Skip to content

Conversation

@jumerckx
Copy link
Collaborator

@jumerckx jumerckx commented Feb 2, 2026

No description provided.

@jumerckx jumerckx self-assigned this Feb 2, 2026
@jumerckx jumerckx requested a review from wsmoses February 2, 2026 13:34

// Propagate sharding if present
if (auto shard = sdy::getShardingPerValue(op)) {
sdy::setShardings(newOp, shard);
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

@avik-pal can you support here I think this [and also recognizemultirotate, in retrospect] may be wrong. Specifically the original slice has only one result, whereas this has multiple, so we need a sharding which replicates it across all values

Copy link
Collaborator

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Yes we need to set for each of the results https://openxla.org/shardy/sdy_dialect#tensorshardingpervalueattr

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Thanks, will try this in a bit!

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

I tried here but unsure if this makes sense: a02f53d (#2023)

@jumerckx jumerckx force-pushed the jm/recognize_multislice branch from 6cce4ac to a02f53d Compare February 2, 2026 20:21
sdy::TensorShardingAttr singleShard = shardings[0];
SmallVector<sdy::TensorShardingAttr> newShardings(totalResults,
singleShard);
sdy::setShardings(newOp, sdy::TensorShardingPerValueAttr::get(
Copy link
Member

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

Looks good, @jumerckx can you also port this to recognizemultirotate?

Copy link
Collaborator Author

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

done

Copy link
Contributor

@github-actions github-actions bot left a comment

Choose a reason for hiding this comment

The reason will be displayed to describe this comment to others. Learn more.

