Skip to content

Conversation

@ivanradanov
Copy link
Collaborator

No description provided.

@ivanradanov ivanradanov requested a review from wsmoses December 16, 2025 14:26
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: 16b7be2 Previous: 0d06322 Ratio
actmtch / JaXPipe / cuda / Primal 0.000002016 s 0.000002016 s 1
actmtch / Jax / cuda / Primal 0.000002016 s 0.000002016 s 1
actmtch / HLOOpt / cuda / Primal 0.000002016 s 0.000002016 s 1
actmtch / PartOpt / cuda / Primal 0.000002016 s 0.000002016 s 1
actmtch / IPartOpt / cuda / Primal 0.000002015 s 0.000002016 s 1.00
actmtch / DefOpt / cuda / Primal 0.000002015 s 0.000002016 s 1.00
actmtch / IDefOpt / cuda / Primal 0.000002015 s 0.000002015 s 1
actmtch / JaXPipe / cuda / Forward 0.00001008 s 0.000010944 s 0.92
actmtch / Jax / cuda / Forward 0.000010176 s 0.000010464 s 0.97
actmtch / HLOOpt / cuda / Forward 0.00001024 s 0.000011552 s 0.89
actmtch / PartOpt / cuda / Forward 0.000010016 s 0.000010912 s 0.92
actmtch / IPartOpt / cuda / Forward 0.000010304 s 0.000011520000000000002 s 0.89
actmtch / DefOpt / cuda / Forward 0.000010335 s 0.000011424 s 0.90
actmtch / IDefOpt / cuda / Forward 0.000010144 s 0.00001184 s 0.86
actmtch / JaXPipe / cuda / PreRev 0.000009824 s 0.000010624 s 0.92
actmtch / JaXPipe / cuda / PostRev 0.000010463 s 0.000010336 s 1.01
actmtch / JaXPipe / cuda / BothRev 0.0000104 s 0.0000104 s 1
actmtch / Jax / cuda / BothRev 0.000010016 s 0.000010688 s 0.94
actmtch / HLOOpt / cuda / PreRev 0.000010336 s 0.000010336 s 1
actmtch / HLOOpt / cuda / PostRev 0.000010207 s 0.000010208 s 1.00
actmtch / HLOOpt / cuda / BothRev 0.000010016 s 0.000010528 s 0.95
actmtch / PartOpt / cuda / PreRev 0.000010433 s 0.00001024 s 1.02
actmtch / PartOpt / cuda / PostRev 0.00001024 s 0.000010847 s 0.94
actmtch / PartOpt / cuda / BothRev 0.0000104 s 0.000010496 s 0.99
actmtch / IPartOpt / cuda / PreRev 0.000010336 s 0.000010592 s 0.98
actmtch / IPartOpt / cuda / PostRev 0.0000104 s 0.000010464 s 0.99
actmtch / IPartOpt / cuda / BothRev 0.0000104 s 0.000010591 s 0.98
actmtch / DefOpt / cuda / PreRev 0.000010527 s 0.000010689 s 0.98
actmtch / DefOpt / cuda / PostRev 0.000010144 s 0.000010304 s 0.98
actmtch / DefOpt / cuda / BothRev 0.000010112 s 0.000010208 s 0.99
actmtch / IDefOpt / cuda / PreRev 0.000010496 s 0.000010528 s 1.00
actmtch / IDefOpt / cuda / PostRev 0.000010208 s 0.000010208 s 1
actmtch / IDefOpt / cuda / BothRev 0.00001024 s 0.000010272 s 1.00
actmtch / JaXPipe / tpu / Primal 5.63325e-7 s 5.631e-7 s 1.00
actmtch / Jax / tpu / Primal 5.96775e-7 s 5.96675e-7 s 1.00
actmtch / HLOOpt / tpu / Primal 0.00000209765 s 0.000002098725 s 1.00
actmtch / PartOpt / tpu / Primal 5.96475e-7 s 5.9665e-7 s 1.00
actmtch / IPartOpt / tpu / Primal 5.52525e-7 s 5.52425e-7 s 1.00
actmtch / DefOpt / tpu / Primal 0.000002154225 s 0.000002163725 s 1.00
actmtch / IDefOpt / tpu / Primal 0.0000021087 s 0.00000210695 s 1.00
actmtch / JaXPipe / tpu / Forward 0.000003817525 s 0.000003815725 s 1.00
actmtch / Jax / tpu / Forward 0.000001209375 s 0.0000012058000000000002 s 1.00
actmtch / HLOOpt / tpu / Forward 0.0000039481500000000005 s 0.0000039318 s 1.00
actmtch / PartOpt / tpu / Forward 0.000003911225 s 0.000003918225 s 1.00
actmtch / IPartOpt / tpu / Forward 0.00000395565 s 0.000003930175 s 1.01
actmtch / DefOpt / tpu / Forward 0.0000039117 s 0.000003920199999999999 s 1.00
actmtch / IDefOpt / tpu / Forward 0.0000039437 s 0.000003930175 s 1.00
actmtch / JaXPipe / tpu / PreRev 0.00000347665 s 0.0000034626750000000003 s 1.00
actmtch / JaXPipe / tpu / PostRev 0.00000163625 s 0.000001652625 s 0.99
actmtch / JaXPipe / tpu / BothRev 0.00000347965 s 0.0000034682 s 1.00
actmtch / Jax / tpu / BothRev 0.00000163515 s 0.0000016323 s 1.00
actmtch / HLOOpt / tpu / PreRev 0.0000034804 s 0.000003474925 s 1.00
actmtch / HLOOpt / tpu / PostRev 0.0000034102 s 0.00000340775 s 1.00
actmtch / HLOOpt / tpu / BothRev 0.0000034808 s 0.000003469675 s 1.00
actmtch / PartOpt / tpu / PreRev 0.000003424 s 0.0000034048 s 1.01
actmtch / PartOpt / tpu / PostRev 0.000001593625 s 0.0000016085 s 0.99
actmtch / PartOpt / tpu / BothRev 0.000003418675 s 0.0000034112 s 1.00
actmtch / IPartOpt / tpu / PreRev 0.000003470125 s 0.0000034773 s 1.00
actmtch / IPartOpt / tpu / PostRev 0.0000016353750000000002 s 0.000001637625 s 1.00
actmtch / IPartOpt / tpu / BothRev 0.000003476225 s 0.0000034829500000000003 s 1.00
actmtch / DefOpt / tpu / PreRev 0.0000034158 s 0.00000341875 s 1.00
actmtch / DefOpt / tpu / PostRev 0.00000341495 s 0.00000341295 s 1.00
actmtch / DefOpt / tpu / BothRev 0.0000034144000000000003 s 0.00000342575 s 1.00
actmtch / IDefOpt / tpu / PreRev 0.00000348695 s 0.0000034794750000000003 s 1.00
actmtch / IDefOpt / tpu / PostRev 0.0000034130500000000003 s 0.0000034012 s 1.00
actmtch / IDefOpt / tpu / BothRev 0.000003473625 s 0.000003471725 s 1.00
actmtch / JaXPipe / cpu / Primal 0.000013372 s 0.00000702896004440845 s 1.90
actmtch / Jax / cpu / Primal 0.000013422 s 0.000007090619947121013 s 1.89
actmtch / HLOOpt / cpu / Primal 0.000013898 s 0.00000910397997358814 s 1.53
actmtch / PartOpt / cpu / Primal 0.000013384 s 0.000007699900033912854 s 1.74
actmtch / IPartOpt / cpu / Primal 0.000012885 s 0.000008126540051307529 s 1.59
actmtch / DefOpt / cpu / Primal 0.000013938 s 0.000008300439876620658 s 1.68
actmtch / IDefOpt / cpu / Primal 0.000014018 s 0.000008272720024251611 s 1.69
actmtch / JaXPipe / cpu / Forward 0.000019152 s 0.000012635160037461902 s 1.52
actmtch / Jax / cpu / Forward 0.000018323 s 0.000010485599959793037 s 1.75
actmtch / HLOOpt / cpu / Forward 0.00001908 s 0.000012656719991355204 s 1.51
actmtch / PartOpt / cpu / Forward 0.000019069 s 0.000011976340065302791 s 1.59
actmtch / IPartOpt / cpu / Forward 0.000018627 s 0.000012473439983295976 s 1.49
actmtch / DefOpt / cpu / Forward 0.000018997 s 0.000012138799938838929 s 1.56
actmtch / IDefOpt / cpu / Forward 0.000019386 s 0.000011991440060228342 s 1.62
actmtch / JaXPipe / cpu / PreRev 0.00001949 s 0.000012268239970580908 s 1.59
actmtch / JaXPipe / cpu / PostRev 0.000017723 s 0.000010976959965773856 s 1.61
actmtch / JaXPipe / cpu / BothRev 0.00001896 s 0.000012177059979876504 s 1.56
actmtch / Jax / cpu / BothRev 0.000017857000000000002 s 0.00001028339993354166 s 1.74
actmtch / HLOOpt / cpu / PreRev 0.000019524 s 0.00001290521991904825 s 1.51
actmtch / HLOOpt / cpu / PostRev 0.000019221 s 0.000014558360035152874 s 1.32
