.6893ms	0.2234ms	4.4767 KOps/s	4.6542 KOps/s	$\color{#d91a1a}-3.82\%$
test_vmap_mlp_speed[True-True]	0.9700ms	0.8885ms	1.1254 KOps/s	1.1369 KOps/s	$\color{#d91a1a}-1.01\%$
test_vmap_mlp_speed[True-False]	0.7681ms	0.4653ms	2.1494 KOps/s	2.1761 KOps/s	$\color{#d91a1a}-1.23\%$
test_vmap_mlp_speed[False-True]	1.5787ms	0.7809ms	1.2806 KOps/s	1.3111 KOps/s	$\color{#d91a1a}-2.32\%$
test_vmap_mlp_speed[False-False]	0.7039ms	0.3826ms	2.6137 KOps/s	2.6167 KOps/s	$\color{#d91a1a}-0.11\%$
test_vmap_mlp_speed_decorator[True-True]	2.2493ms	1.5739ms	635.3515 Ops/s	643.6554 Ops/s	$\color{#d91a1a}-1.29\%$
test_vmap_mlp_speed_decorator[True-False]	1.0332ms	0.5455ms	1.8333 KOps/s	1.8423 KOps/s	$\color{#d91a1a}-0.49\%$
test_vmap_mlp_speed_decorator[False-True]	1.9228ms	1.3604ms	735.1042 Ops/s	749.8734 Ops/s	$\color{#d91a1a}-1.97\%$
test_vmap_mlp_speed_decorator[False-False]	0.8358ms	0.4247ms	2.3548 KOps/s	2.3844 KOps/s	$\color{#d91a1a}-1.24\%$

[github-actions]

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

Total Benchmarks: 127. Improved: $\large\color{#35bf28}1$. Worsened: $\large\color{#d91a1a}13$.

Expand to view detailed results

Name	Max	Mean	Ops	Ops on Repo HEAD	Change
test_plain_set_nested	0.4567ms	12.8286μs	77.9507 KOps/s	78.3521 KOps/s	$\color{#d91a1a}-0.51\%$
test_plain_set_stack_nested	0.1961ms	0.1159ms	8.6247 KOps/s	8.3774 KOps/s	$\color{#35bf28}+2.95\%$
test_plain_set_nested_inplace	31.6400μs	15.6392μs	63.9417 KOps/s	65.8399 KOps/s	$\color{#d91a1a}-2.88\%$
test_plain_set_stack_nested_inplace	0.1759ms	0.1447ms	6.9127 KOps/s	7.0749 KOps/s	$\color{#d91a1a}-2.29\%$
test_items	23.6400μs	4.7129μs	212.1835 KOps/s	210.7644 KOps/s	$\color{#35bf28}+0.67\%$
test_items_nested	0.3823ms	0.3384ms	2.9549 KOps/s	2.9603 KOps/s	$\color{#d91a1a}-0.18\%$
test_items_nested_locked	0.3599ms	0.3373ms	2.9647 KOps/s	2.9468 KOps/s	$\color{#35bf28}+0.61\%$
test_items_nested_leaf	0.2235ms	0.1978ms	5.0549 KOps/s	5.0043 KOps/s	$\color{#35bf28}+1.01\%$
test_items_stack_nested	1.5352ms	1.4793ms	675.9928 Ops/s	677.6973 Ops/s	$\color{#d91a1a}-0.25\%$
test_items_stack_nested_leaf	1.3825ms	1.3186ms	758.3694 Ops/s	764.7870 Ops/s	$\color{#d91a1a}-0.84\%$
test_items_stack_nested_locked	0.8913ms	0.8229ms	1.2153 KOps/s	1.2391 KOps/s	$\color{#d91a1a}-1.93\%$
test_keys	22.7600μs	4.6217μs	216.3690 KOps/s	212.8548 KOps/s	$\color{#35bf28}+1.65\%$
test_keys_nested	0.5788ms	90.9418μs	10.9960 KOps/s	11.0957 KOps/s	$\color{#d91a1a}-0.90\%$
test_keys_nested_locked	0.1112ms	89.8591μs	11.1285 KOps/s	11.1918 KOps/s	$\color{#d91a1a}-0.57\%$
test_keys_nested_leaf	42.6074ms	86.6209μs	11.5446 KOps/s	12.1860 KOps/s	$\textbf{\color{#d91a1a}-5.26\%}$
test_keys_stack_nested	1.3389ms	1.3016ms	768.2778 Ops/s	763.8173 Ops/s	$\color{#35bf28}+0.58\%$