EnzymeJAX Benchmarks

Details
Benchmark suite Current: a02f53d Previous: 3a9f897 Ratio
actmtch / JaXPipe / cpu / Primal 0.00000756341999704091 s 0.000007733859993095393 s 0.98
actmtch / Jax / cpu / Primal 0.000007082039994656952 s 0.000007217959982881439 s 0.98
actmtch / HLOOpt / cpu / Primal 0.000010064340012831962 s 0.00001069129997631535 s 0.94
actmtch / PartOpt / cpu / Primal 0.00000649998000199048 s 0.000007246099958138075 s 0.90
actmtch / IPartOpt / cpu / Primal 0.000006439779992888361 s 0.000007042160050332314 s 0.91
actmtch / DefOpt / cpu / Primal 0.000007414620010877116 s 0.000011662419983622386 s 0.64
actmtch / IDefOpt / cpu / Primal 0.000007664599997951882 s 0.000007531500014010817 s 1.02
actmtch / JaXPipe / cpu / Forward 0.00001123953999240257 s 0.0000115438599823392 s 0.97
actmtch / Jax / cpu / Forward 0.00000972533999629377 s 0.000010759360029624076 s 0.90
actmtch / HLOOpt / cpu / Forward 0.00001574148001282083 s 0.00001625256000806985 s 0.97
actmtch / PartOpt / cpu / Forward 0.0000153223200118191 s 0.000016039159936553916 s 0.96
actmtch / IPartOpt / cpu / Forward 0.000010404920003566075 s 0.000010876940004891366 s 0.96
actmtch / DefOpt / cpu / Forward 0.000014776780012653037 s 0.000016272039974865038 s 0.91
actmtch / IDefOpt / cpu / Forward 0.0000107448200037652 s 0.000011126260033051948 s 0.97
actmtch / JaXPipe / cpu / PreRev 0.000011881699992954964 s 0.000012882199980595033 s 0.92
actmtch / JaXPipe / cpu / PostRev 0.000011111459998573992 s 0.000011677599968606956 s 0.95
actmtch / JaXPipe / cpu / BothRev 0.00001541293999252957 s 0.000014915620004103403 s 1.03
actmtch / Jax / cpu / BothRev 0.00000990509999837741 s 0.000011236720019951465 s 0.88
actmtch / HLOOpt / cpu / PreRev 0.000011278160006895631 s 0.000012836999976570953 s 0.88
actmtch / HLOOpt / cpu / PostRev 0.000011307839997698466 s 0.000012310039946896725 s 0.92
actmtch / HLOOpt / cpu / BothRev 0.000013429919995360251 s 0.000015116180020413594 s 0.89
actmtch / PartOpt / cpu / PreRev 0.000011267820002558438 s 0.00001279665995753021 s 0.88
actmtch / PartOpt / cpu / PostRev 0.000010799159999805852 s 0.00001154188003965828 s 0.94
actmtch / PartOpt / cpu / BothRev 0.000011575799992442628 s 0.000012728359988614102 s 0.91
actmtch / IPartOpt / cpu / PreRev 0.000014993680013049016 s 0.000013585119968411164 s 1.10
actmtch / IPartOpt / cpu / PostRev 0.000010603139994600497 s 0.000011491159984871046 s 0.92
actmtch / IPartOpt / cpu / BothRev 0.000011769900008857804 s 0.000012637340014407529 s 0.93
actmtch / DefOpt / cpu / PreRev 0.000011963880001530923 s 0.00001295805999689037 s 0.92
actmtch / DefOpt / cpu / PostRev 0.000011982520013589238 s 0.000012994319995414116 s 0.92
actmtch / DefOpt / cpu / BothRev 0.000011685580000175831 s 0.00001293110000915476 s 0.90
actmtch / IDefOpt / cpu / PreRev 0.00001180694000368021 s 0.000013025639982515712 s 0.91
actmtch / IDefOpt / cpu / PostRev 0.000012207980005314313 s 0.000012815459986086352 s 0.95
actmtch / IDefOpt / cpu / BothRev 0.000011668719996578148 s 0.000012133060017731625 s 0.96
actmtch / JaXPipe / cuda / Primal 0.000002016 s 0.000002047 s 0.98
actmtch / Jax / cuda / Primal 0.000002047 s 0.000002015 s 1.02
actmtch / HLOOpt / cuda / Primal 0.000002047 s 0.000002016 s 1.02
actmtch / PartOpt / cuda / Primal 0.000002047 s 0.000002047 s 1
actmtch / IPartOpt / cuda / Primal 0.000002016 s 0.000002016 s 1
actmtch / DefOpt / cuda / Primal 0.000002047 s 0.000002016 s 1.02
actmtch / IDefOpt / cuda / Primal 0.000002047 s 0.000002047 s 1
actmtch / JaXPipe / cuda / Forward 0.000011776 s 0.000010369 s 1.14
actmtch / Jax / cuda / Forward 0.000011264 s 0.000010496 s 1.07
actmtch / HLOOpt / cuda / Forward 0.000010304 s 0.0000104 s 0.99
actmtch / PartOpt / cuda / Forward 0.000010464 s 0.000010176 s 1.03
actmtch / IPartOpt / cuda / Forward 0.00001024 s 0.000010496 s 0.98
actmtch / DefOpt / cuda / Forward 0.000010336 s 0.000010112 s 1.02
actmtch / IDefOpt / cuda / Forward 0.000010113 s 0.000009888 s 1.02
actmtch / JaXPipe / cuda / PreRev 0.00001024 s 0.000010527 s 0.97
actmtch / JaXPipe / cuda / PostRev 0.00001024 s 0.000009888 s 1.04
actmtch / JaXPipe / cuda / BothRev 0.000009248 s 0.000010112 s 0.91
actmtch / Jax / cuda / BothRev 0.00001008 s 0.000010304 s 0.98
actmtch / HLOOpt / cuda / PreRev 0.000009696 s 0.000010176 s 0.95
actmtch / HLOOpt / cuda / PostRev 0.00001024 s 0.000010144 s 1.01
actmtch / HLOOpt / cuda / BothRev 0.000010336 s 0.000009984 s 1.04
actmtch / PartOpt / cuda / PreRev 0.000010112 s 0.000010336 s 0.98
actmtch / PartOpt / cuda / PostRev 0.000010368 s 0.000010112 s 1.03
actmtch / PartOpt / cuda / BothRev 0.000010016 s 0.000010208 s 0.98
actmtch / IPartOpt / cuda / PreRev 0.000010336 s 0.000010176 s 1.02
actmtch / IPartOpt / cuda / PostRev 0.000010016 s 0.000010209 s 0.98
actmtch / IPartOpt / cuda / BothRev 0.000009888 s 0.000010143 s 0.97
actmtch / DefOpt / cuda / PreRev 0.000010176 s 0.00001024 s 0.99
actmtch / DefOpt / cuda / PostRev 0.000010336 s 0.00001056 s 0.98
actmtch / DefOpt / cuda / BothRev 0.000010208 s 0.000010144 s 1.01
actmtch / IDefOpt / cuda / PreRev 0.000010271 s 0.00001056 s 0.97
actmtch / IDefOpt / cuda / PostRev 0.000010176 s 0.000010112 s 1.01
actmtch / IDefOpt / cuda / BothRev 0.000010432 s 0.000010272 s 1.02
actmtch / JaXPipe / tpu / Primal 5.82925e-7 s 5.632000000000001e-7 s 1.04
actmtch / Jax / tpu / Primal 5.735e-7 s 6.06825e-7 s 0.95
actmtch / HLOOpt / tpu / Primal 0.000002169775 s 0.0000021014 s 1.03
actmtch / PartOpt / tpu / Primal 5.7325e-7 s 6.06525e-7 s 0.95
actmtch / IPartOpt / tpu / Primal 5.854e-7 s 5.623750000000001e-7 s 1.04
actmtch / DefOpt / tpu / Primal 0.00000205475 s 0.0000021606 s 0.95
actmtch / IDefOpt / tpu / Primal 0.0000021649 s 0.000002102475 s 1.03
actmtch / JaXPipe / tpu / Forward 0.000003862975 s 0.0000038357 s 1.01
actmtch / Jax / tpu / Forward 0.000001241925 s 0.000001215625 s 1.02
actmtch / HLOOpt / tpu / Forward 0.000003674625 s 0.0000039359 s 0.93
actmtch / PartOpt / tpu / Forward 0.000003893825 s 0.0000039167 s 0.99
actmtch / IPartOpt / tpu / Forward 0.0000036619 s 0.00000393865 s 0.93
actmtch / DefOpt / tpu / Forward 0.000003897375 s 0.00000391775 s 0.99
actmtch / IDefOpt / tpu / Forward 0.000003675925 s 0.000003933225 s 0.93
actmtch / JaXPipe / tpu / PreRev 0.0000037605 s 0.00000347695 s 1.08
actmtch / JaXPipe / tpu / PostRev 0.00000162295 s 0.0000016438 s 0.99
actmtch / JaXPipe / tpu / BothRev 0.0000037561 s 0.0000034815 s 1.08
actmtch / Jax / tpu / BothRev 0.000001616925 s 0.0000016392 s 0.99
actmtch / HLOOpt / tpu / PreRev 0.00000374915 s 0.00000348105 s 1.08
actmtch / HLOOpt / tpu / PostRev 0.000003448125 s 0.00000340075 s 1.01
actmtch / HLOOpt / tpu / BothRev 0.0000037474 s 0.0000034776 s 1.08
actmtch / PartOpt / tpu / PreRev 0.0000034447 s 0.0000034090500000000004 s 1.01
actmtch / PartOpt / tpu / PostRev 0.000001675975 s 0.0000015899 s 1.05
actmtch / PartOpt / tpu / BothRev 0.0000034446 s 0.000003414025 s 1.01
actmtch / IPartOpt / tpu / PreRev 0.000003744925 s 0.0000034937750000000003 s 1.07
actmtch / IPartOpt / tpu / PostRev 0.0000016282749999999997 s 0.00000164185 s 0.99
actmtch / IPartOpt / tpu / BothRev 0.00000374005 s 0.0000034793750000000005 s 1.07
actmtch / DefOpt / tpu / PreRev 0.000003454525 s 0.0000033998 s 1.02
actmtch / DefOpt / tpu / PostRev 0.000003668975 s 0.000003417725 s 1.07
actmtch / DefOpt / tpu / BothRev 0.0000034479 s 0.0000034063500000000004 s 1.01
actmtch / IDefOpt / tpu / PreRev 0.000003745925 s 0.000003474375 s 1.08
actmtch / IDefOpt / tpu / PostRev 0.000003440325 s 0.000003406225 s 1.01
actmtch / IDefOpt / tpu / BothRev 0.000003752575 s 0.00000346605 s 1.08
actmtch / JaXPipe / cpu / Primal 0.000012929 s 0.000007733859993095393 s 1.67
actmtch / Jax / cpu / Primal 0.000013449 s 0.000007217959982881439 s 1.86
actmtch / HLOOpt / cpu / Primal 0.000013907 s 0.00001069129997631535 s 1.30
actmtch / PartOpt / cpu / Primal 0.000013117 s 0.000007246099958138075 s 1.81
actmtch / IPartOpt / cpu / Primal 0.000013181999999999998 s 0.000007042160050332314 s 1.87
actmtch / DefOpt / cpu / Primal 0.000013717 s 0.000011662419983622386 s 1.18
actmtch / IDefOpt / cpu / Primal 0.000013907 s 0.000007531500014010817 s 1.85
actmtch / JaXPipe / cpu / Forward 0.000019495 s 0.0000115438599823392 s 1.69
actmtch / Jax / cpu / Forward 0.000017757 s 0.000010759360029624076 s 1.65
actmtch / HLOOpt / cpu / Forward 0.000018749 s 0.00001625256000806985 s 1.15
actmtch / PartOpt / cpu / Forward 0.000019507 s 0.000016039159936553916 s 1.22
actmtch / IPartOpt / cpu / Forward 0.00001963 s 0.000010876940004891366 s 1.80
actmtch / DefOpt / cpu / Forward 0.000019148 s 0.000016272039974865038 s 1.18
actmtch / IDefOpt / cpu / Forward 0.000019428 s 0.000011126260033051948 s 1.75
actmtch / JaXPipe / cpu / PreRev 0.000019074 s 0.000012882199980595033 s 1.48
actmtch / JaXPipe / cpu / PostRev 0.000018121 s 0.000011677599968606956 s 1.55
actmtch / JaXPipe / cpu / BothRev 0.000019900000000000003 s 0.000014915620004103403 s 1.33
actmtch / Jax / cpu / BothRev 0.000017576000000000002 s 0.000011236720019951465 s 1.56
actmtch / HLOOpt / cpu / PreRev 0.000018922000000000003 s 0.000012836999976570953 s 1.47
actmtch / HLOOpt / cpu / PostRev 0.000019004 s 0.000012310039946896725 s 1.54
actmtch / HLOOpt / cpu / BothRev 0.000019718 s 0.000015116180020413594 s 1.30
actmtch / PartOpt / cpu / PreRev 0.000018567 s 0.00001279665995753021 s 1.45
actmtch / PartOpt / cpu / PostRev 0.000017434000000000003 s 0.00001154188003965828 s 1.51
actmtch / PartOpt / cpu / BothRev 0.000019821 s 0.000012728359988614102 s 1.56
actmtch / IPartOpt / cpu / PreRev 0.000019461 s 0.000013585119968411164 s 1.43
actmtch / IPartOpt / cpu / PostRev 0.000018733 s 0.000011491159984871046 s 1.63
actmtch / IPartOpt / cpu / BothRev 0.000021168 s 0.000012637340014407529 s 1.68
actmtch / DefOpt / cpu / PreRev 0.000020537 s 0.00001295805999689037 s 1.58
actmtch / DefOpt / cpu / PostRev 0.000020174 s 0.000012994319995414116 s 1.55
actmtch / DefOpt / cpu / BothRev 0.000020395 s 0.00001293110000915476 s 1.58
actmtch / IDefOpt / cpu / PreRev 0.000019011 s 0.000013025639982515712 s 1.46
actmtch / IDefOpt / cpu / PostRev 0.000019125 s 0.000012815459986086352 s 1.49
actmtch / IDefOpt / cpu / BothRev 0.000019777 s 0.000012133060017731625 s 1.63
add_one / JaXPipe / cpu / Primal 0.000007761200001823455 s 0.000007868599977882695 s 0.99
add_one / Jax / cpu / Primal 0.000007347579999077424 s 0.00000741951998861623 s 0.99
add_one / HLOOpt / cpu / Primal 0.000010474620009972569 s 0.000010513959987292765 s 1.00
add_one / PartOpt / cpu / Primal 0.000006494919989563641 s 0.000007595719998789718 s 0.86
add_one / IPartOpt / cpu / Primal 0.000007159340004818659 s 0.000007777159989927895 s 0.92
add_one / DefOpt / cpu / Primal 0.00001166422000324019 s 0.0000112859800356091 s 1.03
add_one / IDefOpt / cpu / Primal 0.000006871599991882249 s 0.000007331239994528005 s 0.94
add_one / JaXPipe / cpu / Forward 0.000010595259993806394 s 0.000012304339988986613 s 0.86
add_one / Jax / cpu / Forward 0.000010391159996743226 s 0.00001202503997774329 s 0.86
add_one / HLOOpt / cpu / Forward 0.000014891980001721094 s 0.000013736400032939855 s 1.08
add_one / PartOpt / cpu / Forward 0.000015648139990389608 s 0.000016401239990955217 s 0.95
add_one / IPartOpt / cpu / Forward 0.000010140320005120883 s 0.000011726520024240016 s 0.86
add_one / DefOpt / cpu / Forward 0.000015035219983019487 s 0.000016018619999158545 s 0.94
add_one / IDefOpt / cpu / Forward 0.000010401800004729011 s 0.000011999660018773284 s 0.87
add_one / JaXPipe / cpu / PreRev 0.000012283580006169358 s 0.000013027700024395015 s 0.94
add_one / JaXPipe / cpu / PostRev 0.000012171880000551029 s 0.000012822980006603756 s 0.95
add_one / JaXPipe / cpu / BothRev 0.000016786479986876656 s 0.000016642000000501868 s 1.01
add_one / Jax / cpu / BothRev 0.000012075240008471156 s 0.000012804839980162796 s 0.94
add_one / HLOOpt / cpu / PreRev 0.00001234626000496064 s 0.000014124400022410555 s 0.87
add_one / HLOOpt / cpu / PostRev 0.00001614499999959662 s 0.00001704923999568564 s 0.95
add_one / HLOOpt / cpu / BothRev 0.000014181719993757724 s 0.00001847005998570239 s 0.77
add_one / PartOpt / cpu / PreRev 0.000012468140000692076 s 0.000012570299977596731 s 0.99
add_one / PartOpt / cpu / PostRev 0.000012127179991239243 s 0.00001297972001339076 s 0.93
add_one / PartOpt / cpu / BothRev 0.000012058080010319829 s 0.000012936780049130902 s 0.93
add_one / IPartOpt / cpu / PreRev 0.000012238579993208988 s 0.000017668959999355137 s 0.69
add_one / IPartOpt / cpu / PostRev 0.00001201559999572055 s 0.000012812400054826868 s 0.94
add_one / IPartOpt / cpu / BothRev 0.0000123378800071805 s 0.000012878280012955656 s 0.96
add_one / DefOpt / cpu / PreRev 0.000011873019989252498 s 0.000012653959993258468 s 0.94
add_one / DefOpt / cpu / PostRev 0.000012468199997783811 s 0.000013062259986327264 s 0.95
add_one / DefOpt / cpu / BothRev 0.000011786079994635656 s 0.000012414299990268771 s 0.95
add_one / IDefOpt / cpu / PreRev 0.00001146060000792204 s 0.000013297360037540785 s 0.86
add_one / IDefOpt / cpu / PostRev 0.0000126086599948394 s 0.00001240428003256966 s 1.02
add_one / IDefOpt / cpu / BothRev 0.000012307079996389805 s 0.0000127792000057525 s 0.96
add_one / JaXPipe / cuda / Primal 0.0000019200000000000003 s 0.000001951 s 0.98
add_one / Jax / cuda / Primal 0.000001951 s 0.000001951 s 1
add_one / HLOOpt / cuda / Primal 0.0000019200000000000003 s 0.000001951 s 0.98
add_one / PartOpt / cuda / Primal 0.000001951 s 0.000001951 s 1
add_one / IPartOpt / cuda / Primal 0.0000019200000000000003 s 0.000001951 s 0.98
add_one / DefOpt / cuda / Primal 0.000001951 s 0.000001951 s 1
add_one / IDefOpt / cuda / Primal 0.0000019200000000000003 s 0.000001919 s 1.00
add_one / JaXPipe / cuda / Forward 0.000010432 s 0.000010176 s 1.03
add_one / Jax / cuda / Forward 0.000010688 s 0.00000992 s 1.08
add_one / HLOOpt / cuda / Forward 0.00001056 s 0.0000112 s 0.94
add_one / PartOpt / cuda / Forward 0.000010336 s 0.000011424 s 0.90
add_one / IPartOpt / cuda / Forward 0.000010432 s 0.00001136 s 0.92
add_one / DefOpt / cuda / Forward 0.000010432 s 0.000011455999999999998 s 0.91
add_one / IDefOpt / cuda / Forward 0.0000104 s 0.000011679 s 0.89
add_one / JaXPipe / cuda / PreRev 0.000027328 s 0.000026848 s 1.02
add_one / JaXPipe / cuda / PostRev 0.000025984 s 0.00002496 s 1.04
add_one / JaXPipe / cuda / BothRev 0.000025664 s 0.000028895 s 0.89
add_one / Jax / cuda / BothRev 0.000025376 s 0.000029056 s 0.87
add_one / HLOOpt / cuda / PreRev 0.000025984 s 0.000025568 s 1.02
add_one / HLOOpt / cuda / PostRev 0.000025087 s 0.000033024 s 0.76
add_one / HLOOpt / cuda / BothRev 0.000025088 s 0.000025856 s 0.97
add_one / PartOpt / cuda / PreRev 0.000025408 s 0.0000264 s 0.96
add_one / PartOpt / cuda / PostRev 0.000025856 s 0.00002624 s 0.99
add_one / PartOpt / cuda / BothRev 0.000026207 s 0.000025568 s 1.02
add_one / IPartOpt / cuda / PreRev 0.000025952 s 0.000025888 s 1.00
add_one / IPartOpt / cuda / PostRev 0.000026048 s 0.000025856 s 1.01
add_one / IPartOpt / cuda / BothRev 0.000036479 s 0.000026112 s 1.40
add_one / DefOpt / cuda / PreRev 0.00002544 s 0.00002576 s 0.99
add_one / DefOpt / cuda / PostRev 0.000025727 s 0.000025568 s 1.01
add_one / DefOpt / cuda / BothRev 0.000025759 s 0.0000264 s 0.98
add_one / IDefOpt / cuda / PreRev 0.000025888 s 0.000026016 s 1.00
add_one / IDefOpt / cuda / PostRev 0.000025504 s 0.000026209 s 0.97
add_one / IDefOpt / cuda / BothRev 0.000025792 s 0.000025471 s 1.01
add_one / JaXPipe / tpu / Primal 0.000001452625 s 0.000001421075 s 1.02
add_one / Jax / tpu / Primal 0.0000014521999999999998 s 0.000001401925 s 1.04
add_one / HLOOpt / tpu / Primal 0.00000144645 s 0.000001429425 s 1.01
add_one / PartOpt / tpu / Primal 0.000001457825 s 0.000001411375 s 1.03
add_one / IPartOpt / tpu / Primal 0.0000014502250000000002 s 0.000001426825 s 1.02
add_one / DefOpt / tpu / Primal 0.0000014610500000000002 s 0.000001401575 s 1.04
add_one / IDefOpt / tpu / Primal 0.000001454475 s 0.00000142725 s 1.02
add_one / JaXPipe / tpu / Forward 0.0000019132 s 0.000001856925 s 1.03
add_one / Jax / tpu / Forward 0.000001870525 s 0.000001847075 s 1.01
add_one / HLOOpt / tpu / Forward 0.000001900175 s 0.0000018547250000000005 s 1.02
add_one / PartOpt / tpu / Forward 0.00000186445 s 0.000001840325 s 1.01
add_one / IPartOpt / tpu / Forward 0.000001903525 s 0.0000018468 s 1.03
add_one / DefOpt / tpu / Forward 0.0000018628 s 0.00000183415 s 1.02
add_one / IDefOpt / tpu / Forward 0.00000191455 s 0.000001852925 s 1.03
add_one / JaXPipe / tpu / PreRev 0.000002253275 s 0.000002232875 s 1.01
add_one / JaXPipe / tpu / PostRev 0.0000022977 s 0.0000022385 s 1.03
add_one / JaXPipe / tpu / BothRev 0.000002258575 s 0.000002232425 s 1.01
add_one / Jax / tpu / BothRev 0.000002304225 s 0.000002235125 s 1.03
add_one / HLOOpt / tpu / PreRev 0.000002264075 s 0.00000223105 s 1.01
add_one / HLOOpt / tpu / PostRev 0.00000229775 s 0.00000224865 s 1.02
add_one / HLOOpt / tpu / BothRev 0.0000022589 s 0.00000223125 s 1.01
add_one / PartOpt / tpu / PreRev 0.0000023003 s 0.00000225265 s 1.02
add_one / PartOpt / tpu / PostRev 0.000002264575 s 0.0000022428 s 1.01
add_one / PartOpt / tpu / BothRev 0.00000230155 s 0.0000022354750000000004 s 1.03
add_one / IPartOpt / tpu / PreRev 0.0000022529 s 0.0000022359000000000005 s 1.01
add_one / IPartOpt / tpu / PostRev 0.000002290775 s 0.00000225115 s 1.02
add_one / IPartOpt / tpu / BothRev 0.00000226445 s 0.0000022322 s 1.01
add_one / DefOpt / tpu / PreRev 0.00000230435 s 0.000002243475 s 1.03
add_one / DefOpt / tpu / PostRev 0.000002252625 s 0.0000022352 s 1.01
add_one / DefOpt / tpu / BothRev 0.000002293075 s 0.0000022375 s 1.02
add_one / IDefOpt / tpu / PreRev 0.0000022683 s 0.000002236975 s 1.01
add_one / IDefOpt / tpu / PostRev 0.000002294125 s 0.0000022343 s 1.03
add_one / IDefOpt / tpu / BothRev 0.000002264325 s 0.0000022451 s 1.01
add_one / JaXPipe / cpu / Primal 0.000013666 s 0.000007868599977882695 s 1.74
add_one / Jax / cpu / Primal 0.000013327 s 0.00000741951998861623 s 1.80
add_one / HLOOpt / cpu / Primal 0.000012997 s 0.000010513959987292765 s 1.24
add_one / PartOpt / cpu / Primal 0.000013167 s 0.000007595719998789718 s 1.73
add_one / IPartOpt / cpu / Primal 0.000012796 s 0.000007777159989927895 s 1.65
add_one / DefOpt / cpu / Primal 0.0000127 s 0.0000112859800356091 s 1.13
add_one / IDefOpt / cpu / Primal 0.000012699 s 0.000007331239994528005 s 1.73
add_one / JaXPipe / cpu / Forward 0.000017386999999999998 s 0.000012304339988986613 s 1.41
add_one / Jax / cpu / Forward 0.000017539 s 0.00001202503997774329 s 1.46
add_one / HLOOpt / cpu / Forward 0.000017621000000000003 s 0.000013736400032939855 s 1.28
add_one / PartOpt / cpu / Forward 0.000017927 s 0.000016401239990955217 s 1.09
add_one / IPartOpt / cpu / Forward 0.000017805 s 0.000011726520024240016 s 1.52
add_one / DefOpt / cpu / Forward 0.000018138 s 0.000016018619999158545 s 1.13
add_one / IDefOpt / cpu / Forward 0.000017542 s 0.000011999660018773284 s 1.46
add_one / JaXPipe / cpu / PreRev 0.0000203 s 0.000013027700024395015 s 1.56
add_one / JaXPipe / cpu / PostRev 0.000020275 s 0.000012822980006603756 s 1.58
add_one / JaXPipe / cpu / BothRev 0.000020403 s 0.000016642000000501868 s 1.23
add_one / Jax / cpu / BothRev 0.00002 s 0.000012804839980162796 s 1.56
add_one / HLOOpt / cpu / PreRev 0.000020867 s 0.000014124400022410555 s 1.48
add_one / HLOOpt / cpu / PostRev 0.00002049 s 0.00001704923999568564 s 1.20
add_one / HLOOpt / cpu / BothRev 0.000021379 s 0.00001847005998570239 s 1.16
add_one / PartOpt / cpu / PreRev 0.000020639 s 0.000012570299977596731 s 1.64
add_one / PartOpt / cpu / PostRev 0.000021531 s 0.00001297972001339076 s 1.66
add_one / PartOpt / cpu / BothRev 0.000020029 s 0.000012936780049130902 s 1.55
add_one / IPartOpt / cpu / PreRev 0.000019992 s 0.000017668959999355137 s 1.13
add_one / IPartOpt / cpu / PostRev 0.00001998 s 0.000012812400054826868 s 1.56