actmtch / HLOOpt / cpu / BothRev 0.000019206 s 0.000011904499988304451 s 1.61
actmtch / PartOpt / cpu / PreRev 0.00001939 s 0.000012542979893623851 s 1.55
actmtch / PartOpt / cpu / PostRev 0.000017353 s 0.000011923499969270778 s 1.46
actmtch / PartOpt / cpu / BothRev 0.000019404000000000003 s 0.000013031420112383784 s 1.49
actmtch / IPartOpt / cpu / PreRev 0.000019296 s 0.000012351440091151744 s 1.56
actmtch / IPartOpt / cpu / PostRev 0.000017395000000000002 s 0.000010943919951387216 s 1.59
actmtch / IPartOpt / cpu / BothRev 0.000019646 s 0.000012477419968490722 s 1.57
actmtch / DefOpt / cpu / PreRev 0.000019149 s 0.000012202039979456458 s 1.57
actmtch / DefOpt / cpu / PostRev 0.000019233 s 0.000012676960086537292 s 1.52
actmtch / DefOpt / cpu / BothRev 0.000019185 s 0.00001223014001880074 s 1.57
actmtch / IDefOpt / cpu / PreRev 0.000019215 s 0.00001243941997017828 s 1.54
actmtch / IDefOpt / cpu / PostRev 0.000018949 s 0.000012198299955343828 s 1.55
actmtch / IDefOpt / cpu / BothRev 0.000019742 s 0.000011905400006071432 s 1.66
add_one / JaXPipe / cuda / Primal 0.0000019200000000000003 s 0.0000019200000000000003 s 1
add_one / Jax / cuda / Primal 0.0000019200000000000003 s 0.0000019200000000000003 s 1
add_one / HLOOpt / cuda / Primal 0.0000019200000000000003 s 0.0000019200000000000003 s 1
add_one / PartOpt / cuda / Primal 0.0000019200000000000003 s 0.0000019200000000000003 s 1
add_one / IPartOpt / cuda / Primal 0.0000019200000000000003 s 0.0000019200000000000003 s 1
add_one / DefOpt / cuda / Primal 0.0000019200000000000003 s 0.0000019200000000000003 s 1
add_one / IDefOpt / cuda / Primal 0.0000019200000000000003 s 0.0000019200000000000003 s 1
add_one / JaXPipe / cuda / Forward 0.000010304 s 0.000010272 s 1.00
add_one / Jax / cuda / Forward 0.000010432 s 0.000010272 s 1.02
add_one / HLOOpt / cuda / Forward 0.000010336 s 0.000010336 s 1
add_one / PartOpt / cuda / Forward 0.000010271 s 0.000010112 s 1.02
add_one / IPartOpt / cuda / Forward 0.00001024 s 0.000010272 s 1.00
add_one / DefOpt / cuda / Forward 0.000009984 s 0.000010304 s 0.97
add_one / IDefOpt / cuda / Forward 0.00001008 s 0.000010016 s 1.01
add_one / JaXPipe / cuda / PreRev 0.000025888 s 0.000025568 s 1.01
add_one / JaXPipe / cuda / PostRev 0.000025536 s 0.000025023 s 1.02
add_one / JaXPipe / cuda / BothRev 0.000025696 s 0.000025056 s 1.03
add_one / Jax / cuda / BothRev 0.000025888 s 0.000025312 s 1.02
add_one / HLOOpt / cuda / PreRev 0.00002576 s 0.000025216 s 1.02
add_one / HLOOpt / cuda / PostRev 0.000025632 s 0.000024768 s 1.03
add_one / HLOOpt / cuda / BothRev 0.000025408 s 0.000024992 s 1.02
add_one / PartOpt / cuda / PreRev 0.000025569 s 0.000025408 s 1.01
add_one / PartOpt / cuda / PostRev 0.000024992 s 0.000024896 s 1.00
add_one / PartOpt / cuda / BothRev 0.000025729 s 0.000025087 s 1.03
add_one / IPartOpt / cuda / PreRev 0.00002544 s 0.000025376 s 1.00
add_one / IPartOpt / cuda / PostRev 0.000025472000000000003 s 0.000025408 s 1.00
add_one / IPartOpt / cuda / BothRev 0.000024448 s 0.000025312 s 0.97
add_one / DefOpt / cuda / PreRev 0.000025504 s 0.000025824 s 0.99
add_one / DefOpt / cuda / PostRev 0.000026144 s 0.0000256 s 1.02
add_one / DefOpt / cuda / BothRev 0.000025504 s 0.000026368 s 0.97
add_one / IDefOpt / cuda / PreRev 0.000025184 s 0.000025471 s 0.99
add_one / IDefOpt / cuda / PostRev 0.00002592 s 0.00002544 s 1.02
add_one / IDefOpt / cuda / BothRev 0.000025664 s 0.000025792 s 1.00
add_one / JaXPipe / tpu / Primal 0.0000014295749999999998 s 0.0000014340249999999998 s 1.00
add_one / Jax / tpu / Primal 0.000001404625 s 0.000001401425 s 1.00
add_one / HLOOpt / tpu / Primal 0.0000014306 s 0.000001425525 s 1.00
add_one / PartOpt / tpu / Primal 0.0000014030000000000002 s 0.000001415625 s 0.99
add_one / IPartOpt / tpu / Primal 0.0000014320250000000005 s 0.00000142825 s 1.00
add_one / DefOpt / tpu / Primal 0.0000014068 s 0.0000014157 s 0.99
add_one / IDefOpt / tpu / Primal 0.000001429825 s 0.000001426675 s 1.00
add_one / JaXPipe / tpu / Forward 0.000001856925 s 0.0000018621 s 1.00
add_one / Jax / tpu / Forward 0.00000184955 s 0.00000184395 s 1.00
add_one / HLOOpt / tpu / Forward 0.0000018464 s 0.000001853375 s 1.00
add_one / PartOpt / tpu / Forward 0.000001838275 s 0.000001845425 s 1.00
add_one / IPartOpt / tpu / Forward 0.00000184925 s 0.000001850575 s 1.00
add_one / DefOpt / tpu / Forward 0.00000184045 s 0.00000185095 s 0.99
add_one / IDefOpt / tpu / Forward 0.000001851225 s 0.00000184935 s 1.00
add_one / JaXPipe / tpu / PreRev 0.0000022357 s 0.000002234425 s 1.00
add_one / JaXPipe / tpu / PostRev 0.00000224335 s 0.000002245875 s 1.00
add_one / JaXPipe / tpu / BothRev 0.000002245675 s 0.000002231825 s 1.01
add_one / Jax / tpu / BothRev 0.000002232575 s 0.000002238975 s 1.00
add_one / HLOOpt / tpu / PreRev 0.00000223485 s 0.000002237225 s 1.00
add_one / HLOOpt / tpu / PostRev 0.0000022358 s 0.00000223695 s 1.00
add_one / HLOOpt / tpu / BothRev 0.000002238325 s 0.0000022443250000000003 s 1.00
add_one / PartOpt / tpu / PreRev 0.000002238525 s 0.0000022425750000000003 s 1.00
add_one / PartOpt / tpu / PostRev 0.000002236525 s 0.00000224105 s 1.00
add_one / PartOpt / tpu / BothRev 0.0000022431 s 0.0000022365750000000003 s 1.00
add_one / IPartOpt / tpu / PreRev 0.0000022303 s 0.000002238025 s 1.00
add_one / IPartOpt / tpu / PostRev 0.0000022403500000000005 s 0.000002244225 s 1.00
add_one / IPartOpt / tpu / BothRev 0.00000223075 s 0.00000223555 s 1.00
add_one / DefOpt / tpu / PreRev 0.00000224955 s 0.000002244225 s 1.00
add_one / DefOpt / tpu / PostRev 0.00000223095 s 0.0000022363750000000003 s 1.00
add_one / DefOpt / tpu / BothRev 0.000002254725 s 0.0000022566750000000003 s 1.00
add_one / IDefOpt / tpu / PreRev 0.000002235925 s 0.0000022435250000000003 s 1.00
add_one / IDefOpt / tpu / PostRev 0.0000022467 s 0.000002242425 s 1.00
add_one / IDefOpt / tpu / BothRev 0.00000223015 s 0.000002233975 s 1.00
add_one / JaXPipe / cpu / Primal 0.000013385 s 0.000007482279979740269 s 1.79
add_one / Jax / cpu / Primal 0.000013297 s 0.000007295579998753965 s 1.82
add_one / HLOOpt / cpu / Primal 0.000012982 s 0.0000072367599568678995 s 1.79
add_one / PartOpt / cpu / Primal 0.000012878 s 0.000006717680007568561 s 1.92
add_one / IPartOpt / cpu / Primal 0.00001299 s 0.000007430179957736982 s 1.75
add_one / DefOpt / cpu / Primal 0.000013048 s 0.000006753920042683603 s 1.93
add_one / IDefOpt / cpu / Primal 0.000012777 s 0.000007092920041031902 s 1.80