test_keys_stack_nested_leaf	1.3866ms	1.2982ms	770.2936 Ops/s	767.4561 Ops/s	$\color{#35bf28}+0.37\%$
test_keys_stack_nested_locked	0.7084ms	0.6244ms	1.6015 KOps/s	1.6303 KOps/s	$\color{#d91a1a}-1.77\%$
test_values	10.3937μs	1.8829μs	531.0966 KOps/s	526.9657 KOps/s	$\color{#35bf28}+0.78\%$
test_values_nested	58.6310μs	43.2600μs	23.1160 KOps/s	23.1413 KOps/s	$\color{#d91a1a}-0.11\%$
test_values_nested_locked	73.4110μs	43.0648μs	23.2208 KOps/s	23.1102 KOps/s	$\color{#35bf28}+0.48\%$
test_values_nested_leaf	0.1101ms	37.5784μs	26.6110 KOps/s	26.7814 KOps/s	$\color{#d91a1a}-0.64\%$
test_values_stack_nested	1.1960ms	1.1407ms	876.6554 Ops/s	890.2080 Ops/s	$\color{#d91a1a}-1.52\%$
test_values_stack_nested_leaf	1.1655ms	1.1291ms	885.6979 Ops/s	899.1234 Ops/s	$\color{#d91a1a}-1.49\%$
test_values_stack_nested_locked	0.5819ms	0.4978ms	2.0087 KOps/s	2.0517 KOps/s	$\color{#d91a1a}-2.09\%$
test_membership	3.9200μs	0.9449μs	1.0583 MOps/s	1.0578 MOps/s	$\color{#35bf28}+0.05\%$
test_membership_nested	16.2510μs	2.2105μs	452.3838 KOps/s	445.6173 KOps/s	$\color{#35bf28}+1.52\%$
test_membership_nested_leaf	13.3050μs	2.1401μs	467.2646 KOps/s	465.3064 KOps/s	$\color{#35bf28}+0.42\%$
test_membership_stacked_nested	45.3010μs	10.8143μs	92.4699 KOps/s	90.9826 KOps/s	$\color{#35bf28}+1.63\%$
test_membership_stacked_nested_leaf	30.6410μs	10.8942μs	91.7919 KOps/s	90.6306 KOps/s	$\color{#35bf28}+1.28\%$
test_membership_nested_last	38.2710μs	4.6523μs	214.9485 KOps/s	215.5704 KOps/s	$\color{#d91a1a}-0.29\%$
test_membership_nested_leaf_last	27.0510μs	4.6591μs	214.6334 KOps/s	216.4922 KOps/s	$\color{#d91a1a}-0.86\%$
test_membership_stacked_nested_last	0.2047ms	0.1343ms	7.4478 KOps/s	7.5429 KOps/s	$\color{#d91a1a}-1.26\%$
test_membership_stacked_nested_leaf_last	81.8610μs	12.8309μs	77.9369 KOps/s	78.2151 KOps/s	$\color{#d91a1a}-0.36\%$
test_nested_getleaf	31.0910μs	8.3599μs	119.6193 KOps/s	119.3469 KOps/s	$\color{#35bf28}+0.23\%$
test_nested_get	22.3410μs	7.9436μs	125.8880 KOps/s	125.9552 KOps/s	$\color{#d91a1a}-0.05\%$
test_stacked_getleaf	0.6361ms	0.5739ms	1.7425 KOps/s	1.7231 KOps/s	$\color{#35bf28}+1.13\%$
test_stacked_get	0.6532ms	0.5352ms	1.8684 KOps/s	1.8468 KOps/s	$\color{#35bf28}+1.17\%$
test_nested_getitemleaf	27.8000μs	8.4531μs	118.2997 KOps/s	118.1335 KOps/s	$\color{#35bf28}+0.14\%$
test_nested_getitem	30.6100μs	8.0113μs	124.8231 KOps/s	125.1091 KOps/s	$\color{#d91a1a}-0.23\%$
test_stacked_getitemleaf	0.6319ms	0.5778ms	1.7306 KOps/s	1.7434 KOps/s	$\color{#d91a1a}-0.73\%$
test_stacked_getitem	0.6607ms	0.5362ms	1.8651 KOps/s	1.8465 KOps/s	$\color{#35bf28}+1.01\%$
test_lock_nested	4.4657ms	0.4632ms	2.1590 KOps/s	2.1483 KOps/s	$\color{#35bf28}+0.50\%$
test_lock_stack_nested	71.5859ms	6.6803ms	149.6929 Ops/s	149.2541 Ops/s	$\color{#35bf28}+0.29\%$