add_one / IPartOpt / cpu / BothRev 0.000020275 s 0.000012878280012955656 s 1.57
add_one / DefOpt / cpu / PreRev 0.000019697 s 0.000012653959993258468 s 1.56
add_one / DefOpt / cpu / PostRev 0.000020108 s 0.000013062259986327264 s 1.54
add_one / DefOpt / cpu / BothRev 0.000020102 s 0.000012414299990268771 s 1.62
add_one / IDefOpt / cpu / PreRev 0.000019429 s 0.000013297360037540785 s 1.46
add_one / IDefOpt / cpu / PostRev 0.000019685 s 0.00001240428003256966 s 1.59
add_one / IDefOpt / cpu / BothRev 0.00001947 s 0.0000127792000057525 s 1.52
add_two / JaXPipe / cpu / Primal 0.000007928140000785788 s 0.000007723359985902789 s 1.03
add_two / Jax / cpu / Primal 0.000006638699996983632 s 0.00000696983996022027 s 0.95
add_two / HLOOpt / cpu / Primal 0.000011250580000705667 s 0.000010501580027266754 s 1.07
add_two / PartOpt / cpu / Primal 0.000006794200005515449 s 0.000007528760006607626 s 0.90
add_two / IPartOpt / cpu / Primal 0.000007167560002017126 s 0.000007299680046344292 s 0.98
add_two / DefOpt / cpu / Primal 0.000011758559996906116 s 0.00001088045998585585 s 1.08
add_two / IDefOpt / cpu / Primal 0.000007010839995018614 s 0.000007032020039332565 s 1.00
add_two / JaXPipe / cpu / Forward 0.000010698520000005374 s 0.00001198598002702056 s 0.89
add_two / Jax / cpu / Forward 0.000010640840005180509 s 0.000011226719998376213 s 0.95
add_two / HLOOpt / cpu / Forward 0.000015000320006492984 s 0.000016235859975495258 s 0.92
add_two / PartOpt / cpu / Forward 0.00001506551999000294 s 0.00001550988001326914 s 0.97
add_two / IPartOpt / cpu / Forward 0.000011008760006916418 s 0.00001129524000134552 s 0.97
add_two / DefOpt / cpu / Forward 0.000014892639999288804 s 0.00001601973995093431 s 0.93
add_two / IDefOpt / cpu / Forward 0.000010668099996564706 s 0.000011547620006240325 s 0.92
add_two / JaXPipe / cpu / PreRev 0.000015182940007889556 s 0.00001604412000233424 s 0.95
add_two / JaXPipe / cpu / PostRev 0.000014385699989816204 s 0.000014845979967503808 s 0.97
add_two / JaXPipe / cpu / BothRev 0.000014681379989269771 s 0.000015111439970496575 s 0.97
add_two / Jax / cpu / BothRev 0.000014747539994459658 s 0.000015581239968014415 s 0.95
add_two / HLOOpt / cpu / PreRev 0.000014457459990353529 s 0.000015555060026599677 s 0.93
add_two / HLOOpt / cpu / PostRev 0.00001485273998696357 s 0.000015376660021502174 s 0.97
add_two / HLOOpt / cpu / BothRev 0.000016687999996065627 s 0.000016302340027323227 s 1.02
add_two / PartOpt / cpu / PreRev 0.000014322499998797866 s 0.000015145600027608452 s 0.95
add_two / PartOpt / cpu / PostRev 0.000014512039999772242 s 0.000014814679998380598 s 0.98
add_two / PartOpt / cpu / BothRev 0.00001473678000138534 s 0.0000159910200000013 s 0.92
add_two / IPartOpt / cpu / PreRev 0.000014869079993786729 s 0.000016129919977174723 s 0.92
add_two / IPartOpt / cpu / PostRev 0.000014934120003999852 s 0.000015059539964568103 s 0.99
add_two / IPartOpt / cpu / BothRev 0.000014514140004848742 s 0.000015380740060209063 s 0.94
add_two / DefOpt / cpu / PreRev 0.000015040559992485214 s 0.00001517411998065654 s 0.99
add_two / DefOpt / cpu / PostRev 0.00001465556000084689 s 0.000015974059970176313 s 0.92
add_two / DefOpt / cpu / BothRev 0.000015040280018183694 s 0.000015469319996554987 s 0.97
add_two / IDefOpt / cpu / PreRev 0.000014672060001430508 s 0.00001592369992977183 s 0.92
add_two / IDefOpt / cpu / PostRev 0.000014792000006309537 s 0.000015961200042511336 s 0.93
add_two / IDefOpt / cpu / BothRev 0.000014957959995172133 s 0.000015459220012417063 s 0.97
add_two / JaXPipe / cuda / Primal 0.000001951 s 0.000001951 s 1
add_two / Jax / cuda / Primal 0.000001951 s 0.000001951 s 1
add_two / HLOOpt / cuda / Primal 0.000001951 s 0.000001951 s 1
add_two / PartOpt / cuda / Primal 0.000001951 s 0.0000019200000000000003 s 1.02
add_two / IPartOpt / cuda / Primal 0.000001951 s 0.000001951 s 1
add_two / DefOpt / cuda / Primal 0.0000019200000000000003 s 0.000001951 s 0.98
add_two / IDefOpt / cuda / Primal 0.000001951 s 0.000001951 s 1
add_two / JaXPipe / cuda / Forward 0.000009952 s 0.000009984 s 1.00
add_two / Jax / cuda / Forward 0.00001024 s 0.000009984 s 1.03
add_two / HLOOpt / cuda / Forward 0.00001008 s 0.000009984 s 1.01
add_two / PartOpt / cuda / Forward 0.000010016 s 0.0000104 s 0.96
add_two / IPartOpt / cuda / Forward 0.000010016 s 0.000010368 s 0.97
add_two / DefOpt / cuda / Forward 0.000009856 s 0.000009824 s 1.00
add_two / IDefOpt / cuda / Forward 0.000009792 s 0.000010015 s 0.98
add_two / JaXPipe / cuda / PreRev 0.00003264 s 0.000032928 s 0.99
add_two / JaXPipe / cuda / PostRev 0.000031936 s 0.000032832 s 0.97
add_two / JaXPipe / cuda / BothRev 0.000031776 s 0.000032767999999999995 s 0.97
add_two / Jax / cuda / BothRev 0.000031424 s 0.000033376 s 0.94
add_two / HLOOpt / cuda / PreRev 0.00003248 s 0.000032992 s 0.98
add_two / HLOOpt / cuda / PostRev 0.000033248 s 0.000033056 s 1.01
add_two / HLOOpt / cuda / BothRev 0.000032864 s 0.000034848 s 0.94
add_two / PartOpt / cuda / PreRev 0.000032992 s 0.000033343 s 0.99
add_two / PartOpt / cuda / PostRev 0.000032416 s 0.000032576 s 1.00
add_two / PartOpt / cuda / BothRev 0.000032896000000000005 s 0.000032672 s 1.01
add_two / IPartOpt / cuda / PreRev 0.000033312 s 0.000032897 s 1.01
add_two / IPartOpt / cuda / PostRev 0.000033184 s 0.000032832 s 1.01
add_two / IPartOpt / cuda / BothRev 0.000043456 s 0.000033055 s 1.31
add_two / DefOpt / cuda / PreRev 0.000032544 s 0.000032991 s 0.99
add_two / DefOpt / cuda / PostRev 0.000034112 s 0.00003328 s 1.03
add_two / DefOpt / cuda / BothRev 0.000033152000000000004 s 0.000038144 s 0.87
add_two / IDefOpt / cuda / PreRev 0.000034656 s 0.000033536000000000006 s 1.03
add_two / IDefOpt / cuda / PostRev 0.00003408 s 0.0000376 s 0.91
add_two / IDefOpt / cuda / BothRev 0.000033056 s 0.000033311 s 0.99
add_two / JaXPipe / tpu / Primal 0.000001395075 s 0.0000014412 s 0.97
add_two / Jax / tpu / Primal 0.0000014105500000000002 s 0.000001480425 s 0.95
add_two / HLOOpt / tpu / Primal 0.0000013944500000000005 s 0.00000142565 s 0.98
add_two / PartOpt / tpu / Primal 0.000001401025 s 0.000001473 s 0.95
add_two / IPartOpt / tpu / Primal 0.0000013995 s 0.000001433925 s 0.98
add_two / DefOpt / tpu / Primal 0.0000014139 s 0.000001476275 s 0.96
add_two / IDefOpt / tpu / Primal 0.00000139005 s 0.000001433925 s 0.97
add_two / JaXPipe / tpu / Forward 0.0000017845749999999998 s 0.000001823425 s 0.98
add_two / Jax / tpu / Forward 0.00000182525 s 0.000001831975 s 1.00
add_two / HLOOpt / tpu / Forward 0.000001793225 s 0.00000183065 s 0.98
add_two / PartOpt / tpu / Forward 0.000001832425 s 0.00000182505 s 1.00
add_two / IPartOpt / tpu / Forward 0.000001804175 s 0.00000183225 s 0.98
add_two / DefOpt / tpu / Forward 0.000001823575 s 0.000001836275 s 0.99
add_two / IDefOpt / tpu / Forward 0.0000017948249999999998 s 0.000001826925 s 0.98
add_two / JaXPipe / tpu / PreRev 0.0000028249 s 0.00000283795 s 1.00
add_two / JaXPipe / tpu / PostRev 0.000002735375 s 0.00000275535 s 0.99
add_two / JaXPipe / tpu / BothRev 0.0000028071 s 0.0000028325 s 0.99
add_two / Jax / tpu / BothRev 0.00000273835 s 0.000002751725 s 1.00
add_two / HLOOpt / tpu / PreRev 0.0000028137000000000003 s 0.0000028398250000000004 s 0.99
add_two / HLOOpt / tpu / PostRev 0.000002724225 s 0.0000027475 s 0.99
add_two / HLOOpt / tpu / BothRev 0.000002818825 s 0.00000284185 s 0.99
add_two / PartOpt / tpu / PreRev 0.000002728675 s 0.000002757825 s 0.99
add_two / PartOpt / tpu / PostRev 0.000002817875 s 0.000002824325 s 1.00
add_two / PartOpt / tpu / BothRev 0.0000027294000000000005 s 0.0000027582250000000006 s 0.99
add_two / IPartOpt / tpu / PreRev 0.000002814325 s 0.00000283555 s 0.99
add_two / IPartOpt / tpu / PostRev 0.000002733575 s 0.000002745575 s 1.00
add_two / IPartOpt / tpu / BothRev 0.000002818025 s 0.000002840825 s 0.99
add_two / DefOpt / tpu / PreRev 0.00000273515 s 0.000002755975 s 0.99
add_two / DefOpt / tpu / PostRev 0.000002819775 s 0.0000028389 s 0.99
add_two / DefOpt / tpu / BothRev 0.000002727725 s 0.0000027441 s 0.99
add_two / IDefOpt / tpu / PreRev 0.0000028079 s 0.0000028370250000000006 s 0.99
add_two / IDefOpt / tpu / PostRev 0.00000272425 s 0.0000027601 s 0.99
add_two / IDefOpt / tpu / BothRev 0.00000280715 s 0.000002846225 s 0.99
add_two / JaXPipe / cpu / Primal 0.000013476 s 0.000007723359985902789 s 1.74
add_two / Jax / cpu / Primal 0.000013404 s 0.00000696983996022027 s 1.92
add_two / HLOOpt / cpu / Primal 0.000013378 s 0.000010501580027266754 s 1.27
add_two / PartOpt / cpu / Primal 0.000012966 s 0.000007528760006607626 s 1.72
add_two / IPartOpt / cpu / Primal 0.000013526 s 0.000007299680046344292 s 1.85
add_two / DefOpt / cpu / Primal 0.000013365 s 0.00001088045998585585 s 1.23
add_two / IDefOpt / cpu / Primal 0.000013258 s 0.000007032020039332565 s 1.89
add_two / JaXPipe / cpu / Forward 0.000018375 s 0.00001198598002702056 s 1.53
add_two / Jax / cpu / Forward 0.000017758 s 0.000011226719998376213 s 1.58
add_two / HLOOpt / cpu / Forward 0.000017978 s 0.000016235859975495258 s 1.11
add_two / PartOpt / cpu / Forward 0.000018255 s 0.00001550988001326914 s 1.18
add_two / IPartOpt / cpu / Forward 0.000018343 s 0.00001129524000134552 s 1.62
add_two / DefOpt / cpu / Forward 0.000018222 s 0.00001601973995093431 s 1.14
add_two / IDefOpt / cpu / Forward 0.000018044 s 0.000011547620006240325 s 1.56
add_two / JaXPipe / cpu / PreRev 0.000025639 s 0.00001604412000233424 s 1.60
add_two / JaXPipe / cpu / PostRev 0.000025921 s 0.000014845979967503808 s 1.75
add_two / JaXPipe / cpu / BothRev 0.000026127 s 0.000015111439970496575 s 1.73
add_two / Jax / cpu / BothRev 0.000024385 s 0.000015581239968014415 s 1.57
add_two / HLOOpt / cpu / PreRev 0.000024822 s 0.000015555060026599677 s 1.60
add_two / HLOOpt / cpu / PostRev 0.000025743 s 0.000015376660021502174 s 1.67
add_two / HLOOpt / cpu / BothRev 0.000025137 s 0.000016302340027323227 s 1.54
add_two / PartOpt / cpu / PreRev 0.000024957 s 0.000015145600027608452 s 1.65
add_two / PartOpt / cpu / PostRev 0.000025996 s 0.000014814679998380598 s 1.75
add_two / PartOpt / cpu / BothRev 0.000025281 s 0.0000159910200000013 s 1.58
add_two / IPartOpt / cpu / PreRev 0.000025158 s 0.000016129919977174723 s 1.56
add_two / IPartOpt / cpu / PostRev 0.00002554 s 0.000015059539964568103 s 1.70
add_two / IPartOpt / cpu / BothRev 0.00002555 s 0.000015380740060209063 s 1.66
add_two / DefOpt / cpu / PreRev 0.000025359 s 0.00001517411998065654 s 1.67
add_two / DefOpt / cpu / PostRev 0.000024429 s 0.000015974059970176313 s 1.53
add_two / DefOpt / cpu / BothRev 0.00002445 s 0.000015469319996554987 s 1.58
add_two / IDefOpt / cpu / PreRev 0.000024202 s 0.00001592369992977183 s 1.52
add_two / IDefOpt / cpu / PostRev 0.000025936 s 0.000015961200042511336 s 1.62
add_two / IDefOpt / cpu / BothRev 0.000026333 s 0.000015459220012417063 s 1.70
cache / JaXPipe / cpu / Primal 0.000006704759996409848 s 0.000006956840015845956 s 0.96
cache / Jax / cpu / Primal 0.000007148000001961918 s 0.000007311219933399116 s 0.98
cache / HLOOpt / cpu / Primal 0.0000067763000106424445 s 0.000007593460013595177 s 0.89
cache / PartOpt / cpu / Primal 0.000007507680008984608 s 0.000007535879994975403 s 1.00
cache / IPartOpt / cpu / Primal 0.000006921660001353303 s 0.000007371540013991762 s 0.94
cache / DefOpt / cpu / Primal 0.000006444840003041463 s 0.000007683940011702362 s 0.84
cache / IDefOpt / cpu / Primal 0.0000065485799996167775 s 0.0000073235999934695425 s 0.89
cache / JaXPipe / cpu / Forward 0.00001400896000177454 s 0.000014550459982274333 s 0.96
cache / Jax / cpu / Forward 0.000014287819994933673 s 0.000014931959995010404 s 0.96
cache / HLOOpt / cpu / Forward 0.00001953021999952398 s 0.00001945838004758116 s 1.00
cache / PartOpt / cpu / Forward 0.000020570959986798697 s 0.00001901271997667209 s 1.08
cache / IPartOpt / cpu / Forward 0.000015427460000410064 s 0.0000151107200235856 s 1.02
cache / DefOpt / cpu / Forward 0.00002085963999661544 s 0.00001980758002900984 s 1.05
cache / IDefOpt / cpu / Forward 0.000014792300007684389 s 0.000014846979993308196 s 1.00
cache / JaXPipe / cpu / PreRev 0.000017505580012766585 s 0.000015902700033620933 s 1.10
cache / JaXPipe / cpu / PostRev 0.000021765780002169776 s 0.000021121619975019717 s 1.03
cache / JaXPipe / cpu / BothRev 0.000022876399989399945 s 0.000017116039998654742 s 1.34
cache / Jax / cpu / BothRev 0.00002155626000103439 s 0.00002173714004129579 s 0.99
cache / HLOOpt / cpu / PreRev 0.00001658504000488392 s 0.000017123339985118948 s 0.97
cache / HLOOpt / cpu / PostRev 0.000020419739998942533 s 0.00001864448002379504 s 1.10
cache / HLOOpt / cpu / BothRev 0.0000212658799887322 s 0.000018996239978150696 s 1.12
cache / PartOpt / cpu / PreRev 0.00001629732000083095 s 0.00001642888004425913 s 0.99
cache / PartOpt / cpu / PostRev 0.00002127841999708835 s 0.000020945439973729663 s 1.02
cache / PartOpt / cpu / BothRev 0.00001688822000460277 s 0.000016367300049751065 s 1.03
cache / IPartOpt / cpu / PreRev 0.00002261333999058479 s 0.000017679780003163616 s 1.28
cache / IPartOpt / cpu / PostRev 0.00002229039999292581 s 0.000022093920060797247 s 1.01
cache / IPartOpt / cpu / BothRev 0.000016651019996061224 s 0.000016750399972806917 s 0.99
cache / DefOpt / cpu / PreRev 0.00001773136000792874 s 0.000016780159985501086 s 1.06
cache / DefOpt / cpu / PostRev 0.00002216581999391565 s 0.00001688983998974436 s 1.31
cache / DefOpt / cpu / BothRev 0.000016958400003659335 s 0.000016799259938125033 s 1.01
cache / IDefOpt / cpu / PreRev 0.000016649199992571084 s 0.000017757959994924022 s 0.94
cache / IDefOpt / cpu / PostRev 0.000016108600002553432 s 0.00001715016000161995 s 0.94
cache / IDefOpt / cpu / BothRev 0.000015865459997712604 s 0.0000166604799869674 s 0.95
cache / JaXPipe / cuda / Primal 0.000002304 s 0.000002304 s 1
cache / Jax / cuda / Primal 0.000002272 s 0.000002272 s 1
cache / HLOOpt / cuda / Primal 0.000002272 s 0.000002272 s 1
cache / PartOpt / cuda / Primal 0.00000224 s 0.000002272 s 0.99
cache / IPartOpt / cuda / Primal 0.00000224 s 0.000002272 s 0.99
cache / DefOpt / cuda / Primal 0.00000224 s 0.00000224 s 1
cache / IDefOpt / cuda / Primal 0.000002304 s 0.000002304 s 1
cache / JaXPipe / cuda / Forward 0.000002336 s 0.000002335 s 1.00
cache / Jax / cuda / Forward 0.000002272 s 0.000002272 s 1
cache / HLOOpt / cuda / Forward 0.000002335 s 0.000002335 s 1
cache / PartOpt / cuda / Forward 0.000002303 s 0.000002304 s 1.00
cache / IPartOpt / cuda / Forward 0.000002303 s 0.000002304 s 1.00
cache / DefOpt / cuda / Forward 0.00000224 s 0.00000224 s 1
cache / IDefOpt / cuda / Forward 0.000002303 s 0.000002272 s 1.01
cache / JaXPipe / cuda / PreRev 0.000013184 s 0.000013184 s 1
cache / JaXPipe / cuda / PostRev 0.000011616 s 0.00001136 s 1.02
cache / JaXPipe / cuda / BothRev 0.000013215 s 0.000013216 s 1.00
cache / Jax / cuda / BothRev 0.0000112 s 0.000011488 s 0.97
cache / HLOOpt / cuda / PreRev 0.000013216 s 0.000013215 s 1.00
cache / HLOOpt / cuda / PostRev 0.000013216 s 0.000013183 s 1.00
cache / HLOOpt / cuda / BothRev 0.000013216 s 0.000013216 s 1
cache / PartOpt / cuda / PreRev 0.000013248 s 0.000013216 s 1.00
cache / PartOpt / cuda / PostRev 0.000011872 s 0.000011488 s 1.03
cache / PartOpt / cuda / BothRev 0.000013215 s 0.000013216 s 1.00
cache / IPartOpt / cuda / PreRev 0.000013215 s 0.000013216 s 1.00
cache / IPartOpt / cuda / PostRev 0.000011488 s 0.00001184 s 0.97
cache / IPartOpt / cuda / BothRev 0.000013215 s 0.000013184 s 1.00
cache / DefOpt / cuda / PreRev 0.000013248 s 0.000013216 s 1.00
cache / DefOpt / cuda / PostRev 0.000013151 s 0.000013152 s 1.00
cache / DefOpt / cuda / BothRev 0.000013248 s 0.000013248 s 1
cache / IDefOpt / cuda / PreRev 0.000013216 s 0.000013183 s 1.00
cache / IDefOpt / cuda / PostRev 0.000013184 s 0.000013152 s 1.00
cache / IDefOpt / cuda / BothRev 0.000013216 s 0.000013215 s 1.00
cache / JaXPipe / tpu / Primal 0.00000245955 s 0.000002456125 s 1.00
cache / Jax / tpu / Primal 0.00000245175 s 0.000002448675 s 1.00
cache / HLOOpt / tpu / Primal 0.000002459325 s 0.0000024729 s 0.99
cache / PartOpt / tpu / Primal 0.000002456225 s 0.0000024683 s 1.00
cache / IPartOpt / tpu / Primal 0.0000024477 s 0.00000246125 s 0.99
cache / DefOpt / tpu / Primal 0.000002479225 s 0.00000244685 s 1.01
cache / IDefOpt / tpu / Primal 0.0000024449 s 0.000002470775 s 0.99
cache / JaXPipe / tpu / Forward 0.0000035318 s 0.0000035407250000000003 s 1.00
cache / Jax / tpu / Forward 0.0000035532750000000003 s 0.00000353345 s 1.01
cache / HLOOpt / tpu / Forward 0.0000035856 s 0.0000035537000000000004 s 1.01
cache / PartOpt / tpu / Forward 0.00000355915 s 0.0000035360750000000003 s 1.01
cache / IPartOpt / tpu / Forward 0.00000356035 s 0.0000035455750000000004 s 1.00
cache / DefOpt / tpu / Forward 0.0000035548000000000003 s 0.00000352575 s 1.01
cache / IDefOpt / tpu / Forward 0.0000035819000000000004 s 0.0000035553 s 1.01
cache / JaXPipe / tpu / PreRev 0.0000039224 s 0.000003930274999999999 s 1.00
cache / JaXPipe / tpu / PostRev 0.000004900875 s 0.00000496115 s 0.99
cache / JaXPipe / tpu / BothRev 0.000003972925 s 0.00000393385 s 1.01
cache / Jax / tpu / BothRev 0.000004954275 s 0.000004990125000000001 s 0.99
cache / HLOOpt / tpu / PreRev 0.000003975025 s 0.000003932450000000001 s 1.01
cache / HLOOpt / tpu / PostRev 0.000004081975 s 0.0000041239 s 0.99
cache / HLOOpt / tpu / BothRev 0.000003976225 s 0.0000039407 s 1.01
cache / PartOpt / tpu / PreRev 0.0000041101 s 0.000004137224999999999 s 0.99
cache / PartOpt / tpu / PostRev 0.0000049157 s 0.000004995249999999999 s 0.98
cache / PartOpt / tpu / BothRev 0.00000408545 s 0.0000041481 s 0.98
cache / IPartOpt / tpu / PreRev 0.00000397885 s 0.00000394215 s 1.01
cache / IPartOpt / tpu / PostRev 0.000004943449999999999 s 0.00000498895 s 0.99
cache / IPartOpt / tpu / BothRev 0.0000039803 s 0.000003957575 s 1.01
cache / DefOpt / tpu / PreRev 0.00000410195 s 0.000004119575 s 1.00
cache / DefOpt / tpu / PostRev 0.000003984275 s 0.000003933349999999999 s 1.01
cache / DefOpt / tpu / BothRev 0.0000040772000000000005 s 0.0000041150500000000005 s 0.99
cache / IDefOpt / tpu / PreRev 0.0000039691 s 0.000003921825 s 1.01
cache / IDefOpt / tpu / PostRev 0.000004084225000000001 s 0.0000041251 s 0.99
cache / IDefOpt / tpu / BothRev 0.000003978174999999999 s 0.0000039568 s 1.01
cache / JaXPipe / cpu / Primal 0.000012655 s 0.000006956840015845956 s 1.82
cache / Jax / cpu / Primal 0.000012979 s 0.000007311219933399116 s 1.78
cache / HLOOpt / cpu / Primal 0.000013026 s 0.000007593460013595177 s 1.72
cache / PartOpt / cpu / Primal 0.000013354 s 0.000007535879994975403 s 1.77
cache / IPartOpt / cpu / Primal 0.000012584 s 0.000007371540013991762 s 1.71
cache / DefOpt / cpu / Primal 0.000012949 s 0.000007683940011702362 s 1.69
cache / IDefOpt / cpu / Primal 0.00001242 s 0.0000073235999934695425 s 1.70
cache / JaXPipe / cpu / Forward 0.000018539 s 0.000014550459982274333 s 1.27
cache / Jax / cpu / Forward 0.00001799 s 0.000014931959995010404 s 1.20
cache / HLOOpt / cpu / Forward 0.000018149 s 0.00001945838004758116 s 0.93
cache / PartOpt / cpu / Forward 0.000017330000000000002 s 0.00001901271997667209 s 0.91
cache / IPartOpt / cpu / Forward 0.000017509 s 0.0000151107200235856 s 1.16
cache / DefOpt / cpu / Forward 0.000017406999999999998 s 0.00001980758002900984 s 0.88
cache / IDefOpt / cpu / Forward 0.000017398000000000002 s 0.000014846979993308196 s 1.17
cache / JaXPipe / cpu / PreRev 0.000018572 s 0.000015902700033620933 s 1.17
cache / JaXPipe / cpu / PostRev 0.000019719 s 0.000021121619975019717 s 0.93
cache / JaXPipe / cpu / BothRev 0.000017916999999999998 s 0.000017116039998654742 s 1.05
cache / Jax / cpu / BothRev 0.000031298 s 0.00002173714004129579 s 1.44
cache / HLOOpt / cpu / PreRev 0.000018112 s 0.000017123339985118948 s 1.06
cache / HLOOpt / cpu / PostRev 0.000017624 s 0.00001864448002379504 s 0.95
cache / HLOOpt / cpu / BothRev 0.000035535 s 0.000018996239978150696 s 1.87
cache / PartOpt / cpu / PreRev 0.00003759 s 0.00001642888004425913 s 2.29
cache / PartOpt / cpu / PostRev 0.00002901 s 0.000020945439973729663 s 1.39