add_one / JaXPipe / cpu / Forward 0.000017578 s 0.000010866239972529 s 1.62
add_one / Jax / cpu / Forward 0.000017602 s 0.000011250040042796172 s 1.56
add_one / HLOOpt / cpu / Forward 0.000017697 s 0.000011172939975949704 s 1.58
add_one / PartOpt / cpu / Forward 0.000017697 s 0.000011107540030934617 s 1.59
add_one / IPartOpt / cpu / Forward 0.000017434999999999998 s 0.000010974180022458312 s 1.59
add_one / DefOpt / cpu / Forward 0.000017402 s 0.000011392880023777253 s 1.53
add_one / IDefOpt / cpu / Forward 0.000017668 s 0.00001110596007492859 s 1.59
add_one / JaXPipe / cpu / PreRev 0.000019806 s 0.000013583340059994953 s 1.46
add_one / JaXPipe / cpu / PostRev 0.000019673 s 0.00001288485997065436 s 1.53
add_one / JaXPipe / cpu / BothRev 0.000019475 s 0.000013069859996903689 s 1.49
add_one / Jax / cpu / BothRev 0.000019681 s 0.000012902259913971648 s 1.53
add_one / HLOOpt / cpu / PreRev 0.000019523 s 0.000012942099983774824 s 1.51
add_one / HLOOpt / cpu / PostRev 0.00001981 s 0.000014932679914636537 s 1.33
add_one / HLOOpt / cpu / BothRev 0.000019556 s 0.000012773839953297284 s 1.53
add_one / PartOpt / cpu / PreRev 0.000019586 s 0.00001311142006670707 s 1.49
add_one / PartOpt / cpu / PostRev 0.000019508 s 0.000012460100024327402 s 1.57
add_one / PartOpt / cpu / BothRev 0.000019421 s 0.000013669600029970752 s 1.42
add_one / IPartOpt / cpu / PreRev 0.000019492 s 0.000012843979930039496 s 1.52
add_one / IPartOpt / cpu / PostRev 0.000019463 s 0.00001249579998329864 s 1.56
add_one / IPartOpt / cpu / BothRev 0.000019325 s 0.000012642160054383569 s 1.53
add_one / DefOpt / cpu / PreRev 0.0000195 s 0.000012758939992636445 s 1.53
add_one / DefOpt / cpu / PostRev 0.000019511 s 0.000012898780059913406 s 1.51
add_one / DefOpt / cpu / BothRev 0.000019662 s 0.000012417840061971218 s 1.58
add_one / IDefOpt / cpu / PreRev 0.000019662 s 0.000013013019943173276 s 1.51
add_one / IDefOpt / cpu / PostRev 0.000019615 s 0.000012688639926636823 s 1.55
add_one / IDefOpt / cpu / BothRev 0.000019683 s 0.000012303200019232463 s 1.60
add_two / JaXPipe / cuda / Primal 0.0000019200000000000003 s 0.0000019200000000000003 s 1
add_two / Jax / cuda / Primal 0.0000019200000000000003 s 0.0000019200000000000003 s 1
add_two / HLOOpt / cuda / Primal 0.0000019200000000000003 s 0.0000019200000000000003 s 1
add_two / PartOpt / cuda / Primal 0.0000019200000000000003 s 0.0000019200000000000003 s 1
add_two / IPartOpt / cuda / Primal 0.0000019200000000000003 s 0.0000019200000000000003 s 1
add_two / DefOpt / cuda / Primal 0.0000019200000000000003 s 0.0000019200000000000003 s 1
add_two / IDefOpt / cuda / Primal 0.000001919 s 0.0000019200000000000003 s 1.00
add_two / JaXPipe / cuda / Forward 0.000009984 s 0.00001024 s 0.97
add_two / Jax / cuda / Forward 0.000010048 s 0.000010048 s 1
add_two / HLOOpt / cuda / Forward 0.000010144 s 0.000009631 s 1.05
add_two / PartOpt / cuda / Forward 0.000009888 s 0.000010112 s 0.98
add_two / IPartOpt / cuda / Forward 0.000009952 s 0.000010143 s 0.98
add_two / DefOpt / cuda / Forward 0.00001008 s 0.000010176 s 0.99
add_two / IDefOpt / cuda / Forward 0.000015296 s 0.000009952 s 1.54
add_two / JaXPipe / cuda / PreRev 0.000032448 s 0.000031873 s 1.02
add_two / JaXPipe / cuda / PostRev 0.000032096 s 0.000032576 s 0.99
add_two / JaXPipe / cuda / BothRev 0.000032224 s 0.000033024 s 0.98
add_two / Jax / cuda / BothRev 0.000033088 s 0.000033152000000000004 s 1.00
add_two / HLOOpt / cuda / PreRev 0.000032993 s 0.000033343 s 0.99
add_two / HLOOpt / cuda / PostRev 0.000031968 s 0.000032606999999999995 s 0.98
add_two / HLOOpt / cuda / BothRev 0.000032992 s 0.000032928 s 1.00
add_two / PartOpt / cuda / PreRev 0.000032544 s 0.000032704 s 1.00
add_two / PartOpt / cuda / PostRev 0.000031776 s 0.000032767999999999995 s 0.97
add_two / PartOpt / cuda / BothRev 0.000032544 s 0.000032608 s 1.00
add_two / IPartOpt / cuda / PreRev 0.000032385 s 0.000033024 s 0.98
add_two / IPartOpt / cuda / PostRev 0.000032032 s 0.000033696 s 0.95
add_two / IPartOpt / cuda / BothRev 0.000032608 s 0.000033504 s 0.97
add_two / DefOpt / cuda / PreRev 0.000033536000000000006 s 0.00003344 s 1.00
add_two / DefOpt / cuda / PostRev 0.000032064 s 0.000033536000000000006 s 0.96
add_two / DefOpt / cuda / BothRev 0.000032033 s 0.000033408 s 0.96
add_two / IDefOpt / cuda / PreRev 0.000033632 s 0.000033024 s 1.02
add_two / IDefOpt / cuda / PostRev 0.000032896000000000005 s 0.000032736 s 1.00
add_two / IDefOpt / cuda / BothRev 0.000032608 s 0.000033376 s 0.98
add_two / JaXPipe / tpu / Primal 0.000001438975 s 0.0000014327 s 1.00
add_two / Jax / tpu / Primal 0.0000014689 s 0.00000148525 s 0.99
add_two / HLOOpt / tpu / Primal 0.00000143015 s 0.000001432125 s 1.00
add_two / PartOpt / tpu / Primal 0.000001474925 s 0.0000014746000000000002 s 1.00
add_two / IPartOpt / tpu / Primal 0.000001432625 s 0.000001427125 s 1.00
add_two / DefOpt / tpu / Primal 0.0000014809 s 0.000001486675 s 1.00
add_two / IDefOpt / tpu / Primal 0.0000014341999999999998 s 0.000001434 s 1.00
add_two / JaXPipe / tpu / Forward 0.000001823225 s 0.000001838025 s 0.99
add_two / Jax / tpu / Forward 0.000001829025 s 0.000001850425 s 0.99
add_two / HLOOpt / tpu / Forward 0.0000018356 s 0.000001825675 s 1.01
add_two / PartOpt / tpu / Forward 0.000001820275 s 0.00000183435 s 0.99
add_two / IPartOpt / tpu / Forward 0.0000018284 s 0.00000182775 s 1.00
add_two / DefOpt / tpu / Forward 0.00000182515 s 0.000001821225 s 1.00
add_two / IDefOpt / tpu / Forward 0.00000182685 s 0.000001830125 s 1.00
add_two / JaXPipe / tpu / PreRev 0.0000028352 s 0.000002833475 s 1.00
add_two / JaXPipe / tpu / PostRev 0.00000274935 s 0.000002741825 s 1.00
add_two / JaXPipe / tpu / BothRev 0.00000282795 s 0.0000028442 s 0.99
add_two / Jax / tpu / BothRev 0.0000027508500000000005 s 0.0000027538750000000004 s 1.00
add_two / HLOOpt / tpu / PreRev 0.000002834325 s 0.00000282865 s 1.00
add_two / HLOOpt / tpu / PostRev 0.000002744625 s 0.00000275385 s 1.00
add_two / HLOOpt / tpu / BothRev 0.00000283115 s 0.000002856775 s 0.99
add_two / PartOpt / tpu / PreRev 0.00000275215 s 0.0000027495750000000003 s 1.00
add_two / PartOpt / tpu / PostRev 0.0000028335 s 0.0000028459 s 1.00
add_two / PartOpt / tpu / BothRev 0.00000274645 s 0.00000275325 s 1.00
add_two / IPartOpt / tpu / PreRev 0.000002841475 s 0.00000283295 s 1.00
add_two / IPartOpt / tpu / PostRev 0.00000275105 s 0.000002750475 s 1.00