test_unlock_nested	1.3050ms	0.4381ms	2.2828 KOps/s	1.9912 KOps/s	$\textbf{\color{#35bf28}+14.64\%}$
test_unlock_stack_nested	67.7370ms	7.4377ms	134.4493 Ops/s	135.1676 Ops/s	$\color{#d91a1a}-0.53\%$
test_flatten_speed	0.5211ms	0.1877ms	5.3289 KOps/s	5.3870 KOps/s	$\color{#d91a1a}-1.08\%$
test_unflatten_speed	0.3925ms	0.3604ms	2.7745 KOps/s	2.7689 KOps/s	$\color{#35bf28}+0.20\%$
test_common_ops	1.0912ms	0.6359ms	1.5725 KOps/s	1.6002 KOps/s	$\color{#d91a1a}-1.73\%$
test_creation	37.1400μs	1.9575μs	510.8514 KOps/s	508.4584 KOps/s	$\color{#35bf28}+0.47\%$
test_creation_empty	33.8310μs	7.1542μs	139.7780 KOps/s	136.7844 KOps/s	$\color{#35bf28}+2.19\%$
test_creation_nested_1	32.5010μs	9.5509μs	104.7023 KOps/s	104.0046 KOps/s	$\color{#35bf28}+0.67\%$
test_creation_nested_2	71.2920μs	12.1751μs	82.1350 KOps/s	81.9819 KOps/s	$\color{#35bf28}+0.19\%$
test_clone	93.9120μs	14.8206μs	67.4735 KOps/s	71.4441 KOps/s	$\textbf{\color{#d91a1a}-5.56\%}$
test_getitem[int]	66.4520μs	12.2180μs	81.8466 KOps/s	81.7625 KOps/s	$\color{#35bf28}+0.10\%$
test_getitem[slice_int]	48.2710μs	23.8226μs	41.9770 KOps/s	42.4479 KOps/s	$\color{#d91a1a}-1.11\%$
test_getitem[range]	68.2810μs	41.5742μs	24.0534 KOps/s	25.4460 KOps/s	$\textbf{\color{#d91a1a}-5.47\%}$
test_getitem[tuple]	42.7710μs	20.4790μs	48.8305 KOps/s	48.5173 KOps/s	$\color{#35bf28}+0.65\%$
test_getitem[list]	0.2495ms	38.4304μs	26.0211 KOps/s	27.4512 KOps/s	$\textbf{\color{#d91a1a}-5.21\%}$
test_setitem_dim[int]	43.3110μs	27.3272μs	36.5935 KOps/s	38.6876 KOps/s	$\textbf{\color{#d91a1a}-5.41\%}$
test_setitem_dim[slice_int]	82.0320μs	47.7186μs	20.9562 KOps/s	21.5820 KOps/s	$\color{#d91a1a}-2.90\%$
test_setitem_dim[range]	90.2520μs	63.1909μs	15.8251 KOps/s	15.9215 KOps/s	$\color{#d91a1a}-0.61\%$
test_setitem_dim[tuple]	60.4110μs	39.9395μs	25.0379 KOps/s	25.7163 KOps/s	$\color{#d91a1a}-2.64\%$
test_setitem	0.1034ms	19.1817μs	52.1331 KOps/s	55.3744 KOps/s	$\textbf{\color{#d91a1a}-5.85\%}$
test_set	99.4120μs	18.5442μs	53.9252 KOps/s	57.3446 KOps/s	$\textbf{\color{#d91a1a}-5.96\%}$
test_set_shared	0.5668ms	0.1027ms	9.7327 KOps/s	9.9260 KOps/s	$\color{#d91a1a}-1.95\%$
test_update	0.1135ms	22.9937μs	43.4902 KOps/s	45.8345 KOps/s	$\textbf{\color{#d91a1a}-5.11\%}$
test_update_nested	0.1284ms	32.1817μs	31.0735 KOps/s	32.1069 KOps/s	$\color{#d91a1a}-3.22\%$
test_set_nested	98.6820μs	19.9733μs	50.0668 KOps/s	52.6531 KOps/s	$\color{#d91a1a}-4.91\%$
test_set_nested_new	0.1090ms	23.9336μs	41.7822 KOps/s	43.8990 KOps/s	$\color{#d91a1a}-4.82\%$
test_select	76.8920μs	46.3370μs	21.5810 KOps/s	21.2858 KOps/s	$\color{#35bf28}+1.39\%$
test_to	73.6220μs	52.3354μs	19.1075 KOps/s	18.8297 KOps/s	$\color{#35bf28}+1.48\%$
test_to_nonblocking	70.3810μs	34.8765μs	28.6726 KOps/s	28.4916 KOps/s	$\color{#35bf28}+0.64\%$