cache / PartOpt / cpu / BothRev 0.000027251 s 0.000016367300049751065 s 1.66
cache / IPartOpt / cpu / PreRev 0.000024226 s 0.000017679780003163616 s 1.37
cache / IPartOpt / cpu / PostRev 0.00002642 s 0.000022093920060797247 s 1.20
cache / IPartOpt / cpu / BothRev 0.000024396 s 0.000016750399972806917 s 1.46
cache / DefOpt / cpu / PreRev 0.000024562 s 0.000016780159985501086 s 1.46
cache / DefOpt / cpu / PostRev 0.000019534 s 0.00001688983998974436 s 1.16
cache / DefOpt / cpu / BothRev 0.000022103 s 0.000016799259938125033 s 1.32
cache / IDefOpt / cpu / PreRev 0.000024436 s 0.000017757959994924022 s 1.38
cache / IDefOpt / cpu / PostRev 0.00002701 s 0.00001715016000161995 s 1.57
cache / IDefOpt / cpu / BothRev 0.000024496 s 0.0000166604799869674 s 1.47
Concat / JaXPipe / cpu / Primal 0.000007837380001092243 s 0.00000811730000350508 s 0.97
Concat / Jax / cpu / Primal 0.00000711994000539562 s 0.000007682900013605832 s 0.93
Concat / HLOOpt / cpu / Primal 0.00001024896000217268 s 0.00001060113999301393 s 0.97
Concat / PartOpt / cpu / Primal 0.000006625159996929142 s 0.000007756900004096678 s 0.85
Concat / IPartOpt / cpu / Primal 0.000006695799997942231 s 0.000007213979988591745 s 0.93
Concat / DefOpt / cpu / Primal 0.0000106093799968221 s 0.000011252420026721666 s 0.94
Concat / IDefOpt / cpu / Primal 0.00000657244000421997 s 0.000007434619956256938 s 0.88
Concat / JaXPipe / cpu / Forward 0.000011313520003568556 s 0.00001102167998396908 s 1.03
Concat / Jax / cpu / Forward 0.000010954759991363972 s 0.000011529920011525971 s 0.95
Concat / HLOOpt / cpu / Forward 0.000014874840005631996 s 0.000015114119942154502 s 0.98
Concat / PartOpt / cpu / Forward 0.0000151896799889073 s 0.000015529919983237052 s 0.98
Concat / IPartOpt / cpu / Forward 0.00001100988000189318 s 0.00001101205997656507 s 1.00
Concat / DefOpt / cpu / Forward 0.000015059980000842188 s 0.00001623475995984336 s 0.93
Concat / IDefOpt / cpu / Forward 0.000010687780002172075 s 0.000011439519985287916 s 0.93
Concat / JaXPipe / cpu / PreRev 0.000011720900001819244 s 0.00001323297995440953 s 0.89
Concat / JaXPipe / cpu / PostRev 0.000012682240005688072 s 0.000013689120060007552 s 0.93
Concat / JaXPipe / cpu / BothRev 0.00001273031999289742 s 0.00001329070005340327 s 0.96
Concat / Jax / cpu / BothRev 0.000012601979994997236 s 0.000013045700006841798 s 0.97
Concat / HLOOpt / cpu / PreRev 0.00001237235999724362 s 0.000013706500003536347 s 0.90
Concat / HLOOpt / cpu / PostRev 0.000016496280006776943 s 0.000017045479953594622 s 0.97
Concat / HLOOpt / cpu / BothRev 0.00001413272000490906 s 0.00001466194001295662 s 0.96
Concat / PartOpt / cpu / PreRev 0.000012438120002116192 s 0.000013567740024882369 s 0.92
Concat / PartOpt / cpu / PostRev 0.00001232705998518213 s 0.000013383999985308037 s 0.92
Concat / PartOpt / cpu / BothRev 0.000012387320002744674 s 0.000013322119930307964 s 0.93
Concat / IPartOpt / cpu / PreRev 0.000017036279996318626 s 0.000014049620031073571 s 1.21
Concat / IPartOpt / cpu / PostRev 0.000012929979993714369 s 0.000013057799988018814 s 0.99
Concat / IPartOpt / cpu / BothRev 0.00001213414000631019 s 0.00001312568005232606 s 0.92
Concat / DefOpt / cpu / PreRev 0.000011710739997852216 s 0.000012955279962625355 s 0.90
Concat / DefOpt / cpu / PostRev 0.00001182920000019294 s 0.000013988820019221748 s 0.85
Concat / DefOpt / cpu / BothRev 0.000012492359996940647 s 0.000013455380039886224 s 0.93
Concat / IDefOpt / cpu / PreRev 0.000012313399995491636 s 0.000013155779970475124 s 0.94
Concat / IDefOpt / cpu / PostRev 0.000012558000007629743 s 0.000012839719993280596 s 0.98
Concat / IDefOpt / cpu / BothRev 0.000011483460014005686 s 0.000014180980006130994 s 0.81
Concat / JaXPipe / cuda / Primal 0.000001951 s 0.000001951 s 1
Concat / Jax / cuda / Primal 0.000001951 s 0.000001951 s 1
Concat / HLOOpt / cuda / Primal 0.000001951 s 0.000001952 s 1.00
Concat / PartOpt / cuda / Primal 0.000001951 s 0.000001951 s 1
Concat / IPartOpt / cuda / Primal 0.000001951 s 0.000001951 s 1
Concat / DefOpt / cuda / Primal 0.000001951 s 0.000001951 s 1
Concat / IDefOpt / cuda / Primal 0.000001951 s 0.000001951 s 1
Concat / JaXPipe / cuda / Forward 0.000010208 s 0.00001072 s 0.95
Concat / Jax / cuda / Forward 0.000010433 s 0.000010624 s 0.98
Concat / HLOOpt / cuda / Forward 0.000010752 s 0.000010432 s 1.03
Concat / PartOpt / cuda / Forward 0.000010208 s 0.000010591 s 0.96
Concat / IPartOpt / cuda / Forward 0.000010112 s 0.000010432 s 0.97
Concat / DefOpt / cuda / Forward 0.000010176 s 0.0000112 s 0.91
Concat / IDefOpt / cuda / Forward 0.00001008 s 0.000010112 s 1.00
Concat / JaXPipe / cuda / PreRev 0.000016448000000000002 s 0.000016352 s 1.01
Concat / JaXPipe / cuda / PostRev 0.000016704 s 0.00001696 s 0.98
Concat / JaXPipe / cuda / BothRev 0.000014977 s 0.000016768000000000003 s 0.89
Concat / Jax / cuda / BothRev 0.000016416 s 0.000016736 s 0.98
Concat / HLOOpt / cuda / PreRev 0.000016736 s 0.000017184 s 0.97
Concat / HLOOpt / cuda / PostRev 0.00001616 s 0.000017056 s 0.95
Concat / HLOOpt / cuda / BothRev 0.000016385 s 0.000017184 s 0.95
Concat / PartOpt / cuda / PreRev 0.000016608 s 0.000017726999999999998 s 0.94
Concat / PartOpt / cuda / PostRev 0.00001664 s 0.000016768000000000003 s 0.99
Concat / PartOpt / cuda / BothRev 0.00001648 s 0.000017056 s 0.97
Concat / IPartOpt / cuda / PreRev 0.00001696 s 0.000017088 s 0.99
Concat / IPartOpt / cuda / PostRev 0.000016864 s 0.000017153 s 0.98
Concat / IPartOpt / cuda / BothRev 0.000016672 s 0.000017024 s 0.98
Concat / DefOpt / cuda / PreRev 0.000017055000000000002 s 0.000016704 s 1.02
Concat / DefOpt / cuda / PostRev 0.000016672 s 0.000016448000000000002 s 1.01
Concat / DefOpt / cuda / BothRev 0.000016831 s 0.000016704 s 1.01
Concat / IDefOpt / cuda / PreRev 0.000016864 s 0.000019008 s 0.89
Concat / IDefOpt / cuda / PostRev 0.0000168 s 0.000017056 s 0.98
Concat / IDefOpt / cuda / BothRev 0.000016991 s 0.000016352 s 1.04
Concat / JaXPipe / tpu / Primal 0.000001479775 s 0.0000015218999999999998 s 0.97
Concat / Jax / tpu / Primal 0.00000149715 s 0.000001528325 s 0.98
Concat / HLOOpt / tpu / Primal 0.000001477775 s 0.000001530275 s 0.97
Concat / PartOpt / tpu / Primal 0.0000014964 s 0.000001534625 s 0.98
Concat / IPartOpt / tpu / Primal 0.0000014830750000000002 s 0.00000152225 s 0.97
Concat / DefOpt / tpu / Primal 0.000001485525 s 0.000001525125 s 0.97
Concat / IDefOpt / tpu / Primal 0.000001478 s 0.000001522775 s 0.97
Concat / JaXPipe / tpu / Forward 0.00000151395 s 0.000001564575 s 0.97
Concat / Jax / tpu / Forward 0.0000015247 s 0.000001564375 s 0.97
Concat / HLOOpt / tpu / Forward 0.00000151305 s 0.000001579 s 0.96
Concat / PartOpt / tpu / Forward 0.00000152975 s 0.000001571225 s 0.97
Concat / IPartOpt / tpu / Forward 0.0000015135750000000002 s 0.000001578375 s 0.96
Concat / DefOpt / tpu / Forward 0.000001501675 s 0.0000015573749999999998 s 0.96
Concat / IDefOpt / tpu / Forward 0.000001515525 s 0.00000157315 s 0.96
Concat / JaXPipe / tpu / PreRev 0.000001999225 s 0.000002009 s 1.00
Concat / JaXPipe / tpu / PostRev 0.000002028625 s 0.00000208995 s 0.97
Concat / JaXPipe / tpu / BothRev 0.00000197715 s 0.000001990725 s 0.99
Concat / Jax / tpu / BothRev 0.0000020058000000000003 s 0.00000207115 s 0.97
Concat / HLOOpt / tpu / PreRev 0.000001950575 s 0.000001998025 s 0.98
Concat / HLOOpt / tpu / PostRev 0.000002002075 s 0.000002079825 s 0.96
Concat / HLOOpt / tpu / BothRev 0.00000195435 s 0.000001992025 s 0.98
Concat / PartOpt / tpu / PreRev 0.0000020054 s 0.000002079425 s 0.96
Concat / PartOpt / tpu / PostRev 0.00000194965 s 0.0000020019500000000004 s 0.97
Concat / PartOpt / tpu / BothRev 0.0000020062 s 0.000002076725 s 0.97
Concat / IPartOpt / tpu / PreRev 0.0000019557 s 0.00000199835 s 0.98
Concat / IPartOpt / tpu / PostRev 0.0000020022 s 0.000002073975 s 0.97
Concat / IPartOpt / tpu / BothRev 0.0000019602 s 0.00000199655 s 0.98
Concat / DefOpt / tpu / PreRev 0.00000201435 s 0.000002067625 s 0.97
Concat / DefOpt / tpu / PostRev 0.0000019586 s 0.00000199465 s 0.98
Concat / DefOpt / tpu / BothRev 0.00000200585 s 0.000002069275 s 0.97
Concat / IDefOpt / tpu / PreRev 0.00000195665 s 0.00000200825 s 0.97
Concat / IDefOpt / tpu / PostRev 0.000002001275 s 0.000002074575 s 0.96
Concat / IDefOpt / tpu / BothRev 0.00000196175 s 0.000002000675 s 0.98
Concat / JaXPipe / cpu / Primal 0.000012889 s 0.00000811730000350508 s 1.59
Concat / Jax / cpu / Primal 0.000013204 s 0.000007682900013605832 s 1.72
Concat / HLOOpt / cpu / Primal 0.000012756 s 0.00001060113999301393 s 1.20
Concat / PartOpt / cpu / Primal 0.000012848 s 0.000007756900004096678 s 1.66
Concat / IPartOpt / cpu / Primal 0.000012849 s 0.000007213979988591745 s 1.78
Concat / DefOpt / cpu / Primal 0.000013003 s 0.000011252420026721666 s 1.16
Concat / IDefOpt / cpu / Primal 0.00001262 s 0.000007434619956256938 s 1.70
Concat / JaXPipe / cpu / Forward 0.000017874000000000002 s 0.00001102167998396908 s 1.62
Concat / Jax / cpu / Forward 0.000017447 s 0.000011529920011525971 s 1.51
Concat / HLOOpt / cpu / Forward 0.000017823 s 0.000015114119942154502 s 1.18
Concat / PartOpt / cpu / Forward 0.000017641 s 0.000015529919983237052 s 1.14
Concat / IPartOpt / cpu / Forward 0.000018216 s 0.00001101205997656507 s 1.65
Concat / DefOpt / cpu / Forward 0.000017828 s 0.00001623475995984336 s 1.10
Concat / IDefOpt / cpu / Forward 0.000018044 s 0.000011439519985287916 s 1.58
Concat / JaXPipe / cpu / PreRev 0.000020415 s 0.00001323297995440953 s 1.54
Concat / JaXPipe / cpu / PostRev 0.000019953 s 0.000013689120060007552 s 1.46
Concat / JaXPipe / cpu / BothRev 0.000019657 s 0.00001329070005340327 s 1.48
Concat / Jax / cpu / BothRev 0.000020352 s 0.000013045700006841798 s 1.56
Concat / HLOOpt / cpu / PreRev 0.00001984 s 0.000013706500003536347 s 1.45
Concat / HLOOpt / cpu / PostRev 0.000020157 s 0.000017045479953594622 s 1.18
Concat / HLOOpt / cpu / BothRev 0.000019415000000000003 s 0.00001466194001295662 s 1.32
Concat / PartOpt / cpu / PreRev 0.000019998 s 0.000013567740024882369 s 1.47
Concat / PartOpt / cpu / PostRev 0.0000207 s 0.000013383999985308037 s 1.55
Concat / PartOpt / cpu / BothRev 0.000021109 s 0.000013322119930307964 s 1.58
Concat / IPartOpt / cpu / PreRev 0.000020448 s 0.000014049620031073571 s 1.46
Concat / IPartOpt / cpu / PostRev 0.000021175 s 0.000013057799988018814 s 1.62
Concat / IPartOpt / cpu / BothRev 0.000021279 s 0.00001312568005232606 s 1.62
Concat / DefOpt / cpu / PreRev 0.000020186 s 0.000012955279962625355 s 1.56
Concat / DefOpt / cpu / PostRev 0.00002026 s 0.000013988820019221748 s 1.45
Concat / DefOpt / cpu / BothRev 0.000019948 s 0.000013455380039886224 s 1.48
Concat / IDefOpt / cpu / PreRev 0.000019963 s 0.000013155779970475124 s 1.52
Concat / IDefOpt / cpu / PostRev 0.000021359 s 0.000012839719993280596 s 1.66
Concat / IDefOpt / cpu / BothRev 0.000021194 s 0.000014180980006130994 s 1.49
const_scatter / JaXPipe / cpu / Primal 0.000007045879999623139 s 0.000007684319989493815 s 0.92
const_scatter / Jax / cpu / Primal 0.000006954280013360403 s 0.000007795560031809146 s 0.89
const_scatter / HLOOpt / cpu / Primal 0.00000660258000380054 s 0.000008156360036082332 s 0.81
const_scatter / PartOpt / cpu / Primal 0.000007046820003324683 s 0.000007077780019244528 s 1.00
const_scatter / IPartOpt / cpu / Primal 0.000007281340010649729 s 0.000007127839962777216 s 1.02
const_scatter / DefOpt / cpu / Primal 0.000007038079993435531 s 0.00001182172003609594 s 0.60
const_scatter / IDefOpt / cpu / Primal 0.000006507680000140681 s 0.000007085820006977883 s 0.92
const_scatter / JaXPipe / cpu / Forward 0.00001025107998657404 s 0.00001117912002882804 s 0.92
const_scatter / Jax / cpu / Forward 0.000010728259994721156 s 0.000012405520028551109 s 0.86
const_scatter / HLOOpt / cpu / Forward 0.000014583319998564548 s 0.000016166460000022198 s 0.90
const_scatter / PartOpt / cpu / Forward 0.0000151262999975188 s 0.000015849359979256405 s 0.95
const_scatter / IPartOpt / cpu / Forward 0.000010512740007015963 s 0.000010997060026056716 s 0.96
const_scatter / DefOpt / cpu / Forward 0.000015138839999053745 s 0.00001573457998347294 s 0.96
const_scatter / IDefOpt / cpu / Forward 0.000009518999988813447 s 0.000010618399992381455 s 0.90
const_scatter / JaXPipe / cpu / PreRev 0.0003046743200138 s 0.0003040449800209 s 1.00
const_scatter / JaXPipe / cpu / PostRev 0.0002929268200091 s 0.0002915093400042 s 1.00
const_scatter / JaXPipe / cpu / BothRev 0.0002854891600031 s 0.0002868209999996 s 1.00
const_scatter / Jax / cpu / BothRev 0.0002860530200018 s 0.0002858066800308 s 1.00
const_scatter / HLOOpt / cpu / PreRev 0.0002875375000098 s 0.0002873173999796 s 1.00
const_scatter / HLOOpt / cpu / PostRev 0.0002893032800034 s 0.0002849928999967 s 1.02
const_scatter / HLOOpt / cpu / BothRev 0.0002882812199959 s 0.0002872712200041 s 1.00
const_scatter / PartOpt / cpu / PreRev 0.0002913089600042 s 0.0002921233799952 s 1.00
const_scatter / PartOpt / cpu / PostRev 0.0002861352599938 s 0.000287245599984 s 1.00
const_scatter / PartOpt / cpu / BothRev 0.0002918820600007 s 0.0002904328599561 s 1.00
const_scatter / IPartOpt / cpu / PreRev 0.0002940502000024 s 0.000291726020041 s 1.01
const_scatter / IPartOpt / cpu / PostRev 0.0002869377000024 s 0.0002852017199893 s 1.01
const_scatter / IPartOpt / cpu / BothRev 0.0002926235800146 s 0.000291159999997 s 1.01
const_scatter / DefOpt / cpu / PreRev 0.0002915703600024 s 0.0002904347400271 s 1.00
const_scatter / DefOpt / cpu / PostRev 0.0002863039400062 s 0.0002867657000024 s 1.00
const_scatter / DefOpt / cpu / BothRev 0.0002901643400059 s 0.0002920737600197 s 0.99
const_scatter / IDefOpt / cpu / PreRev 0.0002911594799957 s 0.0002910701800101 s 1.00
const_scatter / IDefOpt / cpu / PostRev 0.0002891495400081 s 0.000286926560002 s 1.01
const_scatter / IDefOpt / cpu / BothRev 0.0002927561200021 s 0.0003165261199501 s 0.92
const_scatter / JaXPipe / cuda / Primal 0.0000019200000000000003 s 0.0000019200000000000003 s 1
const_scatter / Jax / cuda / Primal 0.0000019200000000000003 s 0.0000019200000000000003 s 1
const_scatter / HLOOpt / cuda / Primal 0.0000019200000000000003 s 0.0000019200000000000003 s 1
const_scatter / PartOpt / cuda / Primal 0.000001919 s 0.0000019200000000000003 s 1.00
const_scatter / IPartOpt / cuda / Primal 0.000001919 s 0.000001887 s 1.02
const_scatter / DefOpt / cuda / Primal 0.0000019200000000000003 s 0.0000019200000000000003 s 1
const_scatter / IDefOpt / cuda / Primal 0.0000019200000000000003 s 0.000001919 s 1.00
const_scatter / JaXPipe / cuda / Forward 0.000009824 s 0.000011103 s 0.88
const_scatter / Jax / cuda / Forward 0.000009888 s 0.000009856 s 1.00
const_scatter / HLOOpt / cuda / Forward 0.000009472 s 0.00001008 s 0.94
const_scatter / PartOpt / cuda / Forward 0.000009952 s 0.00001008 s 0.99
const_scatter / IPartOpt / cuda / Forward 0.000009952 s 0.000009472 s 1.05
const_scatter / DefOpt / cuda / Forward 0.000009856 s 0.000010176 s 0.97
const_scatter / IDefOpt / cuda / Forward 0.000009695 s 0.000010015 s 0.97
const_scatter / JaXPipe / cuda / PreRev 0.000012736 s 0.000012863 s 0.99
const_scatter / JaXPipe / cuda / PostRev 0.000016670999999999997 s 0.000016927999999999998 s 0.98
const_scatter / JaXPipe / cuda / BothRev 0.000017760000000000003 s 0.00001296 s 1.37
const_scatter / Jax / cuda / BothRev 0.000016672 s 0.000016896000000000002 s 0.99
const_scatter / HLOOpt / cuda / PreRev 0.000012513 s 0.0000128 s 0.98
const_scatter / HLOOpt / cuda / PostRev 0.00001328 s 0.000013152 s 1.01
const_scatter / HLOOpt / cuda / BothRev 0.000012544 s 0.000012832 s 0.98
const_scatter / PartOpt / cuda / PreRev 0.000012864 s 0.000013056 s 0.99
const_scatter / PartOpt / cuda / PostRev 0.000016255999999999998 s 0.000017536 s 0.93
const_scatter / PartOpt / cuda / BothRev 0.000012416 s 0.0000128 s 0.97
const_scatter / IPartOpt / cuda / PreRev 0.000012865 s 0.000012704 s 1.01
const_scatter / IPartOpt / cuda / PostRev 0.000016512 s 0.000016929 s 0.98
const_scatter / IPartOpt / cuda / BothRev 0.000012768 s 0.000012736 s 1.00
const_scatter / DefOpt / cuda / PreRev 0.000013472 s 0.00001312 s 1.03
const_scatter / DefOpt / cuda / PostRev 0.000012768 s 0.00001296 s 0.99
const_scatter / DefOpt / cuda / BothRev 0.000012896 s 0.000012832 s 1.00
const_scatter / IDefOpt / cuda / PreRev 0.000013024 s 0.000012832 s 1.01
const_scatter / IDefOpt / cuda / PostRev 0.000012288 s 0.000013408 s 0.92
const_scatter / IDefOpt / cuda / BothRev 0.0000128 s 0.00001312 s 0.98
const_scatter / JaXPipe / tpu / Primal 0.000003806775 s 0.00000382305 s 1.00
const_scatter / Jax / tpu / Primal 0.0000038291750000000005 s 0.000003831675 s 1.00
const_scatter / HLOOpt / tpu / Primal 9.30625e-7 s 9.5125e-7 s 0.98
const_scatter / PartOpt / tpu / Primal 0.000003834125 s 0.000003822175 s 1.00
const_scatter / IPartOpt / tpu / Primal 0.000003829475 s 0.00000379315 s 1.01
const_scatter / DefOpt / tpu / Primal 9.71025e-7 s 9.58825e-7 s 1.01
const_scatter / IDefOpt / tpu / Primal 9.30375e-7 s 9.43525e-7 s 0.99
const_scatter / JaXPipe / tpu / Forward 0.00000192845 s 0.00000192225 s 1.00
const_scatter / Jax / tpu / Forward 0.00000648755 s 0.000006482375 s 1.00
const_scatter / HLOOpt / tpu / Forward 0.0000019194 s 0.00000191365 s 1.00
const_scatter / PartOpt / tpu / Forward 0.000001911225 s 0.000001931925 s 0.99
const_scatter / IPartOpt / tpu / Forward 0.000001921225 s 0.000001927125 s 1.00
const_scatter / DefOpt / tpu / Forward 0.0000019204 s 0.000001929975 s 1.00
const_scatter / IDefOpt / tpu / Forward 0.00000190595 s 0.00000191915 s 0.99
const_scatter / JaXPipe / tpu / PreRev 0.000004299725 s 0.000004329525 s 0.99
const_scatter / JaXPipe / tpu / PostRev 0.000006683200000000001 s 0.000006649024999999999 s 1.01
const_scatter / JaXPipe / tpu / BothRev 0.0000042966000000000005 s 0.000004307675 s 1.00
const_scatter / Jax / tpu / BothRev 0.0000066769750000000006 s 0.0000066769 s 1.00
const_scatter / HLOOpt / tpu / PreRev 0.000004304425 s 0.000004301700000000001 s 1.00
const_scatter / HLOOpt / tpu / PostRev 0.000004307125 s 0.000004335525 s 0.99
const_scatter / HLOOpt / tpu / BothRev 0.0000042995 s 0.0000043162 s 1.00
const_scatter / PartOpt / tpu / PreRev 0.000004320025 s 0.000004306499999999999 s 1.00
const_scatter / PartOpt / tpu / PostRev 0.000006662924999999999 s 0.000006660475 s 1.00
const_scatter / PartOpt / tpu / BothRev 0.000004295625 s 0.00000432115 s 0.99
const_scatter / IPartOpt / tpu / PreRev 0.0000043193 s 0.000004300775 s 1.00
const_scatter / IPartOpt / tpu / PostRev 0.0000066737 s 0.0000066653 s 1.00
const_scatter / IPartOpt / tpu / BothRev 0.000004296725000000001 s 0.00000431405 s 1.00
const_scatter / DefOpt / tpu / PreRev 0.000004318 s 0.0000043047 s 1.00
const_scatter / DefOpt / tpu / PostRev 0.00000430735 s 0.00000429905 s 1.00
const_scatter / DefOpt / tpu / BothRev 0.0000043085 s 0.000004318675 s 1.00
const_scatter / IDefOpt / tpu / PreRev 0.000004308975 s 0.0000043148 s 1.00
const_scatter / IDefOpt / tpu / PostRev 0.00000433655 s 0.0000043003 s 1.01
const_scatter / IDefOpt / tpu / BothRev 0.0000043062500000000005 s 0.0000043025 s 1.00
const_scatter / JaXPipe / cpu / Primal 0.000012664 s 0.000007684319989493815 s 1.65
const_scatter / Jax / cpu / Primal 0.000013181999999999998 s 0.000007795560031809146 s 1.69
const_scatter / HLOOpt / cpu / Primal 0.000013185 s 0.000008156360036082332 s 1.62
const_scatter / PartOpt / cpu / Primal 0.000012721 s 0.000007077780019244528 s 1.80
const_scatter / IPartOpt / cpu / Primal 0.000012458 s 0.000007127839962777216 s 1.75
const_scatter / DefOpt / cpu / Primal 0.000012613 s 0.00001182172003609594 s 1.07
const_scatter / IDefOpt / cpu / Primal 0.000012549 s 0.000007085820006977883 s 1.77
const_scatter / JaXPipe / cpu / Forward 0.000017135 s 0.00001117912002882804 s 1.53
const_scatter / Jax / cpu / Forward 0.00001685 s 0.000012405520028551109 s 1.36
const_scatter / HLOOpt / cpu / Forward 0.000016930000000000002 s 0.000016166460000022198 s 1.05
const_scatter / PartOpt / cpu / Forward 0.000016667 s 0.000015849359979256405 s 1.05