add_two / IPartOpt / tpu / BothRev 0.0000028461999999999995 s 0.0000028408 s 1.00
add_two / DefOpt / tpu / PreRev 0.0000027508500000000005 s 0.0000027521250000000004 s 1.00
add_two / DefOpt / tpu / PostRev 0.000002838825 s 0.00000284555 s 1.00
add_two / DefOpt / tpu / BothRev 0.0000027479500000000005 s 0.000002756675 s 1.00
add_two / IDefOpt / tpu / PreRev 0.000002836525 s 0.0000028440500000000003 s 1.00
add_two / IDefOpt / tpu / PostRev 0.000002750825 s 0.000002742525 s 1.00
add_two / IDefOpt / tpu / BothRev 0.0000028325 s 0.00000283935 s 1.00
add_two / JaXPipe / cpu / Primal 0.00001378 s 0.000007603080048284028 s 1.81
add_two / Jax / cpu / Primal 0.00001369 s 0.000007265580097737256 s 1.88
add_two / HLOOpt / cpu / Primal 0.000013507999999999998 s 0.000007252379946294241 s 1.86
add_two / PartOpt / cpu / Primal 0.00001352 s 0.000007664579989068443 s 1.76
add_two / IPartOpt / cpu / Primal 0.000013443 s 0.000007609659987792838 s 1.77
add_two / DefOpt / cpu / Primal 0.000013431 s 0.000007436000050802249 s 1.81
add_two / IDefOpt / cpu / Primal 0.000013241 s 0.000007664840049983468 s 1.73
add_two / JaXPipe / cpu / Forward 0.000018459 s 0.000011162979917571648 s 1.65
add_two / Jax / cpu / Forward 0.00001794 s 0.000011393439945095453 s 1.57
add_two / HLOOpt / cpu / Forward 0.000018044 s 0.000011876660046254983 s 1.52
add_two / PartOpt / cpu / Forward 0.000017726 s 0.000011154319981869775 s 1.59
add_two / IPartOpt / cpu / Forward 0.000017871999999999998 s 0.000011239279920118862 s 1.59
add_two / DefOpt / cpu / Forward 0.000018431 s 0.000011385859997972148 s 1.62
add_two / IDefOpt / cpu / Forward 0.000018132 s 0.000011184240047441565 s 1.62
add_two / JaXPipe / cpu / PreRev 0.000023186 s 0.000015534419944742694 s 1.49
add_two / JaXPipe / cpu / PostRev 0.000023057 s 0.000015442319981957552 s 1.49
add_two / JaXPipe / cpu / BothRev 0.000023192 s 0.00001556208002511994 s 1.49
add_two / Jax / cpu / BothRev 0.000023047 s 0.00001551052004288067 s 1.49
add_two / HLOOpt / cpu / PreRev 0.000022954 s 0.00001560625998536125 s 1.47
add_two / HLOOpt / cpu / PostRev 0.000023582 s 0.000017657759926805738 s 1.34
add_two / HLOOpt / cpu / BothRev 0.000023197 s 0.00001540101999125909 s 1.51
add_two / PartOpt / cpu / PreRev 0.000022731 s 0.000015437360052601433 s 1.47
add_two / PartOpt / cpu / PostRev 0.000023254 s 0.000015096120041562244 s 1.54
add_two / PartOpt / cpu / BothRev 0.000023164 s 0.00001628869993510307 s 1.42
add_two / IPartOpt / cpu / PreRev 0.000022821 s 0.000015847779923205963 s 1.44
add_two / IPartOpt / cpu / PostRev 0.000023167 s 0.000015286219950212397 s 1.52
add_two / IPartOpt / cpu / BothRev 0.0000233 s 0.000015069540022523144 s 1.55
add_two / DefOpt / cpu / PreRev 0.000023057 s 0.000015951179975672857 s 1.45
add_two / DefOpt / cpu / PostRev 0.000022896 s 0.000014918000033503633 s 1.53
add_two / DefOpt / cpu / BothRev 0.000023424 s 0.000015067820040712831 s 1.55
add_two / IDefOpt / cpu / PreRev 0.000022848 s 0.000015652199981559533 s 1.46
add_two / IDefOpt / cpu / PostRev 0.000023058 s 0.000014937559899408373 s 1.54
add_two / IDefOpt / cpu / BothRev 0.000022987 s 0.00001542553998660878 s 1.49
cache / JaXPipe / cuda / Primal 0.000002304 s 0.000002303 s 1.00
cache / Jax / cuda / Primal 0.000002304 s 0.000002303 s 1.00
cache / HLOOpt / cuda / Primal 0.000002207 s 0.00000224 s 0.99
cache / PartOpt / cuda / Primal 0.00000224 s 0.00000224 s 1
cache / IPartOpt / cuda / Primal 0.000002303 s 0.000002303 s 1
cache / DefOpt / cuda / Primal 0.00000224 s 0.000002272 s 0.99
cache / IDefOpt / cuda / Primal 0.000002208 s 0.000002208 s 1
cache / JaXPipe / cuda / Forward 0.000002335 s 0.000002303 s 1.01
cache / Jax / cuda / Forward 0.000002336 s 0.000002304 s 1.01
cache / HLOOpt / cuda / Forward 0.000002304 s 0.000002335 s 0.99
cache / PartOpt / cuda / Forward 0.000002335 s 0.000002335 s 1
cache / IPartOpt / cuda / Forward 0.000002304 s 0.000002336 s 0.99
cache / DefOpt / cuda / Forward 0.000002273 s 0.000002272 s 1.00
cache / IDefOpt / cuda / Forward 0.000002335 s 0.000002335 s 1
cache / JaXPipe / cuda / PreRev 0.000011264 s 0.00001072 s 1.05
cache / JaXPipe / cuda / PostRev 0.000010975 s 0.000010848 s 1.01
cache / JaXPipe / cuda / BothRev 0.000010944 s 0.000010784 s 1.01
cache / Jax / cuda / BothRev 0.00001088 s 0.000011008 s 0.99
cache / HLOOpt / cuda / PreRev 0.000013472 s 0.000013471 s 1.00
cache / HLOOpt / cuda / PostRev 0.000013408 s 0.000013408 s 1
cache / HLOOpt / cuda / BothRev 0.000013472 s 0.000013472 s 1
cache / PartOpt / cuda / PreRev 0.000011104 s 0.000011104 s 1
cache / PartOpt / cuda / PostRev 0.000010688 s 0.000010911 s 0.98
cache / PartOpt / cuda / BothRev 0.000010848 s 0.000010976 s 0.99
cache / IPartOpt / cuda / PreRev 0.000010976 s 0.000010304 s 1.07
cache / IPartOpt / cuda / PostRev 0.00001104 s 0.000010528 s 1.05
cache / IPartOpt / cuda / BothRev 0.000011297 s 0.000011008 s 1.03
cache / DefOpt / cuda / PreRev 0.000011008 s 0.00001088 s 1.01
cache / DefOpt / cuda / PostRev 0.000010816 s 0.000010912 s 0.99
cache / DefOpt / cuda / BothRev 0.000011169 s 0.000010912 s 1.02
cache / IDefOpt / cuda / PreRev 0.00001072 s 0.000011008 s 0.97
cache / IDefOpt / cuda / PostRev 0.000011008 s 0.000011136 s 0.99
cache / IDefOpt / cuda / BothRev 0.000010688 s 0.000011296 s 0.95
cache / JaXPipe / tpu / Primal 0.0000024818 s 0.00000247055 s 1.00
cache / Jax / tpu / Primal 0.000002462625 s 0.000002459375 s 1.00
cache / HLOOpt / tpu / Primal 0.000002467375 s 0.000002470125 s 1.00
cache / PartOpt / tpu / Primal 0.0000024736 s 0.000002456775 s 1.01
cache / IPartOpt / tpu / Primal 0.00000245645 s 0.00000246025 s 1.00
cache / DefOpt / tpu / Primal 0.0000024712750000000003 s 0.0000024752 s 1.00
cache / IDefOpt / tpu / Primal 0.000002458875 s 0.000002482 s 0.99
cache / JaXPipe / tpu / Forward 0.000003554225 s 0.000003567575 s 1.00
cache / Jax / tpu / Forward 0.000003529175 s 0.0000035388 s 1.00
cache / HLOOpt / tpu / Forward 0.0000035462 s 0.00000355415 s 1.00
cache / PartOpt / tpu / Forward 0.0000035356 s 0.0000035287 s 1.00
cache / IPartOpt / tpu / Forward 0.000003544775 s 0.0000035521000000000003 s 1.00
cache / DefOpt / tpu / Forward 0.000003523875 s 0.000003546825 s 0.99
cache / IDefOpt / tpu / Forward 0.0000035322500000000004 s 0.0000035494750000000004 s 1.00
cache / JaXPipe / tpu / PreRev 0.0000049448 s 0.0000049307 s 1.00
cache / JaXPipe / tpu / PostRev 0.000004956975 s 0.000004935399999999999 s 1.00