test_unbind_speed	0.3927ms	0.3565ms	2.8049 KOps/s	2.8222 KOps/s	$\color{#d91a1a}-0.61\%$
test_unbind_speed_stack0	63.3412ms	5.2621ms	190.0377 Ops/s	191.1943 Ops/s	$\color{#d91a1a}-0.60\%$
test_unbind_speed_stack1	1.2430μs	0.5244μs	1.9068 MOps/s	1.9284 MOps/s	$\color{#d91a1a}-1.12\%$
test_split	54.2278ms	1.8369ms	544.4051 Ops/s	558.3058 Ops/s	$\color{#d91a1a}-2.49\%$
test_chunk	54.0882ms	1.8191ms	549.7133 Ops/s	564.2835 Ops/s	$\color{#d91a1a}-2.58\%$
test_creation[device0]	0.4516ms	0.3097ms	3.2290 KOps/s	3.1815 KOps/s	$\color{#35bf28}+1.49\%$
test_creation[device1]	0.7904ms	0.3115ms	3.2103 KOps/s	3.1816 KOps/s	$\color{#35bf28}+0.90\%$
test_creation_from_tensor	57.2059ms	0.3640ms	2.7472 KOps/s	2.9490 KOps/s	$\textbf{\color{#d91a1a}-6.84\%}$
test_add_one[memmap_tensor0]	0.2546ms	24.5831μs	40.6783 KOps/s	41.9724 KOps/s	$\color{#d91a1a}-3.08\%$
test_add_one[memmap_tensor1]	0.1845ms	75.1746μs	13.3024 KOps/s	13.2829 KOps/s	$\color{#35bf28}+0.15\%$
test_contiguous[memmap_tensor0]	32.1500μs	6.0829μs	164.3948 KOps/s	169.5108 KOps/s	$\color{#d91a1a}-3.02\%$
test_contiguous[memmap_tensor1]	50.8800μs	22.5668μs	44.3129 KOps/s	45.1950 KOps/s	$\color{#d91a1a}-1.95\%$
test_stack[memmap_tensor0]	49.3610μs	21.5523μs	46.3988 KOps/s	50.8129 KOps/s	$\textbf{\color{#d91a1a}-8.69\%}$
test_stack[memmap_tensor1]	0.1623ms	75.4035μs	13.2620 KOps/s	13.4455 KOps/s	$\color{#d91a1a}-1.36\%$
test_memmaptd_index	0.2619ms	0.2262ms	4.4212 KOps/s	4.4292 KOps/s	$\color{#d91a1a}-0.18\%$
test_memmaptd_index_astensor	0.3879ms	0.2811ms	3.5579 KOps/s	3.5861 KOps/s	$\color{#d91a1a}-0.79\%$
test_memmaptd_index_op	0.6509ms	0.5683ms	1.7596 KOps/s	1.8521 KOps/s	$\color{#d91a1a}-4.99\%$
test_reshape_pytree	0.2609ms	21.4457μs	46.6293 KOps/s	47.8269 KOps/s	$\color{#d91a1a}-2.50\%$
test_reshape_td	60.2110μs	30.8421μs	32.4232 KOps/s	33.4734 KOps/s	$\color{#d91a1a}-3.14\%$
test_view_pytree	43.8300μs	21.2342μs	47.0939 KOps/s	48.6227 KOps/s	$\color{#d91a1a}-3.14\%$
test_view_td	19.3600μs	4.1562μs	240.6035 KOps/s	245.4158 KOps/s	$\color{#d91a1a}-1.96\%$
test_unbind_pytree	44.2310μs	26.6925μs	37.4637 KOps/s	38.5549 KOps/s	$\color{#d91a1a}-2.83\%$
test_unbind_td	83.8310μs	57.3718μs	17.4302 KOps/s	17.6359 KOps/s	$\color{#d91a1a}-1.17\%$
test_split_pytree	94.8720μs	25.7152μs	38.8875 KOps/s	42.1588 KOps/s	$\textbf{\color{#d91a1a}-7.76\%}$
test_split_td	71.3520μs	44.9081μs	22.2677 KOps/s	22.7556 KOps/s	$\color{#d91a1a}-2.14\%$
test_add_pytree	56.6310μs	33.6811μs	29.6902 KOps/s	32.1449 KOps/s	$\textbf{\color{#d91a1a}-7.64\%}$
test_add_td	76.2010μs	46.8318μs	21.3530 KOps/s	22.7685 KOps/s	$\textbf{\color{#d91a1a}-6.22\%}$
test_distributed	20.8110μs	5.6930μs	175.6552 KOps/s	176.3318 KOps/s	$\color{#d91a1a}-0.38\%$