const_scatter / IPartOpt / cpu / Forward 0.000017159 s 0.000010997060026056716 s 1.56
const_scatter / DefOpt / cpu / Forward 0.000017054 s 0.00001573457998347294 s 1.08
const_scatter / IDefOpt / cpu / Forward 0.000017233999999999998 s 0.000010618399992381455 s 1.62
const_scatter / JaXPipe / cpu / PreRev 0.00051675 s 0.0003040449800209 s 1.70
const_scatter / JaXPipe / cpu / PostRev 0.000530804 s 0.0002915093400042 s 1.82
const_scatter / JaXPipe / cpu / BothRev 0.00052207 s 0.0002868209999996 s 1.82
const_scatter / Jax / cpu / BothRev 0.000520012 s 0.0002858066800308 s 1.82
const_scatter / HLOOpt / cpu / PreRev 0.000502599 s 0.0002873173999796 s 1.75
const_scatter / HLOOpt / cpu / PostRev 0.000531446 s 0.0002849928999967 s 1.86
const_scatter / HLOOpt / cpu / BothRev 0.000509983 s 0.0002872712200041 s 1.78
const_scatter / PartOpt / cpu / PreRev 0.0005188739999999 s 0.0002921233799952 s 1.78
const_scatter / PartOpt / cpu / PostRev 0.000525949 s 0.000287245599984 s 1.83
const_scatter / PartOpt / cpu / BothRev 0.0005083069999999 s 0.0002904328599561 s 1.75
const_scatter / IPartOpt / cpu / PreRev 0.000507237 s 0.000291726020041 s 1.74
const_scatter / IPartOpt / cpu / PostRev 0.000522571 s 0.0002852017199893 s 1.83
const_scatter / IPartOpt / cpu / BothRev 0.000520849 s 0.000291159999997 s 1.79
const_scatter / DefOpt / cpu / PreRev 0.0005385149999999 s 0.0002904347400271 s 1.85
const_scatter / DefOpt / cpu / PostRev 0.0005356779999999 s 0.0002867657000024 s 1.87
const_scatter / DefOpt / cpu / BothRev 0.000520829 s 0.0002920737600197 s 1.78
const_scatter / IDefOpt / cpu / PreRev 0.000524497 s 0.0002910701800101 s 1.80
const_scatter / IDefOpt / cpu / PostRev 0.000536962 s 0.000286926560002 s 1.87
const_scatter / IDefOpt / cpu / BothRev 0.000523844 s 0.0003165261199501 s 1.65
GenDot / JaXPipe / cpu / Primal 0.000007764539996060193 s 0.000008797720001894049 s 0.88
GenDot / Jax / cpu / Primal 0.000006740919995991135 s 0.000008127900009640144 s 0.83
GenDot / HLOOpt / cpu / Primal 0.000012684460000400576 s 0.000012815599984605796 s 0.99
GenDot / PartOpt / cpu / Primal 0.000007511799999520008 s 0.000008449520009889966 s 0.89
GenDot / IPartOpt / cpu / Primal 0.000007805919999555044 s 0.000008597039959568065 s 0.91
GenDot / DefOpt / cpu / Primal 0.000007267240000601305 s 0.000009416619950570747 s 0.77
GenDot / IDefOpt / cpu / Primal 0.00000814504001482419 s 0.000009196099981636509 s 0.89
GenDot / JaXPipe / cpu / Forward 0.000011676700005409657 s 0.000012595799989867374 s 0.93
GenDot / Jax / cpu / Forward 0.000011092660001850162 s 0.000011937699982809135 s 0.93
GenDot / HLOOpt / cpu / Forward 0.000016936820004502805 s 0.000016989840050882777 s 1.00
GenDot / PartOpt / cpu / Forward 0.00001217976000816634 s 0.0000176848600312951 s 0.69
GenDot / IPartOpt / cpu / Forward 0.00001097131999586054 s 0.000012526099990282091 s 0.88
GenDot / DefOpt / cpu / Forward 0.000017099699987284113 s 0.000017279039984714472 s 0.99
GenDot / IDefOpt / cpu / Forward 0.000010955619995911548 s 0.000012846839963458478 s 0.85
GenDot / JaXPipe / cpu / PreRev 0.000012257100008810084 s 0.000012676279966399309 s 0.97
GenDot / JaXPipe / cpu / PostRev 0.000010664799997357476 s 0.000011538660028236335 s 0.92
GenDot / JaXPipe / cpu / BothRev 0.000011642160015981064 s 0.000016753639984017353 s 0.69
GenDot / Jax / cpu / BothRev 0.000010723319996941428 s 0.000011219420048291796 s 0.96
GenDot / HLOOpt / cpu / PreRev 0.000012119660011649104 s 0.000012427280007614171 s 0.98
GenDot / HLOOpt / cpu / PostRev 0.00001509053999598109 s 0.000012331320021985448 s 1.22
GenDot / HLOOpt / cpu / BothRev 0.000013554020003994085 s 0.000013908260016251006 s 0.97
GenDot / PartOpt / cpu / PreRev 0.000011333700003888225 s 0.000011689480024870137 s 0.97
GenDot / PartOpt / cpu / PostRev 0.000010589520011308195 s 0.000011357859930285486 s 0.93
GenDot / PartOpt / cpu / BothRev 0.000010748720007995871 s 0.0000118991400449886 s 0.90
GenDot / IPartOpt / cpu / PreRev 0.000017186180009503005 s 0.000011938199986616382 s 1.44
GenDot / IPartOpt / cpu / PostRev 0.000010466019994055386 s 0.000011882080034411048 s 0.88
GenDot / IPartOpt / cpu / BothRev 0.000011465680001947476 s 0.000011801500022556866 s 0.97
GenDot / DefOpt / cpu / PreRev 0.000011654240004190798 s 0.000012377880011626983 s 0.94
GenDot / DefOpt / cpu / PostRev 0.000010896979993049173 s 0.00001185160002023622 s 0.92
GenDot / DefOpt / cpu / BothRev 0.000011346240003149432 s 0.000011738000039258622 s 0.97
GenDot / IDefOpt / cpu / PreRev 0.000011781420009810973 s 0.000012479819988584496 s 0.94
GenDot / IDefOpt / cpu / PostRev 0.00001154707999603488 s 0.00001180088001092372 s 0.98
GenDot / IDefOpt / cpu / BothRev 0.000010908919996381885 s 0.000012278259991944652 s 0.89
GenDot / JaXPipe / cuda / Primal 0.000002016 s 0.000002047 s 0.98
GenDot / Jax / cuda / Primal 0.000002016 s 0.000002047 s 0.98
GenDot / HLOOpt / cuda / Primal 0.000002015 s 0.000002015 s 1
GenDot / PartOpt / cuda / Primal 0.000002016 s 0.000002016 s 1
GenDot / IPartOpt / cuda / Primal 0.000002016 s 0.000002047 s 0.98
GenDot / DefOpt / cuda / Primal 0.000002015 s 0.000002015 s 1
GenDot / IDefOpt / cuda / Primal 0.000002016 s 0.000002016 s 1
GenDot / JaXPipe / cuda / Forward 0.000010304 s 0.000010144 s 1.02
GenDot / Jax / cuda / Forward 0.000009856 s 0.000010336 s 0.95
GenDot / HLOOpt / cuda / Forward 0.000010208 s 0.000010016 s 1.02
GenDot / PartOpt / cuda / Forward 0.00000992 s 0.000010272 s 0.97
GenDot / IPartOpt / cuda / Forward 0.00001024 s 0.000009984 s 1.03
GenDot / DefOpt / cuda / Forward 0.00001008 s 0.000010592 s 0.95
GenDot / IDefOpt / cuda / Forward 0.000010176 s 0.000010111 s 1.01
GenDot / JaXPipe / cuda / PreRev 0.000010113 s 0.000010048 s 1.01
GenDot / JaXPipe / cuda / PostRev 0.000010336 s 0.000010592 s 0.98
GenDot / JaXPipe / cuda / BothRev 0.000010112 s 0.000010176 s 0.99
GenDot / Jax / cuda / BothRev 0.000009984 s 0.000009824 s 1.02
GenDot / HLOOpt / cuda / PreRev 0.000009952 s 0.000010111 s 0.98
GenDot / HLOOpt / cuda / PostRev 0.00000944 s 0.000010048 s 0.94
GenDot / HLOOpt / cuda / BothRev 0.000009503 s 0.000010369 s 0.92
GenDot / PartOpt / cuda / PreRev 0.00001024 s 0.00001008 s 1.02
GenDot / PartOpt / cuda / PostRev 0.000010144 s 0.000010464 s 0.97
GenDot / PartOpt / cuda / BothRev 0.000010048 s 0.000010176 s 0.99
GenDot / IPartOpt / cuda / PreRev 0.000010048 s 0.000010271 s 0.98
GenDot / IPartOpt / cuda / PostRev 0.000010176 s 0.000010336 s 0.98
GenDot / IPartOpt / cuda / BothRev 0.000010688 s 0.000010143 s 1.05
GenDot / DefOpt / cuda / PreRev 0.000010528 s 0.0000104 s 1.01
GenDot / DefOpt / cuda / PostRev 0.000010048 s 0.000010144 s 0.99
GenDot / DefOpt / cuda / BothRev 0.000009759 s 0.00000976 s 1.00
GenDot / IDefOpt / cuda / PreRev 0.000010112 s 0.000010208 s 0.99
GenDot / IDefOpt / cuda / PostRev 0.000010049 s 0.000010176 s 0.99
GenDot / IDefOpt / cuda / BothRev 0.000009985 s 0.000009984 s 1.00
GenDot / JaXPipe / tpu / Primal 9.21075e-7 s 9.30075e-7 s 0.99
GenDot / Jax / tpu / Primal 9.403e-7 s 9.3635e-7 s 1.00
GenDot / HLOOpt / tpu / Primal 0.00000161355 s 0.00000158145 s 1.02
GenDot / PartOpt / tpu / Primal 9.40025e-7 s 9.3585e-7 s 1.00
GenDot / IPartOpt / tpu / Primal 9.8745e-7 s 9.3965e-7 s 1.05
GenDot / DefOpt / tpu / Primal 0.0000015063499999999998 s 0.0000014923500000000002 s 1.01
GenDot / IDefOpt / tpu / Primal 0.0000016167500000000002 s 0.000001577 s 1.03
GenDot / JaXPipe / tpu / Forward 0.000003060425 s 0.000003169925 s 0.97
GenDot / Jax / tpu / Forward 0.0000023204 s 0.0000023474 s 0.99
GenDot / HLOOpt / tpu / Forward 0.00000311285 s 0.0000031217 s 1.00
GenDot / PartOpt / tpu / Forward 0.000003116075 s 0.000003223 s 0.97
GenDot / IPartOpt / tpu / Forward 0.000003120175 s 0.00000312165 s 1.00
GenDot / DefOpt / tpu / Forward 0.0000031089 s 0.0000032237 s 0.96
GenDot / IDefOpt / tpu / Forward 0.00000311095 s 0.000003127325 s 0.99
GenDot / JaXPipe / tpu / PreRev 0.000002941875 s 0.000002970675 s 0.99
GenDot / JaXPipe / tpu / PostRev 0.000002382775 s 0.000002402475 s 0.99
GenDot / JaXPipe / tpu / BothRev 0.000002959025 s 0.000002967475 s 1.00
GenDot / Jax / tpu / BothRev 0.0000023764250000000003 s 0.000002400975 s 0.99
GenDot / HLOOpt / tpu / PreRev 0.000002957475 s 0.0000029706 s 1.00
GenDot / HLOOpt / tpu / PostRev 0.0000029856 s 0.000002942875 s 1.01
GenDot / HLOOpt / tpu / BothRev 0.000002950825 s 0.000002962125 s 1.00
GenDot / PartOpt / tpu / PreRev 0.0000029891000000000003 s 0.00000294405 s 1.02
GenDot / PartOpt / tpu / PostRev 0.0000024054000000000003 s 0.0000023922750000000003 s 1.01
GenDot / PartOpt / tpu / BothRev 0.000002982175 s 0.0000029394500000000003 s 1.01
GenDot / IPartOpt / tpu / PreRev 0.000002947325 s 0.000002957075 s 1.00
GenDot / IPartOpt / tpu / PostRev 0.0000023784 s 0.000002408875 s 0.99
GenDot / IPartOpt / tpu / BothRev 0.000002957175 s 0.000002964075 s 1.00
GenDot / DefOpt / tpu / PreRev 0.000002976075 s 0.00000294815 s 1.01
GenDot / DefOpt / tpu / PostRev 0.000002950525 s 0.0000029699500000000003 s 0.99
GenDot / DefOpt / tpu / BothRev 0.000002979625 s 0.000002940925 s 1.01
GenDot / IDefOpt / tpu / PreRev 0.000002950825 s 0.000002966275 s 0.99
GenDot / IDefOpt / tpu / PostRev 0.000002976075 s 0.00000293485 s 1.01
GenDot / IDefOpt / tpu / BothRev 0.0000029458250000000003 s 0.000002959825 s 1.00
GenDot / JaXPipe / cpu / Primal 0.000014801 s 0.000008797720001894049 s 1.68
GenDot / Jax / cpu / Primal 0.000014984 s 0.000008127900009640144 s 1.84
GenDot / HLOOpt / cpu / Primal 0.00002134 s 0.000012815599984605796 s 1.67
GenDot / PartOpt / cpu / Primal 0.000014863 s 0.000008449520009889966 s 1.76
GenDot / IPartOpt / cpu / Primal 0.00001493 s 0.000008597039959568065 s 1.74
GenDot / DefOpt / cpu / Primal 0.000013718 s 0.000009416619950570747 s 1.46
GenDot / IDefOpt / cpu / Primal 0.000014323 s 0.000009196099981636509 s 1.56
GenDot / JaXPipe / cpu / Forward 0.000019645 s 0.000012595799989867374 s 1.56
GenDot / Jax / cpu / Forward 0.00002098 s 0.000011937699982809135 s 1.76
GenDot / HLOOpt / cpu / Forward 0.000019521 s 0.000016989840050882777 s 1.15
GenDot / PartOpt / cpu / Forward 0.000019441 s 0.0000176848600312951 s 1.10
GenDot / IPartOpt / cpu / Forward 0.000019649 s 0.000012526099990282091 s 1.57
GenDot / DefOpt / cpu / Forward 0.000019512 s 0.000017279039984714472 s 1.13
GenDot / IDefOpt / cpu / Forward 0.000019137 s 0.000012846839963458478 s 1.49
GenDot / JaXPipe / cpu / PreRev 0.000019489 s 0.000012676279966399309 s 1.54
GenDot / JaXPipe / cpu / PostRev 0.000021278 s 0.000011538660028236335 s 1.84
GenDot / JaXPipe / cpu / BothRev 0.000019813 s 0.000016753639984017353 s 1.18
GenDot / Jax / cpu / BothRev 0.00002093 s 0.000011219420048291796 s 1.87
GenDot / HLOOpt / cpu / PreRev 0.000018535 s 0.000012427280007614171 s 1.49
GenDot / HLOOpt / cpu / PostRev 0.000019883000000000003 s 0.000012331320021985448 s 1.61
GenDot / HLOOpt / cpu / BothRev 0.000019256 s 0.000013908260016251006 s 1.38
GenDot / PartOpt / cpu / PreRev 0.000019207 s 0.000011689480024870137 s 1.64
GenDot / PartOpt / cpu / PostRev 0.000021351 s 0.000011357859930285486 s 1.88
GenDot / PartOpt / cpu / BothRev 0.000019636 s 0.0000118991400449886 s 1.65
GenDot / IPartOpt / cpu / PreRev 0.000019319 s 0.000011938199986616382 s 1.62
GenDot / IPartOpt / cpu / PostRev 0.000021175 s 0.000011882080034411048 s 1.78
GenDot / IPartOpt / cpu / BothRev 0.000019304 s 0.000011801500022556866 s 1.64
GenDot / DefOpt / cpu / PreRev 0.000019788 s 0.000012377880011626983 s 1.60
GenDot / DefOpt / cpu / PostRev 0.000020156 s 0.00001185160002023622 s 1.70
GenDot / DefOpt / cpu / BothRev 0.000019505 s 0.000011738000039258622 s 1.66
GenDot / IDefOpt / cpu / PreRev 0.000019741000000000003 s 0.000012479819988584496 s 1.58
GenDot / IDefOpt / cpu / PostRev 0.000020126 s 0.00001180088001092372 s 1.71
GenDot / IDefOpt / cpu / BothRev 0.000019763 s 0.000012278259991944652 s 1.61
hlo_ffi / JaXPipe / cpu / Primal 0.00001108872000258998 s 0.000010822440017363988 s 1.02
hlo_ffi / Jax / cpu / Primal 0.000010399159993994544 s 0.000010285600001225249 s 1.01
hlo_ffi / HLOOpt / cpu / Primal 0.00001432895999187167 s 0.000013866080007574056 s 1.03
hlo_ffi / PartOpt / cpu / Primal 0.000009614979999241767 s 0.000009819259985306416 s 0.98
hlo_ffi / IPartOpt / cpu / Primal 0.00001039095999658457 s 0.000010091600024679792 s 1.03
hlo_ffi / DefOpt / cpu / Primal 0.000013756320008724287 s 0.000010069340023619588 s 1.37
hlo_ffi / IDefOpt / cpu / Primal 0.000010297940004875272 s 0.000009854680001808449 s 1.04
hlo_ffi / JaXPipe / cpu / Forward 0.000014509259999613278 s 0.000014822959992670804 s 0.98
hlo_ffi / Jax / cpu / Forward 0.000014524600003369414 s 0.000015186000036919722 s 0.96
hlo_ffi / HLOOpt / cpu / Forward 0.00001482992000546801 s 0.00001533519999611599 s 0.97
hlo_ffi / PartOpt / cpu / Forward 0.000014599199998883705 s 0.000015111639986571393 s 0.97
hlo_ffi / IPartOpt / cpu / Forward 0.000014593340006285871 s 0.000014500600018436671 s 1.01
hlo_ffi / DefOpt / cpu / Forward 0.000014727720008522738 s 0.000014748540043001412 s 1.00
hlo_ffi / IDefOpt / cpu / Forward 0.000014917699993475252 s 0.000014836160007689612 s 1.01
hlo_ffi / JaXPipe / cpu / PreRev 0.000014350100004776322 s 0.0000182156799655786 s 0.79
hlo_ffi / JaXPipe / cpu / PostRev 0.000014715439999690715 s 0.000014539180056090117 s 1.01
hlo_ffi / JaXPipe / cpu / BothRev 0.00001442673999690669 s 0.000014412620002985931 s 1.00
hlo_ffi / Jax / cpu / BothRev 0.00001419042000634363 s 0.000015165359964157688 s 0.94
hlo_ffi / HLOOpt / cpu / PreRev 0.000014442780000081256 s 0.000014253160006774124 s 1.01
hlo_ffi / HLOOpt / cpu / PostRev 0.000014407880003091125 s 0.000014273299966589549 s 1.01
hlo_ffi / HLOOpt / cpu / BothRev 0.000016216299998177418 s 0.00001619226000912022 s 1.00
hlo_ffi / PartOpt / cpu / PreRev 0.000014471740000772116 s 0.000014413839962799104 s 1.00
hlo_ffi / PartOpt / cpu / PostRev 0.000014559339992956666 s 0.000014372299983733682 s 1.01
hlo_ffi / PartOpt / cpu / BothRev 0.000014645680005287432 s 0.000014556479991369995 s 1.01
hlo_ffi / IPartOpt / cpu / PreRev 0.00001413834000004499 s 0.00001478366000810638 s 0.96
hlo_ffi / IPartOpt / cpu / PostRev 0.000014425780011606547 s 0.000014571639958376182 s 0.99
hlo_ffi / IPartOpt / cpu / BothRev 0.000014231559994186682 s 0.00001460164001400699 s 0.97
hlo_ffi / DefOpt / cpu / PreRev 0.000014671740007088376 s 0.000014733459993294671 s 1.00
hlo_ffi / DefOpt / cpu / PostRev 0.000013581799996700284 s 0.000014538120021825309 s 0.93
hlo_ffi / DefOpt / cpu / BothRev 0.000014569300003586247 s 0.000014464760015471256 s 1.01
hlo_ffi / IDefOpt / cpu / PreRev 0.000014410839994525305 s 0.000014284580047387862 s 1.01
hlo_ffi / IDefOpt / cpu / PostRev 0.000014940240002943029 s 0.000014173599975038087 s 1.05
hlo_ffi / IDefOpt / cpu / BothRev 0.000013810539994665306 s 0.00001527166002233571 s 0.90
hlo_ffi / JaXPipe / cuda / Primal 0.000001984 s 0.000001984 s 1
hlo_ffi / Jax / cuda / Primal 0.000001984 s 0.000001984 s 1
hlo_ffi / HLOOpt / cuda / Primal 0.000001984 s 0.000001984 s 1
hlo_ffi / PartOpt / cuda / Primal 0.000002015 s 0.000002015 s 1
hlo_ffi / IPartOpt / cuda / Primal 0.000001984 s 0.000001984 s 1
hlo_ffi / DefOpt / cuda / Primal 0.000001984 s 0.000001984 s 1
hlo_ffi / IDefOpt / cuda / Primal 0.000001983 s 0.000002015 s 0.98
hlo_ffi / JaXPipe / cuda / Forward 0.00000208 s 0.00000208 s 1
hlo_ffi / Jax / cuda / Forward 0.00000208 s 0.00000208 s 1
hlo_ffi / HLOOpt / cuda / Forward 0.000002079 s 0.00000208 s 1.00
hlo_ffi / PartOpt / cuda / Forward 0.00000208 s 0.00000208 s 1
hlo_ffi / IPartOpt / cuda / Forward 0.00000208 s 0.00000208 s 1
hlo_ffi / DefOpt / cuda / Forward 0.00000208 s 0.00000208 s 1
hlo_ffi / IDefOpt / cuda / Forward 0.00000208 s 0.00000208 s 1
hlo_ffi / JaXPipe / cuda / PreRev 0.00000208 s 0.000002047 s 1.02
hlo_ffi / JaXPipe / cuda / PostRev 0.000002079 s 0.000002047 s 1.02
hlo_ffi / JaXPipe / cuda / BothRev 0.00000208 s 0.000002079 s 1.00
hlo_ffi / Jax / cuda / BothRev 0.00000208 s 0.00000208 s 1
hlo_ffi / HLOOpt / cuda / PreRev 0.000002079 s 0.00000208 s 1.00
hlo_ffi / HLOOpt / cuda / PostRev 0.00000208 s 0.00000208 s 1
hlo_ffi / HLOOpt / cuda / BothRev 0.00000208 s 0.00000208 s 1
hlo_ffi / PartOpt / cuda / PreRev 0.00000208 s 0.000002079 s 1.00
hlo_ffi / PartOpt / cuda / PostRev 0.000002048 s 0.00000208 s 0.98
hlo_ffi / PartOpt / cuda / BothRev 0.00000208 s 0.00000208 s 1
hlo_ffi / IPartOpt / cuda / PreRev 0.000002048 s 0.000002048 s 1
hlo_ffi / IPartOpt / cuda / PostRev 0.000002048 s 0.00000208 s 0.98
hlo_ffi / IPartOpt / cuda / BothRev 0.00000208 s 0.000002048 s 1.02
hlo_ffi / DefOpt / cuda / PreRev 0.00000208 s 0.00000208 s 1
hlo_ffi / DefOpt / cuda / PostRev 0.000002047 s 0.000002048 s 1.00
hlo_ffi / DefOpt / cuda / BothRev 0.000002079 s 0.00000208 s 1.00
hlo_ffi / IDefOpt / cuda / PreRev 0.00000208 s 0.000002048 s 1.02
hlo_ffi / IDefOpt / cuda / PostRev 0.00000208 s 0.00000208 s 1
hlo_ffi / IDefOpt / cuda / BothRev 0.000002048 s 0.00000208 s 0.98
hlo_ffi / JaXPipe / tpu / Primal 9.1995e-7 s 9.2325e-7 s 1.00
hlo_ffi / Jax / tpu / Primal 9.49425e-7 s 9.5055e-7 s 1.00
hlo_ffi / HLOOpt / tpu / Primal 8.966250000000001e-7 s 8.98425e-7 s 1.00
hlo_ffi / PartOpt / tpu / Primal 9.5055e-7 s 9.60775e-7 s 0.99
hlo_ffi / IPartOpt / tpu / Primal 9.029e-7 s 9.03925e-7 s 1.00
hlo_ffi / DefOpt / tpu / Primal 9.53025e-7 s 9.5335e-7 s 1.00
hlo_ffi / IDefOpt / tpu / Primal 9.004e-7 s 8.985250000000001e-7 s 1.00
hlo_ffi / JaXPipe / tpu / Forward 9.4885e-7 s 9.49575e-7 s 1.00
hlo_ffi / Jax / tpu / Forward 9.81925e-7 s 9.8175e-7 s 1.00
hlo_ffi / HLOOpt / tpu / Forward 9.739e-7 s 9.74075e-7 s 1.00
hlo_ffi / PartOpt / tpu / Forward 9.343e-7 s 9.33825e-7 s 1.00
hlo_ffi / IPartOpt / tpu / Forward 9.73775e-7 s 9.73925e-7 s 1.00
hlo_ffi / DefOpt / tpu / Forward 9.33725e-7 s 9.339e-7 s 1.00
hlo_ffi / IDefOpt / tpu / Forward 9.742e-7 s 9.7405e-7 s 1.00
hlo_ffi / JaXPipe / tpu / PreRev 9.32325e-7 s 9.32325e-7 s 1
hlo_ffi / JaXPipe / tpu / PostRev 9.64325e-7 s 9.6485e-7 s 1.00
hlo_ffi / JaXPipe / tpu / BothRev 9.5975e-7 s 9.59925e-7 s 1.00
hlo_ffi / Jax / tpu / BothRev 9.65e-7 s 9.65e-7 s 1
hlo_ffi / HLOOpt / tpu / PreRev 9.603e-7 s 9.60075e-7 s 1.00
hlo_ffi / HLOOpt / tpu / PostRev 9.649e-7 s 9.6455e-7 s 1.00
hlo_ffi / HLOOpt / tpu / BothRev 9.6025e-7 s 9.60425e-7 s 1.00
hlo_ffi / PartOpt / tpu / PreRev 9.64575e-7 s 9.65525e-7 s 1.00
hlo_ffi / PartOpt / tpu / PostRev 9.59775e-7 s 9.6045e-7 s 1.00
hlo_ffi / PartOpt / tpu / BothRev 9.64625e-7 s 9.6435e-7 s 1.00
hlo_ffi / IPartOpt / tpu / PreRev 9.596250000000002e-7 s 9.6025e-7 s 1.00
hlo_ffi / IPartOpt / tpu / PostRev 9.64675e-7 s 9.650749999999998e-7 s 1.00
hlo_ffi / IPartOpt / tpu / BothRev 9.6025e-7 s 9.602e-7 s 1.00
hlo_ffi / DefOpt / tpu / PreRev 9.65e-7 s 9.64625e-7 s 1.00
hlo_ffi / DefOpt / tpu / PostRev 9.59975e-7 s 9.6075e-7 s 1.00
hlo_ffi / DefOpt / tpu / BothRev 9.647e-7 s 9.65325e-7 s 1.00
hlo_ffi / IDefOpt / tpu / PreRev 9.596250000000002e-7 s 9.606e-7 s 1.00
hlo_ffi / IDefOpt / tpu / PostRev 9.65025e-7 s 9.65225e-7 s 1.00
hlo_ffi / IDefOpt / tpu / BothRev 9.59925e-7 s 9.600250000000002e-7 s 1.00
hlo_ffi / JaXPipe / cpu / Primal 0.000017979 s 0.000010822440017363988 s 1.66
hlo_ffi / Jax / cpu / Primal 0.000017888000000000002 s 0.000010285600001225249 s 1.74
hlo_ffi / HLOOpt / cpu / Primal 0.000017664 s 0.000013866080007574056 s 1.27
hlo_ffi / PartOpt / cpu / Primal 0.000017627 s 0.000009819259985306416 s 1.80
hlo_ffi / IPartOpt / cpu / Primal 0.000017806 s 0.000010091600024679792 s 1.76
hlo_ffi / DefOpt / cpu / Primal 0.000017433 s 0.000010069340023619588 s 1.73
hlo_ffi / IDefOpt / cpu / Primal 0.000017537 s 0.000009854680001808449 s 1.78
hlo_ffi / JaXPipe / cpu / Forward 0.000025136 s 0.000014822959992670804 s 1.70
hlo_ffi / Jax / cpu / Forward 0.000025237 s 0.000015186000036919722 s 1.66
hlo_ffi / HLOOpt / cpu / Forward 0.000024677 s 0.00001533519999611599 s 1.61
hlo_ffi / PartOpt / cpu / Forward 0.00002493 s 0.000015111639986571393 s 1.65