cache / JaXPipe / tpu / BothRev 0.000004976775 s 0.000004954025 s 1.00
cache / Jax / tpu / BothRev 0.00000498915 s 0.000004991775 s 1.00
cache / HLOOpt / tpu / PreRev 0.0000039467 s 0.000003940575 s 1.00
cache / HLOOpt / tpu / PostRev 0.000004124050000000001 s 0.0000041419 s 1.00
cache / HLOOpt / tpu / BothRev 0.000003937574999999999 s 0.0000039466 s 1.00
cache / PartOpt / tpu / PreRev 0.0000049567 s 0.00000498505 s 0.99
cache / PartOpt / tpu / PostRev 0.00000496865 s 0.0000049616 s 1.00
cache / PartOpt / tpu / BothRev 0.00000496925 s 0.000004985 s 1.00
cache / IPartOpt / tpu / PreRev 0.0000049759 s 0.00000495935 s 1.00
cache / IPartOpt / tpu / PostRev 0.000004976175 s 0.000004985075 s 1.00
cache / IPartOpt / tpu / BothRev 0.0000049627 s 0.000004959749999999999 s 1.00
cache / DefOpt / tpu / PreRev 0.000004984075 s 0.0000049680000000000005 s 1.00
cache / DefOpt / tpu / PostRev 0.0000049547 s 0.000004955975 s 1.00
cache / DefOpt / tpu / BothRev 0.00000496015 s 0.00000495815 s 1.00
cache / IDefOpt / tpu / PreRev 0.000004952625 s 0.00000497235 s 1.00
cache / IDefOpt / tpu / PostRev 0.000004961775 s 0.0000049673 s 1.00
cache / IDefOpt / tpu / BothRev 0.0000049675 s 0.0000049585 s 1.00
cache / JaXPipe / cpu / Primal 0.000012766 s 0.000007113319916243199 s 1.79
cache / Jax / cpu / Primal 0.000012616 s 0.000007572799877380019 s 1.67
cache / HLOOpt / cpu / Primal 0.000012608 s 0.000006496139994851546 s 1.94
cache / PartOpt / cpu / Primal 0.000012696 s 0.0000067282200507179366 s 1.89
cache / IPartOpt / cpu / Primal 0.000012595 s 0.000006539540081575979 s 1.93
cache / DefOpt / cpu / Primal 0.000013088 s 0.000007011559991951799 s 1.87
cache / IDefOpt / cpu / Primal 0.000012757 s 0.000006468759947892977 s 1.97
cache / JaXPipe / cpu / Forward 0.000023196000000000003 s 0.000014514539998344844 s 1.60
cache / Jax / cpu / Forward 0.000025577 s 0.00001441554000848555 s 1.77
cache / HLOOpt / cpu / Forward 0.000023504 s 0.000014784280047024368 s 1.59
cache / PartOpt / cpu / Forward 0.000023067 s 0.000014338260025397175 s 1.61
cache / IPartOpt / cpu / Forward 0.000024474 s 0.00001500011998359696 s 1.63
cache / DefOpt / cpu / Forward 0.000023398 s 0.000015464900061488153 s 1.51
cache / IDefOpt / cpu / Forward 0.000023658 s 0.000014286339974205475 s 1.66
cache / JaXPipe / cpu / PreRev 0.000027277 s 0.000016791880061646224 s 1.62
cache / JaXPipe / cpu / PostRev 0.000040874 s 0.0000222802400458022 s 1.83
cache / JaXPipe / cpu / BothRev 0.000024903 s 0.00001686876010353444 s 1.48
cache / Jax / cpu / BothRev 0.000027881 s 0.000021221759961918 s 1.31
cache / HLOOpt / cpu / PreRev 0.000025826000000000003 s 0.000017306019944953732 s 1.49
cache / HLOOpt / cpu / PostRev 0.000018278 s 0.00001918961999763269 s 0.95
cache / HLOOpt / cpu / BothRev 0.000018195 s 0.00001752650006892509 s 1.04
cache / PartOpt / cpu / PreRev 0.000017261 s 0.000016286500012938632 s 1.06
cache / PartOpt / cpu / PostRev 0.000020198 s 0.000020723759989778044 s 0.97
cache / PartOpt / cpu / BothRev 0.000017698 s 0.000017491439957666445 s 1.01
cache / IPartOpt / cpu / PreRev 0.000017885999999999998 s 0.0000170234999677632 s 1.05
cache / IPartOpt / cpu / PostRev 0.000019997 s 0.00002054747998045059 s 0.97
cache / IPartOpt / cpu / BothRev 0.0000187 s 0.000017071119982574602 s 1.10
cache / DefOpt / cpu / PreRev 0.000017942 s 0.000016248899992206134 s 1.10
cache / DefOpt / cpu / PostRev 0.000017944000000000003 s 0.00001709864003714756 s 1.05
cache / DefOpt / cpu / BothRev 0.00002393 s 0.000016460659990116254 s 1.45
cache / IDefOpt / cpu / PreRev 0.000017306 s 0.000016951420202531154 s 1.02
cache / IDefOpt / cpu / PostRev 0.000017391 s 0.000016882059899216984 s 1.03
cache / IDefOpt / cpu / BothRev 0.000018095 s 0.000016809840071800864 s 1.08
Concat / JaXPipe / cuda / Primal 0.0000019200000000000003 s 0.0000019200000000000003 s 1
Concat / Jax / cuda / Primal 0.0000019200000000000003 s 0.0000019200000000000003 s 1
Concat / HLOOpt / cuda / Primal 0.0000019200000000000003 s 0.0000019200000000000003 s 1
Concat / PartOpt / cuda / Primal 0.000001919 s 0.0000019200000000000003 s 1.00
Concat / IPartOpt / cuda / Primal 0.0000019200000000000003 s 0.0000019200000000000003 s 1
Concat / DefOpt / cuda / Primal 0.0000019200000000000003 s 0.0000019200000000000003 s 1
Concat / IDefOpt / cuda / Primal 0.0000019200000000000003 s 0.0000019200000000000003 s 1
Concat / JaXPipe / cuda / Forward 0.00001008 s 0.000010432 s 0.97
Concat / Jax / cuda / Forward 0.00001008 s 0.000010432 s 0.97
Concat / HLOOpt / cuda / Forward 0.000010112 s 0.000010272 s 0.98
Concat / PartOpt / cuda / Forward 0.000010304 s 0.000010528 s 0.98
Concat / IPartOpt / cuda / Forward 0.000010048 s 0.000010464 s 0.96
Concat / DefOpt / cuda / Forward 0.000009952 s 0.000010207 s 0.98
Concat / IDefOpt / cuda / Forward 0.000010208 s 0.000010303 s 0.99
Concat / JaXPipe / cuda / PreRev 0.000016672 s 0.000017056 s 0.98
Concat / JaXPipe / cuda / PostRev 0.000017217 s 0.000016864 s 1.02
Concat / JaXPipe / cuda / BothRev 0.000017024 s 0.000016768000000000003 s 1.02
Concat / Jax / cuda / BothRev 0.000016864 s 0.0000168 s 1.00
Concat / HLOOpt / cuda / PreRev 0.00001712 s 0.000016896000000000002 s 1.01
Concat / HLOOpt / cuda / PostRev 0.000017216 s 0.000016832 s 1.02
Concat / HLOOpt / cuda / BothRev 0.000016927000000000002 s 0.000016896000000000002 s 1.00
Concat / PartOpt / cuda / PreRev 0.000016608 s 0.000017888000000000002 s 0.93
Concat / PartOpt / cuda / PostRev 0.000016736 s 0.000017216 s 0.97
Concat / PartOpt / cuda / BothRev 0.000016607 s 0.000016063999999999997 s 1.03
Concat / IPartOpt / cuda / PreRev 0.000017375999999999998 s 0.000017152 s 1.01
Concat / IPartOpt / cuda / PostRev 0.00001712 s 0.00001648 s 1.04
Concat / IPartOpt / cuda / BothRev 0.000016672 s 0.000016545 s 1.01
Concat / DefOpt / cuda / PreRev 0.000017024 s 0.000017184 s 0.99
Concat / DefOpt / cuda / PostRev 0.000016927999999999998 s 0.000016768000000000003 s 1.01
Concat / DefOpt / cuda / BothRev 0.000016736 s 0.000016832 s 0.99
Concat / IDefOpt / cuda / PreRev 0.000017184 s 0.000017665 s 0.97
Concat / IDefOpt / cuda / PostRev 0.000016544 s 0.000016768000000000003 s 0.99
Concat / IDefOpt / cuda / BothRev 0.000016896000000000002 s 0.000017344 s 0.97
Concat / JaXPipe / tpu / Primal 0.0000015334 s 0.000001524275 s 1.01
Concat / Jax / tpu / Primal 0.000001536875 s 0.000001537425 s 1.00
Concat / HLOOpt / tpu / Primal 0.000001536175 s 0.000001522125 s 1.01