test_tdmodule	89.4720μs	16.9690μs	58.9311 KOps/s	58.8565 KOps/s	$\color{#35bf28}+0.13\%$
test_tdmodule_dispatch	0.2298ms	33.2376μs	30.0865 KOps/s	30.1235 KOps/s	$\color{#d91a1a}-0.12\%$
test_tdseq	39.7800μs	20.2016μs	49.5011 KOps/s	48.4943 KOps/s	$\color{#35bf28}+2.08\%$
test_tdseq_dispatch	0.1357ms	36.3542μs	27.5071 KOps/s	27.1658 KOps/s	$\color{#35bf28}+1.26\%$
test_instantiation_functorch	1.7490ms	1.7130ms	583.7809 Ops/s	595.4224 Ops/s	$\color{#d91a1a}-1.96\%$
test_instantiation_td	1.8487ms	1.1924ms	838.6647 Ops/s	847.1615 Ops/s	$\color{#d91a1a}-1.00\%$
test_exec_functorch	0.2055ms	0.1625ms	6.1541 KOps/s	6.2629 KOps/s	$\color{#d91a1a}-1.74\%$
test_exec_functional_call	0.2208ms	0.1643ms	6.0867 KOps/s	6.2422 KOps/s	$\color{#d91a1a}-2.49\%$
test_exec_td	0.1883ms	0.1542ms	6.4845 KOps/s	6.6362 KOps/s	$\color{#d91a1a}-2.29\%$
test_exec_td_decorator	1.0352ms	0.2261ms	4.4232 KOps/s	4.4516 KOps/s	$\color{#d91a1a}-0.64\%$
test_vmap_mlp_speed[True-True]	1.1796ms	1.0949ms	913.3494 Ops/s	905.0262 Ops/s	$\color{#35bf28}+0.92\%$
test_vmap_mlp_speed[True-False]	0.7440ms	0.6384ms	1.5665 KOps/s	1.5885 KOps/s	$\color{#d91a1a}-1.39\%$
test_vmap_mlp_speed[False-True]	1.0706ms	1.0082ms	991.8382 Ops/s	991.0404 Ops/s	$\color{#35bf28}+0.08\%$
test_vmap_mlp_speed[False-False]	0.6086ms	0.5549ms	1.8021 KOps/s	1.7826 KOps/s	$\color{#35bf28}+1.09\%$
test_vmap_mlp_speed_decorator[True-True]	2.6210ms	1.8237ms	548.3348 Ops/s	544.1566 Ops/s	$\color{#35bf28}+0.77\%$
test_vmap_mlp_speed_decorator[True-False]	1.1735ms	0.7077ms	1.4131 KOps/s	1.4225 KOps/s	$\color{#d91a1a}-0.66\%$
test_vmap_mlp_speed_decorator[False-True]	2.1100ms	1.6404ms	609.6198 Ops/s	605.7532 Ops/s	$\color{#35bf28}+0.64\%$
test_vmap_mlp_speed_decorator[False-False]	1.0398ms	0.5955ms	1.6793 KOps/s	1.6746 KOps/s	$\color{#35bf28}+0.28\%$
test_vmap_transformer_speed[True-True]	12.9590ms	12.8752ms	77.6687 Ops/s	77.5838 Ops/s	$\color{#35bf28}+0.11\%$
test_vmap_transformer_speed[True-False]	10.7422ms	8.4431ms	118.4392 Ops/s	118.5362 Ops/s	$\color{#d91a1a}-0.08\%$
test_vmap_transformer_speed[False-True]	12.8261ms	12.7340ms	78.5302 Ops/s	78.2186 Ops/s	$\color{#35bf28}+0.40\%$
test_vmap_transformer_speed[False-False]	8.6754ms	8.3677ms	119.5065 Ops/s	119.4152 Ops/s	$\color{#35bf28}+0.08\%$
test_vmap_transformer_speed_decorator[True-True]	46.4936ms	44.8865ms	22.2784 Ops/s	22.8169 Ops/s	$\color{#d91a1a}-2.36\%$
test_vmap_transformer_speed_decorator[True-False]	99.3917ms	22.2757ms	44.8919 Ops/s	44.8315 Ops/s	$\color{#35bf28}+0.13\%$
test_vmap_transformer_speed_decorator[False-True]	44.4971ms	43.3295ms	23.0790 Ops/s	23.0208 Ops/s	$\color{#35bf28}+0.25\%$
test_vmap_transformer_speed_decorator[False-False]	0.1012s	21.9076ms	45.6463 Ops/s	45.4709 Ops/s	$\color{#35bf28}+0.39\%$