hlo_ffi / IPartOpt / cpu / Forward 0.00002467 s 0.000014500600018436671 s 1.70
hlo_ffi / DefOpt / cpu / Forward 0.000024377 s 0.000014748540043001412 s 1.65
hlo_ffi / IDefOpt / cpu / Forward 0.000024297 s 0.000014836160007689612 s 1.64
hlo_ffi / JaXPipe / cpu / PreRev 0.000024492 s 0.0000182156799655786 s 1.34
hlo_ffi / JaXPipe / cpu / PostRev 0.000025612 s 0.000014539180056090117 s 1.76
hlo_ffi / JaXPipe / cpu / BothRev 0.000025452 s 0.000014412620002985931 s 1.77
hlo_ffi / Jax / cpu / BothRev 0.000025201 s 0.000015165359964157688 s 1.66
hlo_ffi / HLOOpt / cpu / PreRev 0.0000248 s 0.000014253160006774124 s 1.74
hlo_ffi / HLOOpt / cpu / PostRev 0.000025613 s 0.000014273299966589549 s 1.79
hlo_ffi / HLOOpt / cpu / BothRev 0.000024749 s 0.00001619226000912022 s 1.53
hlo_ffi / PartOpt / cpu / PreRev 0.000025257 s 0.000014413839962799104 s 1.75
hlo_ffi / PartOpt / cpu / PostRev 0.000024803 s 0.000014372299983733682 s 1.73
hlo_ffi / PartOpt / cpu / BothRev 0.000024518 s 0.000014556479991369995 s 1.68
hlo_ffi / IPartOpt / cpu / PreRev 0.000025363 s 0.00001478366000810638 s 1.72
hlo_ffi / IPartOpt / cpu / PostRev 0.000026213 s 0.000014571639958376182 s 1.80
hlo_ffi / IPartOpt / cpu / BothRev 0.000024919 s 0.00001460164001400699 s 1.71
hlo_ffi / DefOpt / cpu / PreRev 0.000024997 s 0.000014733459993294671 s 1.70
hlo_ffi / DefOpt / cpu / PostRev 0.000026489 s 0.000014538120021825309 s 1.82
hlo_ffi / DefOpt / cpu / BothRev 0.000026393 s 0.000014464760015471256 s 1.82
hlo_ffi / IDefOpt / cpu / PreRev 0.000024252 s 0.000014284580047387862 s 1.70
hlo_ffi / IDefOpt / cpu / PostRev 0.000026493 s 0.000014173599975038087 s 1.87
hlo_ffi / IDefOpt / cpu / BothRev 0.000025773 s 0.00001527166002233571 s 1.69
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / JaXPipe / cpu / Primal 0.001113750200011 s 0.0011677415998747 s 0.95
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / Jax / cpu / Primal 0.0009457276000148 s 0.0010169414000301 s 0.93
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / HLOOpt / cpu / Primal 0.0010015292000161 s 0.0009394241999871 s 1.07
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / PartOpt / cpu / Primal 0.0009697868000102 s 0.000897186199927 s 1.08
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IPartOpt / cpu / Primal 0.0009363741999777 s 0.0008884472000318 s 1.05
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / DefOpt / cpu / Primal 0.0009786775999828 s 0.0009461017998546 s 1.03
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IDefOpt / cpu / Primal 0.0010293989999809 s 0.0009835020000537 s 1.05
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / JaXPipe / cpu / Forward 0.0027043279999816 s 0.0027426424001532 s 0.99
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / Jax / cpu / Forward 0.0025499381999907 s 0.0022887613999046 s 1.11
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / HLOOpt / cpu / Forward 0.0024127479999606 s 0.0021659468000507 s 1.11
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / PartOpt / cpu / Forward 0.0022893057999681 s 0.002168072799941 s 1.06
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IPartOpt / cpu / Forward 0.0024831801999653 s 0.0023066482000103 s 1.08
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / DefOpt / cpu / Forward 0.0024889160000384 s 0.0021572545998424 s 1.15
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IDefOpt / cpu / Forward 0.0024799177999966 s 0.002206446800119 s 1.12
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / JaXPipe / cpu / PreRev 0.0058833122000123 s 0.0067595389999041 s 0.87
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / JaXPipe / cpu / PostRev 0.0067876473999831 s 0.0058482147999711 s 1.16
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / JaXPipe / cpu / BothRev 0.0067923801999768 s 0.0057265669998741 s 1.19
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / Jax / cpu / BothRev 0.0062905406000254 s 0.005685240800085 s 1.11
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / HLOOpt / cpu / PreRev 0.0063517330000195 s 0.0056048894001833 s 1.13
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / HLOOpt / cpu / PostRev 0.0056577950000018 s 0.005353641199963 s 1.06
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / HLOOpt / cpu / BothRev 0.0057180417999688 s 0.0055430060001526 s 1.03
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / PartOpt / cpu / PreRev 0.0053945812000165 s 0.0056180495999797 s 0.96
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / PartOpt / cpu / PostRev 0.0056010369999967 s 0.0057529201998477 s 0.97
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / PartOpt / cpu / BothRev 0.0059857673999886 s 0.0055169804000797 s 1.08
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IPartOpt / cpu / PreRev 0.0051625880000074 s 0.0053986708001502 s 0.96
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IPartOpt / cpu / PostRev 0.0038596079999479 s 0.0054266727999674 s 0.71
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IPartOpt / cpu / BothRev 0.0050896610000108 s 0.0053957413999341 s 0.94
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / DefOpt / cpu / PreRev 0.003423768000016 s 0.0055037852001078 s 0.62
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / DefOpt / cpu / PostRev 0.004659365200041 s 0.0054062096000052 s 0.86
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / DefOpt / cpu / BothRev 0.0040709921999905 s 0.0058279300000322 s 0.70
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IDefOpt / cpu / PreRev 0.005069977800008 s 0.0056282245999682 s 0.90
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IDefOpt / cpu / PostRev 0.0035011601999713 s 0.0055000705999191 s 0.64
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IDefOpt / cpu / BothRev 0.0050842067999838 s 0.0055692312000246 s 0.91
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / JaXPipe / cuda / Primal 0.0002756469999999 s 0.000273632 s 1.01
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / Jax / cuda / Primal 0.000274496 s 0.000273696 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / HLOOpt / cuda / Primal 0.000288767 s 0.000288319 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / PartOpt / cuda / Primal 0.000274175 s 0.000273696 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IPartOpt / cuda / Primal 0.0002757749999999 s 0.000275327 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / DefOpt / cuda / Primal 0.0002892159999999 s 0.0002884479999999 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IDefOpt / cuda / Primal 0.000288831 s 0.000287967 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / JaXPipe / cuda / Forward 0.000559807 s 0.000559071 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / Jax / cuda / Forward 0.000540255 s 0.000539583 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / HLOOpt / cuda / Forward 0.000558175 s 0.000558623 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / PartOpt / cuda / Forward 0.000558943 s 0.000558911 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IPartOpt / cuda / Forward 0.000559295 s 0.000559583 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / DefOpt / cuda / Forward 0.000559775 s 0.000559839 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IDefOpt / cuda / Forward 0.000559774 s 0.000559518 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / JaXPipe / cuda / PreRev 0.001017342 s 0.001017982 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / JaXPipe / cuda / PostRev 0.000987358 s 0.000986878 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / JaXPipe / cuda / BothRev 0.0010095019999999 s 0.001008765 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / Jax / cuda / BothRev 0.000982333 s 0.00098259 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / HLOOpt / cuda / PreRev 0.001007774 s 0.001008734 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / HLOOpt / cuda / PostRev 0.001034429 s 0.001032286 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / HLOOpt / cuda / BothRev 0.001006782 s 0.001008798 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / PartOpt / cuda / PreRev 0.00101011 s 0.0010106219999999 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / PartOpt / cuda / PostRev 0.000974366 s 0.000973885 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / PartOpt / cuda / BothRev 0.001009022 s 0.001009054 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IPartOpt / cuda / PreRev 0.001007934 s 0.001010174 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IPartOpt / cuda / PostRev 0.000975646 s 0.000973214 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IPartOpt / cuda / BothRev 0.001008734 s 0.001008542 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / DefOpt / cuda / PreRev 0.001021822 s 0.001020029 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / DefOpt / cuda / PostRev 0.000957246 s 0.0009578529999999 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / DefOpt / cuda / BothRev 0.001020702 s 0.001020446 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IDefOpt / cuda / PreRev 0.001020669 s 0.001020511 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IDefOpt / cuda / PostRev 0.0010169889999999 s 0.001016861 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IDefOpt / cuda / BothRev 0.00102179 s 0.001020797 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / JaXPipe / tpu / Primal 0.0001306885 s 0.0001266495 s 1.03
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / Jax / tpu / Primal 0.00012426825 s 0.0001280524999999 s 0.97
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / HLOOpt / tpu / Primal 0.00016020525 s 0.0001555445 s 1.03
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / PartOpt / tpu / Primal 0.00013078475 s 0.000135421 s 0.97
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IPartOpt / tpu / Primal 0.0001381985 s 0.000133678 s 1.03
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / DefOpt / tpu / Primal 0.000145043 s 0.00014911425 s 0.97
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IDefOpt / tpu / Primal 0.0001580225 s 0.0001540294999999 s 1.03
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / JaXPipe / tpu / Forward 0.0002136905 s 0.0002146765 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / Jax / tpu / Forward 0.0002626389999999 s 0.0002613837499999 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / HLOOpt / tpu / Forward 0.00022041575 s 0.0002143195 s 1.03
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / PartOpt / tpu / Forward 0.00021502175 s 0.00021728225 s 0.99
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IPartOpt / tpu / Forward 0.0002163485 s 0.0002145055 s 1.01
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / DefOpt / tpu / Forward 0.000217868 s 0.0002170342499999 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IDefOpt / tpu / Forward 0.00021611425 s 0.00021452425 s 1.01
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / JaXPipe / tpu / PreRev 0.00035566175 s 0.00035627475 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / JaXPipe / tpu / PostRev 0.0002566625 s 0.0002561875 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / JaXPipe / tpu / BothRev 0.0003557385 s 0.000357608 s 0.99
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / Jax / tpu / BothRev 0.00025661375 s 0.00025765525 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / HLOOpt / tpu / PreRev 0.000355499 s 0.00035707325 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / HLOOpt / tpu / PostRev 0.00029069625 s 0.000291246 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / HLOOpt / tpu / BothRev 0.00035612575 s 0.0003571442499999 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / PartOpt / tpu / PreRev 0.0003554795 s 0.0003553715 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / PartOpt / tpu / PostRev 0.000271586 s 0.00027444725 s 0.99
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / PartOpt / tpu / BothRev 0.0003552262499999 s 0.000355547 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IPartOpt / tpu / PreRev 0.00035594175 s 0.0003567862499999 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IPartOpt / tpu / PostRev 0.00027160375 s 0.00027269325 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IPartOpt / tpu / BothRev 0.000355519 s 0.0003570915 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / DefOpt / tpu / PreRev 0.00035796775 s 0.000357857 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / DefOpt / tpu / PostRev 0.0002829089999999 s 0.00028378575 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / DefOpt / tpu / BothRev 0.000357477 s 0.0003581902499999 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IDefOpt / tpu / PreRev 0.000358083 s 0.0003591905 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IDefOpt / tpu / PostRev 0.00029808275 s 0.00029874275 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IDefOpt / tpu / BothRev 0.0003582819999999 s 0.00035939375 s 1.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / JaXPipe / cpu / Primal 0.002595555 s 0.0011677415998747 s 2.22
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / Jax / cpu / Primal 0.0025726589999999 s 0.0010169414000301 s 2.53
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / HLOOpt / cpu / Primal 0.002552586 s 0.0009394241999871 s 2.72
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / PartOpt / cpu / Primal 0.002322997 s 0.000897186199927 s 2.59
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IPartOpt / cpu / Primal 0.002666251 s 0.0008884472000318 s 3.00
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / DefOpt / cpu / Primal 0.002523841 s 0.0009461017998546 s 2.67
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IDefOpt / cpu / Primal 0.002878963 s 0.0009835020000537 s 2.93
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / JaXPipe / cpu / Forward 0.0067617979999999 s 0.0027426424001532 s 2.47
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / Jax / cpu / Forward 0.006420085 s 0.0022887613999046 s 2.81
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / HLOOpt / cpu / Forward 0.006253628 s 0.0021659468000507 s 2.89
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / PartOpt / cpu / Forward 0.006462881 s 0.002168072799941 s 2.98
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IPartOpt / cpu / Forward 0.006731484 s 0.0023066482000103 s 2.92
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / DefOpt / cpu / Forward 0.006081068 s 0.0021572545998424 s 2.82
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IDefOpt / cpu / Forward 0.006648748 s 0.002206446800119 s 3.01
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / JaXPipe / cpu / PreRev 0.01113773 s 0.0067595389999041 s 1.65
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / JaXPipe / cpu / PostRev 0.010208568 s 0.0058482147999711 s 1.75
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / JaXPipe / cpu / BothRev 0.010151555 s 0.0057265669998741 s 1.77
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / Jax / cpu / BothRev 0.0099830059999999 s 0.005685240800085 s 1.76
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / HLOOpt / cpu / PreRev 0.00846717 s 0.0056048894001833 s 1.51
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / HLOOpt / cpu / PostRev 0.009074556 s 0.005353641199963 s 1.70
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / HLOOpt / cpu / BothRev 0.008663143 s 0.0055430060001526 s 1.56
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / PartOpt / cpu / PreRev 0.009717012 s 0.0056180495999797 s 1.73
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / PartOpt / cpu / PostRev 0.009632612 s 0.0057529201998477 s 1.67
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / PartOpt / cpu / BothRev 0.0102081269999999 s 0.0055169804000797 s 1.85
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IPartOpt / cpu / PreRev 0.008793828 s 0.0053986708001502 s 1.63
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IPartOpt / cpu / PostRev 0.010030523 s 0.0054266727999674 s 1.85
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IPartOpt / cpu / BothRev 0.009925397 s 0.0053957413999341 s 1.84
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / DefOpt / cpu / PreRev 0.010525616 s 0.0055037852001078 s 1.91
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / DefOpt / cpu / PostRev 0.008936678 s 0.0054062096000052 s 1.65
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / DefOpt / cpu / BothRev 0.010292888 s 0.0058279300000322 s 1.77
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IDefOpt / cpu / PreRev 0.009095053 s 0.0056282245999682 s 1.62
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IDefOpt / cpu / PostRev 0.010640675 s 0.0055000705999191 s 1.93
llama_dim_288_hidden_dim_768_n_layers_6_n_heads_6_n_kv_heads_6_vocab_size_32000_seq_len_256 / IDefOpt / cpu / BothRev 0.009895349 s 0.0055692312000246 s 1.78
scatter_sum / JaXPipe / cpu / Primal 0.00000919253999427383 s 0.000009560899970892933 s 0.96
scatter_sum / Jax / cpu / Primal 0.00000856111999382847 s 0.000008968580013970495 s 0.95
scatter_sum / HLOOpt / cpu / Primal 0.000011400040000353328 s 0.000012460140051189227 s 0.91
scatter_sum / PartOpt / cpu / Primal 0.00000780156000473653 s 0.000009277959989049123 s 0.84
scatter_sum / IPartOpt / cpu / Primal 0.000008322920011778479 s 0.000009176700023090236 s 0.91
scatter_sum / DefOpt / cpu / Primal 0.000007944719989154692 s 0.000009108180020120926 s 0.87
scatter_sum / IDefOpt / cpu / Primal 0.000008278279995010963 s 0.000009239719984179829 s 0.90
scatter_sum / JaXPipe / cpu / Forward 0.00001229448000458433 s 0.000013944200027253828 s 0.88
scatter_sum / Jax / cpu / Forward 0.00001210159999800453 s 0.000013279800023155985 s 0.91
scatter_sum / HLOOpt / cpu / Forward 0.000012145440000495 s 0.0000191325799733022 s 0.63
scatter_sum / PartOpt / cpu / Forward 0.00001185415998861572 s 0.00001404371999342402 s 0.84
scatter_sum / IPartOpt / cpu / Forward 0.000012066780006989577 s 0.000013532699949792004 s 0.89
scatter_sum / DefOpt / cpu / Forward 0.000017448179994516978 s 0.000019775379987549967 s 0.88
scatter_sum / IDefOpt / cpu / Forward 0.000011824080008864257 s 0.000013536360002035509 s 0.87
scatter_sum / JaXPipe / cpu / PreRev 0.000012807599998723165 s 0.000013416699976005476 s 0.95
scatter_sum / JaXPipe / cpu / PostRev 0.000012629379993995826 s 0.000013588860037998528 s 0.93
scatter_sum / JaXPipe / cpu / BothRev 0.000017336919993340415 s 0.000018541219978942537 s 0.94
scatter_sum / Jax / cpu / BothRev 0.000012039100004130888 s 0.000013493200040102238 s 0.89
scatter_sum / HLOOpt / cpu / PreRev 0.00001279300000078365 s 0.000013985340037834249 s 0.91
scatter_sum / HLOOpt / cpu / PostRev 0.00001717212000357904 s 0.00001821657999244053 s 0.94
scatter_sum / HLOOpt / cpu / BothRev 0.000014252419996410026 s 0.000020798099985768203 s 0.69
scatter_sum / PartOpt / cpu / PreRev 0.000012034159999529947 s 0.000014508680023936905 s 0.83
scatter_sum / PartOpt / cpu / PostRev 0.000012169960000392164 s 0.000013988859991513892 s 0.87
scatter_sum / PartOpt / cpu / BothRev 0.000011965279991272836 s 0.000013867720035705131 s 0.86
scatter_sum / IPartOpt / cpu / PreRev 0.0000172866600019006 s 0.000019941060008932256 s 0.87
scatter_sum / IPartOpt / cpu / PostRev 0.000011879779992796105 s 0.000014055500041649794 s 0.85
scatter_sum / IPartOpt / cpu / BothRev 0.00001231825999866487 s 0.000013829800009261817 s 0.89
scatter_sum / DefOpt / cpu / PreRev 0.000012029579991121864 s 0.000014402780007003455 s 0.84
scatter_sum / DefOpt / cpu / PostRev 0.00001255811999726575 s 0.00001407524003298022 s 0.89
scatter_sum / DefOpt / cpu / BothRev 0.00001213786000107575 s 0.000013597080014733363 s 0.89
scatter_sum / IDefOpt / cpu / PreRev 0.000012423539994870224 s 0.000013511479946828331 s 0.92
scatter_sum / IDefOpt / cpu / PostRev 0.000012375340008929924 s 0.000013842139969710842 s 0.89