Concat / PartOpt / tpu / Primal 0.000001521825 s 0.00000153655 s 0.99
Concat / IPartOpt / tpu / Primal 0.000001540875 s 0.000001517 s 1.02
Concat / DefOpt / tpu / Primal 0.000001519575 s 0.0000015408 s 0.99
Concat / IDefOpt / tpu / Primal 0.00000153305 s 0.000001532075 s 1.00
Concat / JaXPipe / tpu / Forward 0.0000015795500000000002 s 0.0000015805250000000002 s 1.00
Concat / Jax / tpu / Forward 0.0000015522250000000002 s 0.00000155565 s 1.00
Concat / HLOOpt / tpu / Forward 0.0000015728249999999998 s 0.000001594875 s 0.99
Concat / PartOpt / tpu / Forward 0.000001565925 s 0.0000015500249999999998 s 1.01
Concat / IPartOpt / tpu / Forward 0.000001577925 s 0.000001577275 s 1.00
Concat / DefOpt / tpu / Forward 0.0000015536 s 0.00000153945 s 1.01
Concat / IDefOpt / tpu / Forward 0.0000015745 s 0.0000015775 s 1.00
Concat / JaXPipe / tpu / PreRev 0.000002002275 s 0.000001998375 s 1.00
Concat / JaXPipe / tpu / PostRev 0.0000020871500000000004 s 0.0000020933 s 1.00
Concat / JaXPipe / tpu / BothRev 0.000002003325 s 0.000001995775 s 1.00
Concat / Jax / tpu / BothRev 0.0000020753250000000003 s 0.0000020737 s 1.00
Concat / HLOOpt / tpu / PreRev 0.000002013225 s 0.000001997775 s 1.01
Concat / HLOOpt / tpu / PostRev 0.000002091225 s 0.0000020863000000000003 s 1.00
Concat / HLOOpt / tpu / BothRev 0.00000202255 s 0.0000019951 s 1.01
Concat / PartOpt / tpu / PreRev 0.0000020752 s 0.000002087125 s 0.99
Concat / PartOpt / tpu / PostRev 0.000002015975 s 0.000001998675 s 1.01
Concat / PartOpt / tpu / BothRev 0.000002083475 s 0.000002087575 s 1.00
Concat / IPartOpt / tpu / PreRev 0.00000201125 s 0.00000200505 s 1.00
Concat / IPartOpt / tpu / PostRev 0.000002072375 s 0.0000020779 s 1.00
Concat / IPartOpt / tpu / BothRev 0.000002018075 s 0.0000020009 s 1.01
Concat / DefOpt / tpu / PreRev 0.00000206945 s 0.00000207885 s 1.00
Concat / DefOpt / tpu / PostRev 0.00000200075 s 0.00000199995 s 1.00
Concat / DefOpt / tpu / BothRev 0.0000020683999999999995 s 0.0000020805 s 0.99
Concat / IDefOpt / tpu / PreRev 0.000002007825 s 0.0000019929 s 1.01
Concat / IDefOpt / tpu / PostRev 0.00000206635 s 0.00000208795 s 0.99
Concat / IDefOpt / tpu / BothRev 0.0000020123750000000004 s 0.000001993925 s 1.01
Concat / JaXPipe / cpu / Primal 0.000013001 s 0.000007628220027982025 s 1.70
Concat / Jax / cpu / Primal 0.000013153 s 0.000006918299950484652 s 1.90
Concat / HLOOpt / cpu / Primal 0.000012833 s 0.000006906620074005332 s 1.86
Concat / PartOpt / cpu / Primal 0.000012929 s 0.000007574280025437474 s 1.71
Concat / IPartOpt / cpu / Primal 0.000012904 s 0.00000827180003398098 s 1.56
Concat / DefOpt / cpu / Primal 0.000012958 s 0.000007524260072386824 s 1.72
Concat / IDefOpt / cpu / Primal 0.00001295 s 0.000007567080010630889 s 1.71
Concat / JaXPipe / cpu / Forward 0.0000182 s 0.000010493979989405488 s 1.73
Concat / Jax / cpu / Forward 0.000017488 s 0.000010956819933198858 s 1.60
Concat / HLOOpt / cpu / Forward 0.000017589 s 0.00001056330000210437 s 1.67
Concat / PartOpt / cpu / Forward 0.000017503999999999997 s 0.000011186020019522404 s 1.56
Concat / IPartOpt / cpu / Forward 0.000017595 s 0.00001060269996742136 s 1.66
Concat / DefOpt / cpu / Forward 0.000017607000000000003 s 0.000010989679976773914 s 1.60
Concat / IDefOpt / cpu / Forward 0.000017643 s 0.000011122739979327889 s 1.59
Concat / JaXPipe / cpu / PreRev 0.000019726 s 0.000013004640004510291 s 1.52
Concat / JaXPipe / cpu / PostRev 0.000019657 s 0.00001280795999264228 s 1.53
Concat / JaXPipe / cpu / BothRev 0.000019136 s 0.000013130419956723928 s 1.46
Concat / Jax / cpu / BothRev 0.00001975 s 0.00001225699990754947 s 1.61
Concat / HLOOpt / cpu / PreRev 0.000019731 s 0.00001261576004253584 s 1.56
Concat / HLOOpt / cpu / PostRev 0.000019326 s 0.000015385800033982377 s 1.26
Concat / HLOOpt / cpu / BothRev 0.000019266 s 0.000012868580015492623 s 1.50
Concat / PartOpt / cpu / PreRev 0.000019798 s 0.000012579099930007944 s 1.57
Concat / PartOpt / cpu / PostRev 0.000019496 s 0.00001343272004305618 s 1.45
Concat / PartOpt / cpu / BothRev 0.000019719 s 0.000013328659861144844 s 1.48
Concat / IPartOpt / cpu / PreRev 0.000019847 s 0.000012759020046360093 s 1.56
Concat / IPartOpt / cpu / PostRev 0.000019652 s 0.000012730760008707876 s 1.54
Concat / IPartOpt / cpu / BothRev 0.000019707 s 0.000012841920033679344 s 1.53
Concat / DefOpt / cpu / PreRev 0.000019742 s 0.00001291278002099716 s 1.53
Concat / DefOpt / cpu / PostRev 0.000019543 s 0.00001354980011456064 s 1.44
Concat / DefOpt / cpu / BothRev 0.000019506 s 0.00001339397997071501 s 1.46
Concat / IDefOpt / cpu / PreRev 0.00002005 s 0.00001281239998206729 s 1.56
Concat / IDefOpt / cpu / PostRev 0.000019686 s 0.000012565760098368628 s 1.57
Concat / IDefOpt / cpu / BothRev 0.000019597 s 0.000013160980088287034 s 1.49
const_scatter / JaXPipe / cuda / Primal 0.000001887 s 0.000001887 s 1
const_scatter / Jax / cuda / Primal 0.000001887 s 0.000001887 s 1
const_scatter / HLOOpt / cuda / Primal 0.000001887 s 0.000001888 s 1.00
const_scatter / PartOpt / cuda / Primal 0.000001887 s 0.000001887 s 1
const_scatter / IPartOpt / cuda / Primal 0.000001887 s 0.000001887 s 1
const_scatter / DefOpt / cuda / Primal 0.000001888 s 0.000001887 s 1.00
const_scatter / IDefOpt / cuda / Primal 0.000001887 s 0.000001888 s 1.00
const_scatter / JaXPipe / cuda / Forward 0.000010464 s 0.000009792 s 1.07
const_scatter / Jax / cuda / Forward 0.000010272 s 0.00001008 s 1.02
const_scatter / HLOOpt / cuda / Forward 0.000010336 s 0.00001008 s 1.03
const_scatter / PartOpt / cuda / Forward 0.000010176 s 0.00000976 s 1.04
const_scatter / IPartOpt / cuda / Forward 0.000009984 s 0.000009857 s 1.01
const_scatter / DefOpt / cuda / Forward 0.000010112 s 0.000009984 s 1.01
const_scatter / IDefOpt / cuda / Forward 0.000010272 s 0.000009888 s 1.04
const_scatter / JaXPipe / cuda / PreRev 0.00001696 s 0.000016929 s 1.00
const_scatter / JaXPipe / cuda / PostRev 0.00001696 s 0.000016544 s 1.03
const_scatter / JaXPipe / cuda / BothRev 0.000016032999999999997 s 0.000016832 s 0.95
const_scatter / Jax / cuda / BothRev 0.000016768000000000003 s 0.00001728 s 0.97
const_scatter / HLOOpt / cuda / PreRev 0.000016704 s 0.000016192 s 1.03
const_scatter / HLOOpt / cuda / PostRev 0.000016896000000000002 s 0.000017087 s 0.99
const_scatter / HLOOpt / cuda / BothRev 0.000016193 s 0.00001616 s 1.00
const_scatter / PartOpt / cuda / PreRev 0.00001648 s 0.000017056 s 0.97
const_scatter / PartOpt / cuda / PostRev 0.000017088 s 0.000015808 s 1.08