scatter_sum / IDefOpt / cpu / BothRev 0.000012727219996122583 s 0.00001488863998929446 s 0.85
scatter_sum / JaXPipe / cuda / Primal 0.000010336 s 0.00001008 s 1.03
scatter_sum / Jax / cuda / Primal 0.000009951 s 0.000010368 s 0.96
scatter_sum / HLOOpt / cuda / Primal 0.000010272 s 0.000010304 s 1.00
scatter_sum / PartOpt / cuda / Primal 0.00001008 s 0.000010016 s 1.01
scatter_sum / IPartOpt / cuda / Primal 0.000010144 s 0.000010048 s 1.01
scatter_sum / DefOpt / cuda / Primal 0.000010048 s 0.000010207 s 0.98
scatter_sum / IDefOpt / cuda / Primal 0.000010176 s 0.000010113 s 1.01
scatter_sum / JaXPipe / cuda / Forward 0.000017312 s 0.000017408 s 0.99
scatter_sum / Jax / cuda / Forward 0.000017056 s 0.000017344 s 0.98
scatter_sum / HLOOpt / cuda / Forward 0.000017152 s 0.000017919999999999998 s 0.96
scatter_sum / PartOpt / cuda / Forward 0.00001728 s 0.000017503999999999997 s 0.99
scatter_sum / IPartOpt / cuda / Forward 0.00001696 s 0.000017664 s 0.96
scatter_sum / DefOpt / cuda / Forward 0.000017375999999999998 s 0.000016864 s 1.03
scatter_sum / IDefOpt / cuda / Forward 0.000017087 s 0.000017503999999999997 s 0.98
scatter_sum / JaXPipe / cuda / PreRev 0.00001728 s 0.00001728 s 1
scatter_sum / JaXPipe / cuda / PostRev 0.000018112 s 0.000016863 s 1.07
scatter_sum / JaXPipe / cuda / BothRev 0.000016927999999999998 s 0.000017472 s 0.97
scatter_sum / Jax / cuda / BothRev 0.000017055000000000002 s 0.00001744 s 0.98
scatter_sum / HLOOpt / cuda / PreRev 0.000016896000000000002 s 0.000016864 s 1.00
scatter_sum / HLOOpt / cuda / PostRev 0.000016864 s 0.000017312 s 0.97
scatter_sum / HLOOpt / cuda / BothRev 0.000017184 s 0.000017984 s 0.96
scatter_sum / PartOpt / cuda / PreRev 0.000016927000000000002 s 0.000017406999999999998 s 0.97
scatter_sum / PartOpt / cuda / PostRev 0.000017089 s 0.000017088 s 1.00
scatter_sum / PartOpt / cuda / BothRev 0.00001904 s 0.000016896000000000002 s 1.13
scatter_sum / IPartOpt / cuda / PreRev 0.000017216 s 0.000016927999999999998 s 1.02
scatter_sum / IPartOpt / cuda / PostRev 0.000017216 s 0.000017247999999999998 s 1.00
scatter_sum / IPartOpt / cuda / BothRev 0.000017152 s 0.000017056 s 1.01
scatter_sum / DefOpt / cuda / PreRev 0.000016704 s 0.00001824 s 0.92
scatter_sum / DefOpt / cuda / PostRev 0.000016576000000000002 s 0.000017184 s 0.96
scatter_sum / DefOpt / cuda / BothRev 0.00001696 s 0.00001696 s 1
scatter_sum / IDefOpt / cuda / PreRev 0.000017312 s 0.000016544 s 1.05
scatter_sum / IDefOpt / cuda / PostRev 0.000017152 s 0.000016992 s 1.01
scatter_sum / IDefOpt / cuda / BothRev 0.000017185 s 0.000017505 s 0.98
scatter_sum / JaXPipe / tpu / Primal 0.000001349 s 0.0000013517 s 1.00
scatter_sum / Jax / tpu / Primal 0.0000013553 s 0.00000141465 s 0.96
scatter_sum / HLOOpt / tpu / Primal 0.000001359475 s 0.000001360975 s 1.00
scatter_sum / PartOpt / tpu / Primal 0.0000013549249999999998 s 0.00000141505 s 0.96
scatter_sum / IPartOpt / tpu / Primal 0.000001359525 s 0.0000013612 s 1.00
scatter_sum / DefOpt / tpu / Primal 0.0000013545250000000002 s 0.0000014150999999999998 s 0.96
scatter_sum / IDefOpt / tpu / Primal 0.000001358925 s 0.00000136135 s 1.00
scatter_sum / JaXPipe / tpu / Forward 0.000002751075 s 0.00000271685 s 1.01
scatter_sum / Jax / tpu / Forward 0.000002780425000000001 s 0.00000274255 s 1.01
scatter_sum / HLOOpt / tpu / Forward 0.00000275745 s 0.0000027148499999999995 s 1.02
scatter_sum / PartOpt / tpu / Forward 0.00000274695 s 0.00000270915 s 1.01
scatter_sum / IPartOpt / tpu / Forward 0.000002753925 s 0.00000271675 s 1.01
scatter_sum / DefOpt / tpu / Forward 0.0000027563500000000003 s 0.00000271685 s 1.01
scatter_sum / IDefOpt / tpu / Forward 0.000002752325 s 0.000002714 s 1.01
scatter_sum / JaXPipe / tpu / PreRev 0.0000027489000000000005 s 0.000002698175 s 1.02
scatter_sum / JaXPipe / tpu / PostRev 0.00000275025 s 0.0000026941500000000003 s 1.02
scatter_sum / JaXPipe / tpu / BothRev 0.0000027641 s 0.000002710675 s 1.02
scatter_sum / Jax / tpu / BothRev 0.0000027978 s 0.000002750575 s 1.02
scatter_sum / HLOOpt / tpu / PreRev 0.00000276045 s 0.0000027181 s 1.02
scatter_sum / HLOOpt / tpu / PostRev 0.0000028073 s 0.00000275205 s 1.02
scatter_sum / HLOOpt / tpu / BothRev 0.000002757625 s 0.000002715975 s 1.02
scatter_sum / PartOpt / tpu / PreRev 0.000002802075 s 0.000002754725 s 1.02
scatter_sum / PartOpt / tpu / PostRev 0.000002759725 s 0.000002710875 s 1.02
scatter_sum / PartOpt / tpu / BothRev 0.0000027971249999999995 s 0.000002753625 s 1.02
scatter_sum / IPartOpt / tpu / PreRev 0.00000276655 s 0.000002717375 s 1.02
scatter_sum / IPartOpt / tpu / PostRev 0.000002799725 s 0.00000275435 s 1.02
scatter_sum / IPartOpt / tpu / BothRev 0.00000276145 s 0.000002715725 s 1.02
scatter_sum / DefOpt / tpu / PreRev 0.0000028019250000000003 s 0.0000027562 s 1.02
scatter_sum / DefOpt / tpu / PostRev 0.000002757775 s 0.000002710525 s 1.02
scatter_sum / DefOpt / tpu / BothRev 0.00000279815 s 0.000002755825 s 1.02
scatter_sum / IDefOpt / tpu / PreRev 0.000002762975 s 0.0000027162 s 1.02
scatter_sum / IDefOpt / tpu / PostRev 0.0000027965250000000003 s 0.00000276105 s 1.01
scatter_sum / IDefOpt / tpu / BothRev 0.00000276455 s 0.000002711775 s 1.02
scatter_sum / JaXPipe / cpu / Primal 0.000022573 s 0.000009560899970892933 s 2.36
scatter_sum / Jax / cpu / Primal 0.00001539 s 0.000008968580013970495 s 1.72
scatter_sum / HLOOpt / cpu / Primal 0.000015858 s 0.000012460140051189227 s 1.27
scatter_sum / PartOpt / cpu / Primal 0.000015678 s 0.000009277959989049123 s 1.69
scatter_sum / IPartOpt / cpu / Primal 0.000015918000000000002 s 0.000009176700023090236 s 1.73
scatter_sum / DefOpt / cpu / Primal 0.000016055 s 0.000009108180020120926 s 1.76
scatter_sum / IDefOpt / cpu / Primal 0.000016115 s 0.000009239719984179829 s 1.74
scatter_sum / JaXPipe / cpu / Forward 0.000023129 s 0.000013944200027253828 s 1.66
scatter_sum / Jax / cpu / Forward 0.000023847 s 0.000013279800023155985 s 1.80
scatter_sum / HLOOpt / cpu / Forward 0.000023139 s 0.0000191325799733022 s 1.21
scatter_sum / PartOpt / cpu / Forward 0.000022775 s 0.00001404371999342402 s 1.62
scatter_sum / IPartOpt / cpu / Forward 0.000022911 s 0.000013532699949792004 s 1.69
scatter_sum / DefOpt / cpu / Forward 0.000022869 s 0.000019775379987549967 s 1.16
scatter_sum / IDefOpt / cpu / Forward 0.000024039 s 0.000013536360002035509 s 1.78
scatter_sum / JaXPipe / cpu / PreRev 0.000022918 s 0.000013416699976005476 s 1.71
scatter_sum / JaXPipe / cpu / PostRev 0.000023396 s 0.000013588860037998528 s 1.72
scatter_sum / JaXPipe / cpu / BothRev 0.000023566 s 0.000018541219978942537 s 1.27
scatter_sum / Jax / cpu / BothRev 0.000021972 s 0.000013493200040102238 s 1.63
scatter_sum / HLOOpt / cpu / PreRev 0.000021782 s 0.000013985340037834249 s 1.56
scatter_sum / HLOOpt / cpu / PostRev 0.000024372 s 0.00001821657999244053 s 1.34
scatter_sum / HLOOpt / cpu / BothRev 0.000024246 s 0.000020798099985768203 s 1.17
scatter_sum / PartOpt / cpu / PreRev 0.000022705 s 0.000014508680023936905 s 1.56
scatter_sum / PartOpt / cpu / PostRev 0.00002316 s 0.000013988859991513892 s 1.66
scatter_sum / PartOpt / cpu / BothRev 0.000024354 s 0.000013867720035705131 s 1.76
scatter_sum / IPartOpt / cpu / PreRev 0.000022455 s 0.000019941060008932256 s 1.13
scatter_sum / IPartOpt / cpu / PostRev 0.00002283 s 0.000014055500041649794 s 1.62
scatter_sum / IPartOpt / cpu / BothRev 0.000023628 s 0.000013829800009261817 s 1.71
scatter_sum / DefOpt / cpu / PreRev 0.000022702 s 0.000014402780007003455 s 1.58
scatter_sum / DefOpt / cpu / PostRev 0.000024276 s 0.00001407524003298022 s 1.72
scatter_sum / DefOpt / cpu / BothRev 0.000023963 s 0.000013597080014733363 s 1.76
scatter_sum / IDefOpt / cpu / PreRev 0.000023456 s 0.000013511479946828331 s 1.74
scatter_sum / IDefOpt / cpu / PostRev 0.000022836 s 0.000013842139969710842 s 1.65
scatter_sum / IDefOpt / cpu / BothRev 0.000023729 s 0.00001488863998929446 s 1.59
slicing / JaXPipe / cpu / Primal 0.000007520179997300147 s 0.000008040859984248528 s 0.94
slicing / Jax / cpu / Primal 0.000006730639991019416 s 0.000007323159998122719 s 0.92
slicing / HLOOpt / cpu / Primal 0.00001012661999766351 s 0.000011351140028637018 s 0.89
slicing / PartOpt / cpu / Primal 0.000006404899993412983 s 0.0000070472600782522935 s 0.91
slicing / IPartOpt / cpu / Primal 0.000006238440003016876 s 0.000006619239975407254 s 0.94
slicing / DefOpt / cpu / Primal 0.000011079560010784917 s 0.000012038580007356358 s 0.92
slicing / IDefOpt / cpu / Primal 0.000006226900013643899 s 0.0000068220599860069344 s 0.91
slicing / JaXPipe / cpu / Forward 0.00001081058000181656 s 0.000011048219948861516 s 0.98
slicing / Jax / cpu / Forward 0.000010313619984572142 s 0.000011258000013185663 s 0.92
slicing / HLOOpt / cpu / Forward 0.000015031039999939822 s 0.000015424760022142436 s 0.97
slicing / PartOpt / cpu / Forward 0.000014974399998664012 s 0.000015206179996312133 s 0.98
slicing / IPartOpt / cpu / Forward 0.000009607760002836583 s 0.00001020377998429467 s 0.94
slicing / DefOpt / cpu / Forward 0.000014747959994565464 s 0.000015816619988981983 s 0.93
slicing / IDefOpt / cpu / Forward 0.000009787320004761569 s 0.000010371740027039776 s 0.94
slicing / JaXPipe / cpu / PreRev 0.000010868859997117395 s 0.000011388220018488935 s 0.95
slicing / JaXPipe / cpu / PostRev 0.000011041899988413206 s 0.000012067180014128098 s 0.92
slicing / JaXPipe / cpu / BothRev 0.000013056340001185164 s 0.000011459180004749214 s 1.14
slicing / Jax / cpu / BothRev 0.000011301960000764666 s 0.000011226620008528698 s 1.01
slicing / HLOOpt / cpu / PreRev 0.000010789840009692853 s 0.00001112885999646096 s 0.97
slicing / HLOOpt / cpu / PostRev 0.000010995919999459148 s 0.000011124339989692087 s 0.99
slicing / HLOOpt / cpu / BothRev 0.00001235998000311156 s 0.000012854619972131332 s 0.96
slicing / PartOpt / cpu / PreRev 0.000011128959999950894 s 0.000011010340012944652 s 1.01
slicing / PartOpt / cpu / PostRev 0.000011097220001374808 s 0.000011743560035029076 s 0.94
slicing / PartOpt / cpu / BothRev 0.0000107541600004879 s 0.000010728519982876606 s 1.00
slicing / IPartOpt / cpu / PreRev 0.000015612580002652977 s 0.000013666479962921586 s 1.14
slicing / IPartOpt / cpu / PostRev 0.00001113192000957497 s 0.00001228439999067632 s 0.91
slicing / IPartOpt / cpu / BothRev 0.000010803460006627577 s 0.00001136875999691256 s 0.95
slicing / DefOpt / cpu / PreRev 0.000010420919998068713 s 0.000011448680015746505 s 0.91
slicing / DefOpt / cpu / PostRev 0.000010505100003683764 s 0.000011924800000997493 s 0.88
slicing / DefOpt / cpu / BothRev 0.000010664220003491207 s 0.000011419660004321488 s 0.93
slicing / IDefOpt / cpu / PreRev 0.00000982672000191087 s 0.000011360259995853997 s 0.87
slicing / IDefOpt / cpu / PostRev 0.000011185039995780244 s 0.000011660339960144484 s 0.96
slicing / IDefOpt / cpu / BothRev 0.00001056673999528357 s 0.00001080735999494209 s 0.98
slicing / JaXPipe / cuda / Primal 0.0000019200000000000003 s 0.0000019200000000000003 s 1
slicing / Jax / cuda / Primal 0.0000019200000000000003 s 0.0000019200000000000003 s 1
slicing / HLOOpt / cuda / Primal 0.000001888 s 0.0000019200000000000003 s 0.98
slicing / PartOpt / cuda / Primal 0.0000019200000000000003 s 0.000001919 s 1.00
slicing / IPartOpt / cuda / Primal 0.0000019200000000000003 s 0.000001888 s 1.02
slicing / DefOpt / cuda / Primal 0.000001887 s 0.0000019200000000000003 s 0.98
slicing / IDefOpt / cuda / Primal 0.000001919 s 0.0000019200000000000003 s 1.00
slicing / JaXPipe / cuda / Forward 0.000010143 s 0.00001008 s 1.01
slicing / Jax / cuda / Forward 0.00000992 s 0.00001008 s 0.98
slicing / HLOOpt / cuda / Forward 0.000010049 s 0.000009887 s 1.02
slicing / PartOpt / cuda / Forward 0.00001008 s 0.000010464 s 0.96
slicing / IPartOpt / cuda / Forward 0.000009952 s 0.00001008 s 0.99
slicing / DefOpt / cuda / Forward 0.00001248 s 0.000010016 s 1.25
slicing / IDefOpt / cuda / Forward 0.00000976 s 0.000009632 s 1.01
slicing / JaXPipe / cuda / PreRev 0.00001232 s 0.000010111 s 1.22
slicing / JaXPipe / cuda / PostRev 0.000009984 s 0.0000104 s 0.96
slicing / JaXPipe / cuda / BothRev 0.000010432 s 0.000011392 s 0.92
slicing / Jax / cuda / BothRev 0.000009696 s 0.000010688 s 0.91
slicing / HLOOpt / cuda / PreRev 0.000010112 s 0.000010593 s 0.95
slicing / HLOOpt / cuda / PostRev 0.000009984 s 0.000009888 s 1.01
slicing / HLOOpt / cuda / BothRev 0.000010176 s 0.000011296 s 0.90
slicing / PartOpt / cuda / PreRev 0.000010432 s 0.000011455999999999998 s 0.91
slicing / PartOpt / cuda / PostRev 0.000010304 s 0.000012032 s 0.86
slicing / PartOpt / cuda / BothRev 0.000010112 s 0.000010144 s 1.00
slicing / IPartOpt / cuda / PreRev 0.000010048 s 0.000010304 s 0.98
slicing / IPartOpt / cuda / PostRev 0.000009984 s 0.000010559 s 0.95
slicing / IPartOpt / cuda / BothRev 0.000010048 s 0.000010752 s 0.93
slicing / DefOpt / cuda / PreRev 0.000010336 s 0.000010656 s 0.97
slicing / DefOpt / cuda / PostRev 0.000010112 s 0.000010112 s 1
slicing / DefOpt / cuda / BothRev 0.000010175 s 0.000010208 s 1.00
slicing / IDefOpt / cuda / PreRev 0.000012672 s 0.000009952 s 1.27
slicing / IDefOpt / cuda / PostRev 0.00001024 s 0.000010368 s 0.99
slicing / IDefOpt / cuda / BothRev 0.00001024 s 0.000010049 s 1.02
slicing / JaXPipe / tpu / Primal 9.582e-7 s 0.0000010259 s 0.93
slicing / Jax / tpu / Primal 9.86025e-7 s 9.6695e-7 s 1.02
slicing / HLOOpt / tpu / Primal 9.60925e-7 s 0.000001033725 s 0.93
slicing / PartOpt / tpu / Primal 9.7325e-7 s 9.66175e-7 s 1.01
slicing / IPartOpt / tpu / Primal 9.58575e-7 s 0.0000010305 s 0.93
slicing / DefOpt / tpu / Primal 9.732e-7 s 9.74725e-7 s 1.00
slicing / IDefOpt / tpu / Primal 9.59075e-7 s 0.00000102215 s 0.94
slicing / JaXPipe / tpu / Forward 0.000001403275 s 0.0000014117 s 0.99
slicing / Jax / tpu / Forward 0.00000140805 s 0.0000014743999999999998 s 0.95
slicing / HLOOpt / tpu / Forward 0.00000151105 s 0.0000015146 s 1.00
slicing / PartOpt / tpu / Forward 0.0000014276 s 0.000001492575 s 0.96
slicing / IPartOpt / tpu / Forward 0.00000151725 s 0.0000015134249999999998 s 1.00
slicing / DefOpt / tpu / Forward 0.00000143505 s 0.0000014924 s 0.96
slicing / IDefOpt / tpu / Forward 0.0000015118500000000002 s 0.000001516525 s 1.00
slicing / JaXPipe / tpu / PreRev 0.000002336375 s 0.0000025597 s 0.91
slicing / JaXPipe / tpu / PostRev 0.0000025186 s 0.00000251315 s 1.00
slicing / JaXPipe / tpu / BothRev 0.0000023565 s 0.000002578425 s 0.91
slicing / Jax / tpu / BothRev 0.0000025368 s 0.0000025517250000000005 s 0.99
slicing / HLOOpt / tpu / PreRev 0.000002350125 s 0.00000258015 s 0.91
slicing / HLOOpt / tpu / PostRev 0.0000025418000000000004 s 0.00000254145 s 1.00
slicing / HLOOpt / tpu / BothRev 0.00000234275 s 0.000002590375 s 0.90
slicing / PartOpt / tpu / PreRev 0.000002536875 s 0.00000254725 s 1.00
slicing / PartOpt / tpu / PostRev 0.000002343375 s 0.00000257835 s 0.91
slicing / PartOpt / tpu / BothRev 0.0000025353 s 0.000002541075 s 1.00
slicing / IPartOpt / tpu / PreRev 0.000002349525 s 0.0000025848 s 0.91
slicing / IPartOpt / tpu / PostRev 0.0000025238500000000004 s 0.000002543075 s 0.99
slicing / IPartOpt / tpu / BothRev 0.0000023561750000000004 s 0.0000025757 s 0.91
slicing / DefOpt / tpu / PreRev 0.0000025337 s 0.000002541775 s 1.00
slicing / DefOpt / tpu / PostRev 0.0000023544499999999995 s 0.000002586525 s 0.91
slicing / DefOpt / tpu / BothRev 0.00000253525 s 0.00000254005 s 1.00
slicing / IDefOpt / tpu / PreRev 0.0000023499 s 0.00000257795 s 0.91
slicing / IDefOpt / tpu / PostRev 0.000002529925 s 0.000002539375 s 1.00
slicing / IDefOpt / tpu / BothRev 0.0000023558 s 0.0000025849 s 0.91
slicing / JaXPipe / cpu / Primal 0.000012754 s 0.000008040859984248528 s 1.59
slicing / Jax / cpu / Primal 0.000012391 s 0.000007323159998122719 s 1.69
slicing / HLOOpt / cpu / Primal 0.000012459 s 0.000011351140028637018 s 1.10
slicing / PartOpt / cpu / Primal 0.000012657 s 0.0000070472600782522935 s 1.80
slicing / IPartOpt / cpu / Primal 0.000012264 s 0.000006619239975407254 s 1.85
slicing / DefOpt / cpu / Primal 0.000012303 s 0.000012038580007356358 s 1.02
slicing / IDefOpt / cpu / Primal 0.000012472 s 0.0000068220599860069344 s 1.83
slicing / JaXPipe / cpu / Forward 0.000017137 s 0.000011048219948861516 s 1.55
slicing / Jax / cpu / Forward 0.000016856 s 0.000011258000013185663 s 1.50
slicing / HLOOpt / cpu / Forward 0.000016376 s 0.000015424760022142436 s 1.06
slicing / PartOpt / cpu / Forward 0.00001665 s 0.000015206179996312133 s 1.09
slicing / IPartOpt / cpu / Forward 0.000016751 s 0.00001020377998429467 s 1.64
slicing / DefOpt / cpu / Forward 0.000017323 s 0.000015816619988981983 s 1.10
slicing / IDefOpt / cpu / Forward 0.000017163 s 0.000010371740027039776 s 1.65
slicing / JaXPipe / cpu / PreRev 0.000019151 s 0.000011388220018488935 s 1.68
slicing / JaXPipe / cpu / PostRev 0.000017515 s 0.000012067180014128098 s 1.45
slicing / JaXPipe / cpu / BothRev 0.000017788000000000003 s 0.000011459180004749214 s 1.55
slicing / Jax / cpu / BothRev 0.000017992 s 0.000011226620008528698 s 1.60
slicing / HLOOpt / cpu / PreRev 0.000017305 s 0.00001112885999646096 s 1.55
slicing / HLOOpt / cpu / PostRev 0.000018201 s 0.000011124339989692087 s 1.64
slicing / HLOOpt / cpu / BothRev 0.000018343 s 0.000012854619972131332 s 1.43
slicing / PartOpt / cpu / PreRev 0.000017597 s 0.000011010340012944652 s 1.60
slicing / PartOpt / cpu / PostRev 0.000018233 s 0.000011743560035029076 s 1.55
slicing / PartOpt / cpu / BothRev 0.000017501 s 0.000010728519982876606 s 1.63
slicing / IPartOpt / cpu / PreRev 0.000017315 s 0.000013666479962921586 s 1.27
slicing / IPartOpt / cpu / PostRev 0.000017864 s 0.00001228439999067632 s 1.45
slicing / IPartOpt / cpu / BothRev 0.000018466 s 0.00001136875999691256 s 1.62
slicing / DefOpt / cpu / PreRev 0.000017344 s 0.000011448680015746505 s 1.51
slicing / DefOpt / cpu / PostRev 0.000017502 s 0.000011924800000997493 s 1.47
slicing / DefOpt / cpu / BothRev 0.000018635 s 0.000011419660004321488 s 1.63
slicing / IDefOpt / cpu / PreRev 0.000017860999999999997 s 0.000011360259995853997 s 1.57
slicing / IDefOpt / cpu / PostRev 0.000018038 s 0.000011660339960144484 s 1.55
slicing / IDefOpt / cpu / BothRev 0.000018764 s 0.00001080735999494209 s 1.74
sum / JaXPipe / cpu / Primal 0.000009326759993655289 s 0.000009742220008774891 s 0.96
sum / Jax / cpu / Primal 0.000007966159998886723 s 0.00000903806000678742 s 0.88
sum / HLOOpt / cpu / Primal 0.000012604039998223016 s 0.000012662220033234917 s 1.00
sum / PartOpt / cpu / Primal 0.000008903919997464982 s 0.000008491360022162553 s 1.05
sum / IPartOpt / cpu / Primal 0.00000861625999732496 s 0.00000859147996379761 s 1.00
sum / DefOpt / cpu / Primal 0.000012823760009723628 s 0.000013167799979783012 s 0.97
sum / IDefOpt / cpu / Primal 0.000008508179998898413 s 0.000008690520016898518 s 0.98
sum / JaXPipe / cpu / Forward 0.000012012240003969057 s 0.000013288580057633223 s 0.90
sum / Jax / cpu / Forward 0.00001125527999647602 s 0.000012854360011260724 s 0.88
sum / HLOOpt / cpu / Forward 0.000016323160002684746 s 0.0000172800400378037 s 0.94
sum / PartOpt / cpu / Forward 0.000011959360003857 s 0.00001706791999822599 s 0.70
sum / IPartOpt / cpu / Forward 0.000011761040011606383 s 0.000012654980000661452 s 0.93
sum / DefOpt / cpu / Forward 0.00001700889999710853 s 0.000017553959951328578 s 0.97
sum / IDefOpt / cpu / Forward 0.00001165500000524844 s 0.000013098059989715692 s 0.89
sum / JaXPipe / cpu / PreRev 0.00001235705999761194 s 0.000012499379963628598 s 0.99
sum / JaXPipe / cpu / PostRev 0.000012491020004290475 s 0.000012826340016545146 s 0.97
sum / JaXPipe / cpu / BothRev 0.000015887439994912712 s 0.000012429079970388556 s 1.28
sum / Jax / cpu / BothRev 0.000012843900010466314 s 0.000012457939992600589 s 1.03
sum / HLOOpt / cpu / PreRev 0.0000120040600018001 s 0.000011955399986618432 s 1.00
sum / HLOOpt / cpu / PostRev 0.000011410200013415306 s 0.000016029180033001465 s 0.71
sum / HLOOpt / cpu / BothRev 0.000013808660000904636 s 0.00001411953999195248 s 0.98
sum / PartOpt / cpu / PreRev 0.000012086099995940458 s 0.000012572819996421458 s 0.96
sum / PartOpt / cpu / PostRev 0.000011975619991062558 s 0.000012195079962111776 s 0.98
sum / PartOpt / cpu / BothRev 0.00001142363999861118 s 0.000011735139933080064 s 0.97
sum / IPartOpt / cpu / PreRev 0.0000120522400061418 s 0.00001590705996932229 s 0.76
sum / IPartOpt / cpu / PostRev 0.000011596380002174555 s 0.000011663459990813863 s 0.99
sum / IPartOpt / cpu / BothRev 0.000012219540003570728 s 0.000011670320027405978 s 1.05
sum / DefOpt / cpu / PreRev 0.000011831439994693938 s 0.00001223856000251544 s 0.97
sum / DefOpt / cpu / PostRev 0.00001200267999593052 s 0.000012280459986868664 s 0.98
sum / DefOpt / cpu / BothRev 0.000011404140004742655 s 0.000011861679968205863 s 0.96
sum / IDefOpt / cpu / PreRev 0.000011652099985894892 s 0.00001202029996420606 s 0.97
sum / IDefOpt / cpu / PostRev 0.000012267099991731811 s 0.000011725960048352135 s 1.05
sum / IDefOpt / cpu / BothRev 0.000012170820009487216 s 0.000012039119974360802 s 1.01
sum / JaXPipe / cuda / Primal 0.00000208 s 0.000002143 s 0.97
sum / Jax / cuda / Primal 0.00000208 s 0.0000021120000000000003 s 0.98
sum / HLOOpt / cuda / Primal 0.00000208 s 0.0000021120000000000003 s 0.98
sum / PartOpt / cuda / Primal 0.00000208 s 0.0000021120000000000003 s 0.98
sum / IPartOpt / cuda / Primal 0.00000208 s 0.0000021120000000000003 s 0.98
sum / DefOpt / cuda / Primal 0.00000208 s 0.0000021120000000000003 s 0.98
sum / IDefOpt / cuda / Primal 0.000002079 s 0.000002143 s 0.97
sum / JaXPipe / cuda / Forward 0.000010464 s 0.000011007 s 0.95
sum / Jax / cuda / Forward 0.000010304 s 0.000015456 s 0.67
sum / HLOOpt / cuda / Forward 0.00001008 s 0.000010815 s 0.93
sum / PartOpt / cuda / Forward 0.000010592 s 0.000011967 s 0.89
sum / IPartOpt / cuda / Forward 0.0000104 s 0.0000104 s 1
sum / DefOpt / cuda / Forward 0.000010432 s 0.000010368 s 1.01
sum / IDefOpt / cuda / Forward 0.0000104 s 0.000011712 s 0.89
sum / JaXPipe / cuda / PreRev 0.00001008 s 0.000009888 s 1.02
sum / JaXPipe / cuda / PostRev 0.000010336 s 0.000010207 s 1.01
sum / JaXPipe / cuda / BothRev 0.0000096 s 0.000009952 s 0.96
sum / Jax / cuda / BothRev 0.00000992 s 0.00001008 s 0.98
sum / HLOOpt / cuda / PreRev 0.00000992 s 0.00000992 s 1
sum / HLOOpt / cuda / PostRev 0.000010047 s 0.000010017 s 1.00
sum / HLOOpt / cuda / BothRev 0.000010208 s 0.000009824 s 1.04
sum / PartOpt / cuda / PreRev 0.000010273 s 0.000009984 s 1.03
sum / PartOpt / cuda / PostRev 0.000010112 s 0.000010208 s 0.99
sum / PartOpt / cuda / BothRev 0.000010272 s 0.000010144 s 1.01
sum / IPartOpt / cuda / PreRev 0.000010016 s 0.000009952 s 1.01
sum / IPartOpt / cuda / PostRev 0.000009984 s 0.000010144 s 0.98
sum / IPartOpt / cuda / BothRev 0.00000992 s 0.000010176 s 0.97
sum / DefOpt / cuda / PreRev 0.000010112 s 0.000009952 s 1.02
sum / DefOpt / cuda / PostRev 0.000009728 s 0.000010176 s 0.96
sum / DefOpt / cuda / BothRev 0.000009984 s 0.000010016 s 1.00
sum / IDefOpt / cuda / PreRev 0.000010016 s 0.000009984 s 1.00
sum / IDefOpt / cuda / PostRev 0.000009952 s 0.000010272 s 0.97
sum / IDefOpt / cuda / BothRev 0.000010336 s 0.00001024 s 1.01
sum / JaXPipe / tpu / Primal 5.1665e-7 s 5.1055e-7 s 1.01
sum / Jax / tpu / Primal 5.565e-7 s 5.583250000000001e-7 s 1.00
sum / HLOOpt / tpu / Primal 5.26675e-7 s 5.2135e-7 s 1.01
sum / PartOpt / tpu / Primal 5.56575e-7 s 5.58275e-7 s 1.00
sum / IPartOpt / tpu / Primal 5.26975e-7 s 5.22425e-7 s 1.01
sum / DefOpt / tpu / Primal 5.566e-7 s 5.584499999999999e-7 s 1.00
sum / IDefOpt / tpu / Primal 5.26625e-7 s 5.213e-7 s 1.01
sum / JaXPipe / tpu / Forward 0.0000015475249999999998 s 0.0000015467249999999998 s 1.00
sum / Jax / tpu / Forward 0.0000015096 s 0.0000014976500000000002 s 1.01
sum / HLOOpt / tpu / Forward 0.00000153035 s 0.0000015402 s 0.99
sum / PartOpt / tpu / Forward 0.000001502225 s 0.0000014953 s 1.00
sum / IPartOpt / tpu / Forward 0.000001535175 s 0.000001534125 s 1.00
sum / DefOpt / tpu / Forward 0.000001507725 s 0.000001501775 s 1.00
sum / IDefOpt / tpu / Forward 0.000001538325 s 0.00000153315 s 1.00
sum / JaXPipe / tpu / PreRev 0.0000010021 s 0.00000105375 s 0.95
sum / JaXPipe / tpu / PostRev 0.0000010339 s 0.000001084475 s 0.95
sum / JaXPipe / tpu / BothRev 0.0000010031 s 0.000001061775 s 0.94
sum / Jax / tpu / BothRev 0.000001036125 s 0.000001087875 s 0.95
sum / HLOOpt / tpu / PreRev 0.00000101465 s 0.00000104835 s 0.97
sum / HLOOpt / tpu / PostRev 0.00000104015 s 0.000001083475 s 0.96
sum / HLOOpt / tpu / BothRev 0.00000101045 s 0.000001045575 s 0.97
sum / PartOpt / tpu / PreRev 0.0000010413500000000002 s 0.00000108545 s 0.96
sum / PartOpt / tpu / PostRev 0.0000010096 s 0.000001057975 s 0.95
sum / PartOpt / tpu / BothRev 0.0000010373000000000002 s 0.0000010851249999999998 s 0.96
sum / IPartOpt / tpu / PreRev 0.0000010097999999999998 s 0.0000010513 s 0.96
sum / IPartOpt / tpu / PostRev 0.00000103305 s 0.000001089675 s 0.95
sum / IPartOpt / tpu / BothRev 0.0000010066 s 0.000001061925 s 0.95
sum / DefOpt / tpu / PreRev 0.0000010381 s 0.000001097775 s 0.95
sum / DefOpt / tpu / PostRev 0.000001004925 s 0.00000105545 s 0.95
sum / DefOpt / tpu / BothRev 0.0000010304 s 0.0000010873249999999998 s 0.95
sum / IDefOpt / tpu / PreRev 0.000001000075 s 0.0000010545 s 0.95
sum / IDefOpt / tpu / PostRev 0.000001034525 s 0.000001090125 s 0.95
sum / IDefOpt / tpu / BothRev 0.000001000575 s 0.00000105155 s 0.95
sum / JaXPipe / cpu / Primal 0.000014732 s 0.000009742220008774891 s 1.51
sum / Jax / cpu / Primal 0.00001438 s 0.00000903806000678742 s 1.59
sum / HLOOpt / cpu / Primal 0.000014384 s 0.000012662220033234917 s 1.14
sum / PartOpt / cpu / Primal 0.000014534 s 0.000008491360022162553 s 1.71
sum / IPartOpt / cpu / Primal 0.000014378 s 0.00000859147996379761 s 1.67
sum / DefOpt / cpu / Primal 0.000014605 s 0.000013167799979783012 s 1.11
sum / IDefOpt / cpu / Primal 0.000014792 s 0.000008690520016898518 s 1.70
sum / JaXPipe / cpu / Forward 0.000020518 s 0.000013288580057633223 s 1.54
sum / Jax / cpu / Forward 0.000020765 s 0.000012854360011260724 s 1.62
sum / HLOOpt / cpu / Forward 0.000020066 s 0.0000172800400378037 s 1.16
sum / PartOpt / cpu / Forward 0.000020638 s 0.00001706791999822599 s 1.21
sum / IPartOpt / cpu / Forward 0.000019555 s 0.000012654980000661452 s 1.55
sum / DefOpt / cpu / Forward 0.000020433 s 0.000017553959951328578 s 1.16
sum / IDefOpt / cpu / Forward 0.000020119 s 0.000013098059989715692 s 1.54
sum / JaXPipe / cpu / PreRev 0.000019066 s 0.000012499379963628598 s 1.53
sum / JaXPipe / cpu / PostRev 0.000019634 s 0.000012826340016545146 s 1.53
sum / JaXPipe / cpu / BothRev 0.000019585000000000003 s 0.000012429079970388556 s 1.58
sum / Jax / cpu / BothRev 0.00001955 s 0.000012457939992600589 s 1.57
sum / HLOOpt / cpu / PreRev 0.000019234000000000003 s 0.000011955399986618432 s 1.61
sum / HLOOpt / cpu / PostRev 0.000019721 s 0.000016029180033001465 s 1.23
sum / HLOOpt / cpu / BothRev 0.000019752 s 0.00001411953999195248 s 1.40
sum / PartOpt / cpu / PreRev 0.000019082 s 0.000012572819996421458 s 1.52
sum / PartOpt / cpu / PostRev 0.000019777 s 0.000012195079962111776 s 1.62
sum / PartOpt / cpu / BothRev 0.00001971 s 0.000011735139933080064 s 1.68
sum / IPartOpt / cpu / PreRev 0.000019383 s 0.00001590705996932229 s 1.22
sum / IPartOpt / cpu / PostRev 0.000019888 s 0.000011663459990813863 s 1.71
sum / IPartOpt / cpu / BothRev 0.000019730000000000003 s 0.000011670320027405978 s 1.69
sum / DefOpt / cpu / PreRev 0.00001851 s 0.00001223856000251544 s 1.51
sum / DefOpt / cpu / PostRev 0.000019464 s 0.000012280459986868664 s 1.58
sum / DefOpt / cpu / BothRev 0.000019227 s 0.000011861679968205863 s 1.62
sum / IDefOpt / cpu / PreRev 0.000018478 s 0.00001202029996420606 s 1.54
sum / IDefOpt / cpu / PostRev 0.000019285 s 0.000011725960048352135 s 1.64
sum / IDefOpt / cpu / BothRev 0.000019605 s 0.000012039119974360802 s 1.63
value_and_grad / JaXPipe / cpu / Primal 0.000015838880005958345 s 0.00001745141998071631 s 0.91
value_and_grad / Jax / cpu / Primal 0.000015497940000841482 s 0.000015697619983257027 s 0.99
value_and_grad / HLOOpt / cpu / Primal 0.000014571780006917834 s 0.00001602679996722145 s 0.91
value_and_grad / PartOpt / cpu / Primal 0.000015079859992965796 s 0.000016012839987524787 s 0.94
value_and_grad / IPartOpt / cpu / Primal 0.0000154496400023163 s 0.000015338099992732167 s 1.01
value_and_grad / DefOpt / cpu / Primal 0.000015043379999042373 s 0.0000166041199281608 s 0.91
value_and_grad / IDefOpt / cpu / Primal 0.000014858900005947362 s 0.000015574700009892696 s 0.95
value_and_grad / JaXPipe / cuda / Primal 0.000033344 s 0.000033568 s 0.99
value_and_grad / Jax / cuda / Primal 0.000033888 s 0.000033696 s 1.01
value_and_grad / HLOOpt / cuda / Primal 0.000033728 s 0.000033696 s 1.00
value_and_grad / PartOpt / cuda / Primal 0.000033632 s 0.000033312 s 1.01
value_and_grad / IPartOpt / cuda / Primal 0.000033984 s 0.00003392 s 1.00
value_and_grad / DefOpt / cuda / Primal 0.000033664 s 0.000033472 s 1.01
value_and_grad / IDefOpt / cuda / Primal 0.000032416 s 0.00003344 s 0.97
value_and_grad / JaXPipe / tpu / Primal 0 s 0 s 1
value_and_grad / Jax / tpu / Primal 0 s 0 s 1
value_and_grad / HLOOpt / tpu / Primal 0 s 0 s 1
value_and_grad / PartOpt / tpu / Primal 0 s 0 s 1
value_and_grad / IPartOpt / tpu / Primal 0 s 0 s 1
value_and_grad / DefOpt / tpu / Primal 0 s 0 s 1
value_and_grad / IDefOpt / tpu / Primal 0 s 0 s 1
value_and_grad / JaXPipe / cpu / Primal 0.000024196 s 0.00001745141998071631 s 1.39
value_and_grad / Jax / cpu / Primal 0.000023369 s 0.000015697619983257027 s 1.49
value_and_grad / HLOOpt / cpu / Primal 0.000023537 s 0.00001602679996722145 s 1.47
value_and_grad / PartOpt / cpu / Primal 0.000023469 s 0.000016012839987524787 s 1.47
value_and_grad / IPartOpt / cpu / Primal 0.000023623 s 0.000015338099992732167 s 1.54
value_and_grad / DefOpt / cpu / Primal 0.00002318 s 0.0000166041199281608 s 1.40
value_and_grad / IDefOpt / cpu / Primal 0.000024123 s 0.000015574700009892696 s 1.55
jaxmd20 / JaXPipe / cuda / Primal 0.001514205 s 0.001519418 s 1.00
jaxmd20 / Jax / cuda / Primal 0.0014199969999999 s 0.001550144 s 0.92
jaxmd20 / HLOOpt / cuda / Primal 0.0011384939999999 s 0.001116446 s 1.02
jaxmd20 / PartOpt / cuda / Primal 0.0013158369999999 s 0.001325692 s 0.99
jaxmd20 / IPartOpt / cuda / Primal 0.001369948 s 0.001323068 s 1.04
jaxmd20 / DefOpt / cuda / Primal 0.000528863 s 0.000562847 s 0.94
jaxmd20 / IDefOpt / cuda / Primal 0.0004933429999999 s 0.000521662 s 0.95
jaxmd20 / JaXPipe / cuda / Forward 0.000816958 s 0.000818078 s 1.00
jaxmd20 / Jax / cuda / Forward 0.001778716 s 0.001826107 s 0.97
jaxmd20 / HLOOpt / cuda / Forward 0.000825054 s 0.000835102 s 0.99
jaxmd20 / PartOpt / cuda / Forward 0.000824191 s 0.000867135 s 0.95
jaxmd20 / IPartOpt / cuda / Forward 0.000845822 s 0.000834141 s 1.01
jaxmd20 / DefOpt / cuda / Forward 0.000819006 s 0.000830559 s 0.99
jaxmd20 / IDefOpt / cuda / Forward 0.000815422 s 0.000836286 s 0.98
jaxmd20 / JaXPipe / cuda / PreRev 0.001655678 s 0.001664188 s 0.99
jaxmd20 / JaXPipe / cuda / PostRev 0.005335381 s 0.005894961 s 0.91
jaxmd20 / JaXPipe / cuda / BothRev 0.0016737249999999 s 0.00167104 s 1.00
jaxmd20 / Jax / cuda / BothRev 0.00527202 s 0.005287989 s 1.00
jaxmd20 / HLOOpt / cuda / PreRev 0.00170166 s 0.001727004 s 0.99
jaxmd20 / HLOOpt / cuda / PostRev 0.005159988 s 0.005237779 s 0.99
jaxmd20 / HLOOpt / cuda / BothRev 0.001632412 s 0.001671901 s 0.98
jaxmd20 / PartOpt / cuda / PreRev 0.0016958359999999 s 0.001737211 s 0.98
jaxmd20 / PartOpt / cuda / PostRev 0.0053760519999999 s 0.005424818 s 0.99
jaxmd20 / PartOpt / cuda / BothRev 0.001676828 s 0.001691996 s 0.99
jaxmd20 / IPartOpt / cuda / PreRev 0.001702236 s 0.001741371 s 0.98
jaxmd20 / IPartOpt / cuda / PostRev 0.005356308 s 0.005383125 s 1.00
jaxmd20 / IPartOpt / cuda / BothRev 0.001631741 s 0.001743899 s 0.94
jaxmd20 / DefOpt / cuda / PreRev 0.001703548 s 0.001727099 s 0.99
jaxmd20 / DefOpt / cuda / PostRev 0.002708794 s 0.002745498 s 0.99
jaxmd20 / DefOpt / cuda / BothRev 0.00164582 s 0.001666843 s 0.99
jaxmd20 / IDefOpt / cuda / PreRev 0.0017023 s 0.0018710999999999 s 0.91
jaxmd20 / IDefOpt / cuda / PostRev 0.001977756 s 0.002012187 s 0.98
jaxmd20 / IDefOpt / cuda / BothRev 0.0016621079999999 s 0.00165782 s 1.00
jaxmd20 / JaXPipe / tpu / Primal 0.009288338125 s 0.009265228125 s 1.00
jaxmd20 / Jax / tpu / Primal 0.009269065625 s 0.009272055 s 1.00
jaxmd20 / HLOOpt / tpu / Primal 0.0091692 s 0.009155241875 s 1.00
jaxmd20 / PartOpt / tpu / Primal 0.009199598125 s 0.009206595625 s 1.00
jaxmd20 / IPartOpt / tpu / Primal 0.009203113125 s 0.009205539375 s 1.00
jaxmd20 / DefOpt / tpu / Primal 0.00874739875 s 0.00875223875 s 1.00
jaxmd20 / IDefOpt / tpu / Primal 0.008637930625 s 0.0086319706249999 s 1.00
jaxmd20 / JaXPipe / tpu / Forward 0.01726621125 s 0.01725431125 s 1.00
jaxmd20 / Jax / tpu / Forward 0.018737995 s 0.01872932625 s 1.00
jaxmd20 / HLOOpt / tpu / Forward 0.01723638625 s 0.01724062375 s 1.00
jaxmd20 / PartOpt / tpu / Forward 0.017268876875 s 0.01726465875 s 1.00
jaxmd20 / IPartOpt / tpu / Forward 0.01726452125 s 0.01725511375 s 1.00
jaxmd20 / DefOpt / tpu / Forward 0.017262946875 s 0.017267023125 s 1.00
jaxmd20 / IDefOpt / tpu / Forward 0.017265083125 s 0.017254851875 s 1.00
jaxmd20 / JaXPipe / tpu / PreRev 0.02535844625 s 0.0253632874999999 s 1.00
jaxmd20 / JaXPipe / tpu / PostRev 0.02187210125 s 0.02187114125 s 1.00
jaxmd20 / JaXPipe / tpu / BothRev 0.02537150875 s 0.0253652781249999 s 1.00
jaxmd20 / Jax / tpu / BothRev 0.021875314375 s 0.021876271875 s 1.00
jaxmd20 / HLOOpt / tpu / PreRev 0.025359616875 s 0.025353813125 s 1.00
jaxmd20 / HLOOpt / tpu / PostRev 0.02097529875 s 0.020961249375 s 1.00
jaxmd20 / HLOOpt / tpu / BothRev 0.025278398125 s 0.025266944375 s 1.00
jaxmd20 / PartOpt / tpu / PreRev 0.02536749875 s 0.0253546081249999 s 1.00
jaxmd20 / PartOpt / tpu / PostRev 0.021533745625 s 0.0215015887499999 s 1.00
jaxmd20 / PartOpt / tpu / BothRev 0.02527951875 s 0.02526319 s 1.00
jaxmd20 / IPartOpt / tpu / PreRev 0.0253673556249999 s 0.025354286875 s 1.00
jaxmd20 / IPartOpt / tpu / PostRev 0.021524795 s 0.021510863125 s 1.00
jaxmd20 / IPartOpt / tpu / BothRev 0.0252794025 s 0.02527126125 s 1.00
jaxmd20 / DefOpt / tpu / PreRev 0.0253659875 s 0.02534782875 s 1.00
jaxmd20 / DefOpt / tpu / PostRev 0.018918110625 s 0.0188936225 s 1.00
jaxmd20 / DefOpt / tpu / BothRev 0.025282743125 s 0.02526132 s 1.00
jaxmd20 / IDefOpt / tpu / PreRev 0.025365989375 s 0.025350525 s 1.00
jaxmd20 / IDefOpt / tpu / PostRev 0.018413318125 s 0.018375641875 s 1.00
jaxmd20 / IDefOpt / tpu / BothRev 0.025285955 s 0.025267529375 s 1.00
jaxmd40 / JaXPipe / cpu / Primal 0.087128366 s 0.073189519 s 1.19
jaxmd40 / Jax / cpu / Primal 0.062924865 s 0.062088033 s 1.01
jaxmd40 / HLOOpt / cpu / Primal 0.093542024 s 0.093305827 s 1.00
jaxmd40 / PartOpt / cpu / Primal 0.0719967099999999 s 0.071012165 s 1.01
jaxmd40 / IPartOpt / cpu / Primal 0.076234853 s 0.076429691 s 1.00
jaxmd40 / DefOpt / cpu / Primal 0.088710815 s 0.097710628 s 0.91
jaxmd40 / IDefOpt / cpu / Primal 0.089218281 s 0.094188508 s 0.95
jaxmd40 / JaXPipe / cpu / Forward 0.165594481 s 0.181294098 s 0.91
jaxmd40 / Jax / cpu / Forward 0.093084362 s 0.0968687699999999 s 0.96
jaxmd40 / HLOOpt / cpu / Forward 0.1701688969999999 s 0.169265531 s 1.01
jaxmd40 / PartOpt / cpu / Forward 0.166875254 s 0.171988099 s 0.97
jaxmd40 / IPartOpt / cpu / Forward 0.17398824 s 0.167338572 s 1.04
jaxmd40 / DefOpt / cpu / Forward 0.17601385 s 0.175491958 s 1.00
jaxmd40 / IDefOpt / cpu / Forward 0.174184795 s 0.178809637 s 0.97
jaxmd40 / JaXPipe / cpu / PreRev 0.239904073 s 0.2251418449999999 s 1.07
jaxmd40 / JaXPipe / cpu / PostRev 0.150545654 s 0.1506595829999999 s 1.00
jaxmd40 / JaXPipe / cpu / BothRev 0.2375609279999999 s 0.233851988 s 1.02
jaxmd40 / Jax / cpu / BothRev 0.150644459 s 0.148079004 s 1.02
jaxmd40 / HLOOpt / cpu / PreRev 0.235047131 s 0.234927513 s 1.00
jaxmd40 / HLOOpt / cpu / PostRev 0.191001227 s 0.185470716 s 1.03
jaxmd40 / HLOOpt / cpu / BothRev 0.25071463 s 0.2584409299999999 s 0.97
jaxmd40 / PartOpt / cpu / PreRev 0.225936246 s 0.22314567 s 1.01
jaxmd40 / PartOpt / cpu / PostRev 0.134630885 s 0.139577358 s 0.96
jaxmd40 / PartOpt / cpu / BothRev 0.258274218 s 0.249477685 s 1.04
jaxmd40 / IPartOpt / cpu / PreRev 0.249821856 s 0.229576528 s 1.09
jaxmd40 / IPartOpt / cpu / PostRev 0.132872998 s 0.1397846999999999 s 0.95
jaxmd40 / IPartOpt / cpu / BothRev 0.241774627 s 0.262264603 s 0.92
jaxmd40 / DefOpt / cpu / PreRev 0.24332245 s 0.2235982 s 1.09
jaxmd40 / DefOpt / cpu / PostRev 0.185615815 s 0.183224511 s 1.01
jaxmd40 / DefOpt / cpu / BothRev 0.235466411 s 0.258078644 s 0.91
jaxmd40 / IDefOpt / cpu / PreRev 0.2314958379999999 s 0.224068647 s 1.03
jaxmd40 / IDefOpt / cpu / PostRev 0.187687464 s 0.172570329 s 1.09
jaxmd40 / IDefOpt / cpu / BothRev 0.254376159 s 0.267373618 s 0.95
neuralgcm_v1/deterministic_2_8_deg_inner_steps_24_outer_steps_4 / JaXPipe / cuda / Primal 1.704326045 s 1.704129989 s 1.00
neuralgcm_v1/deterministic_2_8_deg_inner_steps_24_outer_steps_4 / Jax / cuda / Primal 1.706411658 s 1.706821368 s 1.00
neuralgcm_v1/deterministic_2_8_deg_inner_steps_24_outer_steps_4 / HLOOpt / cuda / Primal 1.718811061 s 1.717827882 s 1.00
neuralgcm_v1/deterministic_2_8_deg_inner_steps_24_outer_steps_4 / PartOpt / cuda / Primal 1.6996707510000002 s 1.699017653 s 1.00
neuralgcm_v1/deterministic_2_8_deg_inner_steps_24_outer_steps_4 / IPartOpt / cuda / Primal 1.697329531 s 1.6982513 s 1.00
neuralgcm_v1/deterministic_2_8_deg_inner_steps_24_outer_steps_4 / DefOpt / cuda / Primal 1.711860507 s 1.710406564 s 1.00
neuralgcm_v1/deterministic_2_8_deg_inner_steps_24_outer_steps_4 / IDefOpt / cuda / Primal 1.961061921 s 1.962014622 s 1.00
neuralgcm_v1/deterministic_2_8_deg_inner_steps_24_outer_steps_4 / JaXPipe / tpu / Primal 3.952643924375 s 4.019778363125 s 0.98
neuralgcm_v1/deterministic_2_8_deg_inner_steps_24_outer_steps_4 / Jax / tpu / Primal 3.03899281125 s 3.0385960193750003 s 1.00
neuralgcm_v1/deterministic_2_8_deg_inner_steps_24_outer_steps_4 / HLOOpt / tpu / Primal 3.12113594625 s 3.12114431125 s 1.00
neuralgcm_v1/deterministic_2_8_deg_inner_steps_24_outer_steps_4 / PartOpt / tpu / Primal 3.059245798125 s 3.058878809375 s 1.00
neuralgcm_v1/deterministic_2_8_deg_inner_steps_24_outer_steps_4 / IPartOpt / tpu / Primal 3.059230625 s 3.05891877625 s 1.00
neuralgcm_v1/deterministic_2_8_deg_inner_steps_24_outer_steps_4 / DefOpt / tpu / Primal 2.2635729500000004 s 2.2633443037500003 s 1.00
neuralgcm_v1/deterministic_2_8_deg_inner_steps_24_outer_steps_4 / IDefOpt / tpu / Primal 4.7431810775 s 4.7428916 s 1.00
neuralgcm_v1/deterministic_2_8_deg_inner_steps_2_outer_steps_2 / JaXPipe / cpu / Primal 6.441740369 s 6.305729684 s 1.02
neuralgcm_v1/deterministic_2_8_deg_inner_steps_2_outer_steps_2 / Jax / cpu / Primal 6.287252185000001 s 6.161372106 s 1.02
neuralgcm_v1/deterministic_2_8_deg_inner_steps_2_outer_steps_2 / HLOOpt / cpu / Primal 6.310646281 s 6.139615052 s 1.03
neuralgcm_v1/deterministic_2_8_deg_inner_steps_2_outer_steps_2 / PartOpt / cpu / Primal 6.440268734 s 6.266638879 s 1.03
neuralgcm_v1/deterministic_2_8_deg_inner_steps_2_outer_steps_2 / IPartOpt / cpu / Primal 6.328368245 s 6.203732882000001 s 1.02
neuralgcm_v1/deterministic_2_8_deg_inner_steps_2_outer_steps_2 / DefOpt / cpu / Primal 2.548313525 s 2.440213385 s 1.04
neuralgcm_v1/deterministic_2_8_deg_inner_steps_2_outer_steps_2 / IDefOpt / cpu / Primal 6.9867942990000005 s 6.919413947 s 1.01

This comment was automatically generated by workflow using github-action-benchmark.

@wsmoses wsmoses force-pushed the jm/recognize_multislice branch from 9fe0697 to 6a1ec6d Compare February 10, 2026 03:04
@wsmoses wsmoses merged commit f7ebe6c into main Feb 10, 2026
11 of 14 checks passed
@wsmoses wsmoses deleted the jm/recognize_multislice branch February 10, 2026 03:20
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

3 participants