const_scatter / PartOpt / cuda / BothRev 0.000016447 s 0.000018817 s 0.87
const_scatter / IPartOpt / cuda / PreRev 0.0000184 s 0.000016864 s 1.09
const_scatter / IPartOpt / cuda / PostRev 0.00001632 s 0.000016929 s 0.96
const_scatter / IPartOpt / cuda / BothRev 0.000016895 s 0.000016224 s 1.04
const_scatter / DefOpt / cuda / PreRev 0.000018752000000000003 s 0.000016288 s 1.15
const_scatter / DefOpt / cuda / PostRev 0.00001824 s 0.000016512 s 1.10
const_scatter / DefOpt / cuda / BothRev 0.00001696 s 0.000018752000000000003 s 0.90
const_scatter / IDefOpt / cuda / PreRev 0.00001664 s 0.00001696 s 0.98
const_scatter / IDefOpt / cuda / PostRev 0.000017184 s 0.000018337 s 0.94
const_scatter / IDefOpt / cuda / BothRev 0.000016544 s 0.000016832 s 0.98
const_scatter / JaXPipe / tpu / Primal 0.000003783475 s 0.000003809775 s 0.99
const_scatter / Jax / tpu / Primal 0.000003802475 s 0.00000383075 s 0.99
const_scatter / HLOOpt / tpu / Primal 0.000003792025 s 0.000003812875 s 0.99
const_scatter / PartOpt / tpu / Primal 0.00000381935 s 0.000003816375 s 1.00
const_scatter / IPartOpt / tpu / Primal 0.00000378165 s 0.000003810475 s 0.99
const_scatter / DefOpt / tpu / Primal 0.0000038116 s 0.000003818025 s 1.00
const_scatter / IDefOpt / tpu / Primal 0.0000038054 s 0.000003819575 s 1.00
const_scatter / JaXPipe / tpu / Forward 0.000006457525 s 0.000006456925 s 1.00
const_scatter / Jax / tpu / Forward 0.00000648975 s 0.0000064949 s 1.00
const_scatter / HLOOpt / tpu / Forward 0.0000064604 s 0.000006442675 s 1.00
const_scatter / PartOpt / tpu / Forward 0.0000064826 s 0.000006506075 s 1.00
const_scatter / IPartOpt / tpu / Forward 0.000006458775 s 0.000006462024999999999 s 1.00
const_scatter / DefOpt / tpu / Forward 0.000006482725 s 0.000006515 s 1.00
const_scatter / IDefOpt / tpu / Forward 0.000006464975 s 0.000006475775000000001 s 1.00
const_scatter / JaXPipe / tpu / PreRev 0.000006604275 s 0.00000669465 s 0.99
const_scatter / JaXPipe / tpu / PostRev 0.000006620075 s 0.000006671425 s 0.99
const_scatter / JaXPipe / tpu / BothRev 0.00000663015 s 0.000006659125 s 1.00
const_scatter / Jax / tpu / BothRev 0.0000066189 s 0.0000066715250000000005 s 0.99
const_scatter / HLOOpt / tpu / PreRev 0.000006604425000000001 s 0.000006665824999999999 s 0.99
const_scatter / HLOOpt / tpu / PostRev 0.000006648225 s 0.000006661775 s 1.00
const_scatter / HLOOpt / tpu / BothRev 0.000006619825000000001 s 0.000006681724999999999 s 0.99
const_scatter / PartOpt / tpu / PreRev 0.000006638199999999999 s 0.0000066773 s 0.99
const_scatter / PartOpt / tpu / PostRev 0.000006605225 s 0.00000667105 s 0.99
const_scatter / PartOpt / tpu / BothRev 0.000006621099999999999 s 0.000006679924999999999 s 0.99
const_scatter / IPartOpt / tpu / PreRev 0.000006602550000000001 s 0.00000667415 s 0.99
const_scatter / IPartOpt / tpu / PostRev 0.000006615299999999999 s 0.000006661175 s 0.99
const_scatter / IPartOpt / tpu / BothRev 0.000006604149999999999 s 0.00000665635 s 0.99
const_scatter / DefOpt / tpu / PreRev 0.000006615875 s 0.0000066613 s 0.99
const_scatter / DefOpt / tpu / PostRev 0.000006617 s 0.000006663225 s 0.99
const_scatter / DefOpt / tpu / BothRev 0.000006632625 s 0.000006669075 s 0.99
const_scatter / IDefOpt / tpu / PreRev 0.0000066213500000000006 s 0.000006667524999999999 s 0.99
const_scatter / IDefOpt / tpu / PostRev 0.000006633775 s 0.000006665875 s 1.00
const_scatter / IDefOpt / tpu / BothRev 0.00000660775 s 0.00000665495 s 0.99
const_scatter / JaXPipe / cpu / Primal 0.000012669 s 0.000007149519988161046 s 1.77
const_scatter / Jax / cpu / Primal 0.000012623 s 0.000007949459886731347 s 1.59
const_scatter / HLOOpt / cpu / Primal 0.000013334 s 0.000008326620045409073 s 1.60
const_scatter / PartOpt / cpu / Primal 0.0000126 s 0.000006578599986823974 s 1.92
const_scatter / IPartOpt / cpu / Primal 0.000012645 s 0.000007133699964469997 s 1.77
const_scatter / DefOpt / cpu / Primal 0.000013197 s 0.000007454360020346939 s 1.77
const_scatter / IDefOpt / cpu / Primal 0.000013228 s 0.000007193499932327541 s 1.84
const_scatter / JaXPipe / cpu / Forward 0.000017942 s 0.000011891360027220798 s 1.51
const_scatter / Jax / cpu / Forward 0.000016781 s 0.00001069080000888789 s 1.57
const_scatter / HLOOpt / cpu / Forward 0.000017967 s 0.000012667159935517704 s 1.42
const_scatter / PartOpt / cpu / Forward 0.000018010000000000002 s 0.000011910540033568397 s 1.51
const_scatter / IPartOpt / cpu / Forward 0.000017712999999999998 s 0.000012706300021818604 s 1.39
const_scatter / DefOpt / cpu / Forward 0.000018010000000000002 s 0.000011037019939976744 s 1.63
const_scatter / IDefOpt / cpu / Forward 0.000017930000000000003 s 0.000011218979943805607 s 1.60
const_scatter / JaXPipe / cpu / PreRev 0.000496989 s 0.000288881239976 s 1.72
const_scatter / JaXPipe / cpu / PostRev 0.000502689 s 0.0002845111800525 s 1.77
const_scatter / JaXPipe / cpu / BothRev 0.000516574 s 0.000285194879998 s 1.81
const_scatter / Jax / cpu / BothRev 0.000509047 s 0.0002820368200627 s 1.80
const_scatter / HLOOpt / cpu / PreRev 0.00049836 s 0.0002914486799454 s 1.71
const_scatter / HLOOpt / cpu / PostRev 0.000514258 s 0.0002887479600394 s 1.78
const_scatter / HLOOpt / cpu / BothRev 0.000521565 s 0.0002859900799921 s 1.82
const_scatter / PartOpt / cpu / PreRev 0.000530882 s 0.0002842538599543 s 1.87
const_scatter / PartOpt / cpu / PostRev 0.000513247 s 0.0002842179198887 s 1.81
const_scatter / PartOpt / cpu / BothRev 0.000496544 s 0.0002973716400265 s 1.67
const_scatter / IPartOpt / cpu / PreRev 0.000503489 s 0.0002858380599536 s 1.76
const_scatter / IPartOpt / cpu / PostRev 0.00048986 s 0.0002828226599194 s 1.73
const_scatter / IPartOpt / cpu / BothRev 0.000532755 s 0.0002875438000046 s 1.85
const_scatter / DefOpt / cpu / PreRev 0.000507676 s 0.0002848754000115 s 1.78
const_scatter / DefOpt / cpu / PostRev 0.000514424 s 0.000288019380132 s 1.79
const_scatter / DefOpt / cpu / BothRev 0.000509288 s 0.0002848568600711 s 1.79
const_scatter / IDefOpt / cpu / PreRev 0.000518629 s 0.0002855306399942 s 1.82
const_scatter / IDefOpt / cpu / PostRev 0.00050351 s 0.0002858346200628 s 1.76
const_scatter / IDefOpt / cpu / BothRev 0.000501344 s 0.0002871412999957 s 1.75
GenDot / JaXPipe / cuda / Primal 0.000002016 s 0.000002015 s 1.00
GenDot / Jax / cuda / Primal 0.000002016 s 0.000002015 s 1.00
GenDot / HLOOpt / cuda / Primal 0.000002015 s 0.000002015 s 1
GenDot / PartOpt / cuda / Primal 0.000002016 s 0.000002015 s 1.00
GenDot / IPartOpt / cuda / Primal 0.000002016 s 0.000002016 s 1
GenDot / DefOpt / cuda / Primal 0.000001984 s 0.000002015 s 0.98
GenDot / IDefOpt / cuda / Primal 0.000002015 s 0.000001984 s 1.02
GenDot / JaXPipe / cuda / Forward 0.000010368 s 0.000010304 s 1.01
GenDot / Jax / cuda / Forward 0.0000104 s 0.000010209 s 1.02
GenDot / HLOOpt / cuda / Forward 0.00000992 s 0.000010304 s 0.96
GenDot / PartOpt / cuda / Forward 0.000010176 s 0.000010272 s 0.99
GenDot / IPartOpt / cuda / Forward 0.000010304 s 0.000009984 s 1.03
GenDot / DefOpt / cuda / Forward 0.000010432 s 0.000010304 s 1.01
GenDot / IDefOpt / cuda / Forward 0.000010048 s 0.00001008 s 1.00
GenDot / JaXPipe / cuda / PreRev 0.000010304 s 0.000010687 s 0.96
GenDot / JaXPipe / cuda / PostRev 0.000010016 s 0.000010144 s 0.99
GenDot / JaXPipe / cuda / BothRev 0.000010112 s 0.00001024 s 0.99
GenDot / Jax / cuda / BothRev 0.000010176 s 0.000010592 s 0.96
GenDot / HLOOpt / cuda / PreRev 0.00001056 s 0.000010464 s 1.01
GenDot / HLOOpt / cuda / PostRev 0.000010112 s 0.000010368 s 0.98
GenDot / HLOOpt / cuda / BothRev 0.00000992 s 0.000009888 s 1.00
GenDot / PartOpt / cuda / PreRev 0.000009856 s 0.000010591 s 0.93
GenDot / PartOpt / cuda / PostRev 0.000010208 s 0.000011136 s 0.92
GenDot / PartOpt / cuda / BothRev 0.000009984 s 0.000010048 s 0.99
GenDot / IPartOpt / cuda / PreRev 0.000010176 s 0.000010432 s 0.98
GenDot / IPartOpt / cuda / PostRev 0.000010208 s 0.000011393 s 0.90
GenDot / IPartOpt / cuda / BothRev 0.000009824 s 0.000009825 s 1.00
GenDot / DefOpt / cuda / PreRev 0.00001024 s 0.000010912 s 0.94
GenDot / DefOpt / cuda / PostRev 0.000009984 s 0.000010113 s 0.99
GenDot / DefOpt / cuda / BothRev 0.000010144 s 0.000010144 s 1
GenDot / IDefOpt / cuda / PreRev 0.000010271 s 0.000010304 s 1.00
GenDot / IDefOpt / cuda / PostRev 0.000010303 s 0.000010336 s 1.00
GenDot / IDefOpt / cuda / BothRev 0.000010112 s 0.000010464 s 0.97
GenDot / JaXPipe / tpu / Primal 9.2945e-7 s 9.294e-7 s 1.00
GenDot / Jax / tpu / Primal 9.25125e-7 s 9.2555e-7 s 1.00
GenDot / HLOOpt / tpu / Primal 0.0000015664 s 0.000001581275 s 0.99
GenDot / PartOpt / tpu / Primal 9.2575e-7 s 9.25675e-7 s 1.00
GenDot / IPartOpt / tpu / Primal 9.294e-7 s 9.297e-7 s 1.00
GenDot / DefOpt / tpu / Primal 0.00000148535 s 0.000001502775 s 0.99
GenDot / IDefOpt / tpu / Primal 0.0000015699 s 0.0000015887 s 0.99
GenDot / JaXPipe / tpu / Forward 0.0000031532 s 0.00000317675 s 0.99
GenDot / Jax / tpu / Forward 0.00000231895 s 0.00000232435 s 1.00
GenDot / HLOOpt / tpu / Forward 0.000003106625 s 0.000003131975 s 0.99
GenDot / PartOpt / tpu / Forward 0.00000320955 s 0.0000032324 s 0.99
GenDot / IPartOpt / tpu / Forward 0.000003103475 s 0.0000031277500000000003 s 0.99
GenDot / DefOpt / tpu / Forward 0.000003212475 s 0.00000322695 s 1.00
GenDot / IDefOpt / tpu / Forward 0.0000031068000000000005 s 0.000003140975 s 0.99
GenDot / JaXPipe / tpu / PreRev 0.0000029597250000000003 s 0.000002981425 s 0.99
GenDot / JaXPipe / tpu / PostRev 0.000002406925 s 0.0000024025000000000003 s 1.00
GenDot / JaXPipe / tpu / BothRev 0.0000029669000000000003 s 0.0000029811 s 1.00
GenDot / Jax / tpu / BothRev 0.0000024024 s 0.000002399675 s 1.00
GenDot / HLOOpt / tpu / PreRev 0.0000029507 s 0.000002974 s 0.99
GenDot / HLOOpt / tpu / PostRev 0.000002947625 s 0.000002929375 s 1.01
GenDot / HLOOpt / tpu / BothRev 0.0000029464 s 0.00000298385 s 0.99
GenDot / PartOpt / tpu / PreRev 0.0000029233 s 0.000002923525 s 1.00
GenDot / PartOpt / tpu / PostRev 0.0000024043 s 0.00000239995 s 1.00
GenDot / PartOpt / tpu / BothRev 0.000002937975 s 0.00000292955 s 1.00
GenDot / IPartOpt / tpu / PreRev 0.000002951175 s 0.000002978925 s 0.99
GenDot / IPartOpt / tpu / PostRev 0.0000024099 s 0.000002407425 s 1.00
GenDot / IPartOpt / tpu / BothRev 0.000002956175 s 0.00000297985 s 0.99
GenDot / DefOpt / tpu / PreRev 0.0000029298000000000003 s 0.0000029243750000000003 s 1.00
GenDot / DefOpt / tpu / PostRev 0.000002962875 s 0.0000029796 s 0.99
GenDot / DefOpt / tpu / BothRev 0.000002932625 s 0.0000029346499999999995 s 1.00
GenDot / IDefOpt / tpu / PreRev 0.000002957775 s 0.0000029920249999999995 s 0.99
GenDot / IDefOpt / tpu / PostRev 0.000002937875 s 0.0000029314000000000004 s 1.00
GenDot / IDefOpt / tpu / BothRev 0.000002965975 s 0.0000029811 s 0.99
GenDot / JaXPipe / cpu / Primal 0.000015233999999999998 s 0.000008721840058569797 s 1.75
GenDot / Jax / cpu / Primal 0.000015128000000000002 s 0.000008433120092377066 s 1.79
GenDot / HLOOpt / cpu / Primal 0.00001388 s 0.000008394980050070444 s 1.65
GenDot / PartOpt / cpu / Primal 0.000015032 s 0.00000837149993458297 s 1.80
GenDot / IPartOpt / cpu / Primal 0.000015295 s 0.00000854325995533145 s 1.79
GenDot / DefOpt / cpu / Primal 0.000013963 s 0.000008057859904511133 s 1.73
GenDot / IDefOpt / cpu / Primal 0.000013919 s 0.000007973839965416118 s 1.75
GenDot / JaXPipe / cpu / Forward 0.000019893 s 0.000011691660020005655 s 1.70
GenDot / Jax / cpu / Forward 0.000020812 s 0.000011326959993311905 s 1.84
GenDot / HLOOpt / cpu / Forward 0.000019095 s 0.000011817120102932676 s 1.62
GenDot / PartOpt / cpu / Forward 0.00001864 s 0.000011394320008548677 s 1.64
GenDot / IPartOpt / cpu / Forward 0.000019555 s 0.00001216484004544327 s 1.61
GenDot / DefOpt / cpu / Forward 0.000019317 s 0.00001196550010718056 s 1.61
GenDot / IDefOpt / cpu / Forward 0.000019205 s 0.00001184767999802716 s 1.62
GenDot / JaXPipe / cpu / PreRev 0.00001945 s 0.000012314019968471258 s 1.58
GenDot / JaXPipe / cpu / PostRev 0.000020175 s 0.000011148480025440222 s 1.81
GenDot / JaXPipe / cpu / BothRev 0.000019256 s 0.000012437979858077596 s 1.55
GenDot / Jax / cpu / BothRev 0.000020528 s 0.000011231699991185453 s 1.83
GenDot / HLOOpt / cpu / PreRev 0.000019207 s 0.000012683259974437532 s 1.51
GenDot / HLOOpt / cpu / PostRev 0.000019352 s 0.000013554239958466496 s 1.43
GenDot / HLOOpt / cpu / BothRev 0.000019169 s 0.000011783660029323074 s 1.63
GenDot / PartOpt / cpu / PreRev 0.000019796 s 0.000011460859968792648 s 1.73
GenDot / PartOpt / cpu / PostRev 0.0000202 s 0.000010740839989011875 s 1.88
GenDot / PartOpt / cpu / BothRev 0.000018923 s 0.000011952399981964843 s 1.58
GenDot / IPartOpt / cpu / PreRev 0.00001943 s 0.00001170288000139408 s 1.66
GenDot / IPartOpt / cpu / PostRev 0.000020156 s 0.000011515260030137145 s 1.75
GenDot / IPartOpt / cpu / BothRev 0.000019228 s 0.000012036140014970444 s 1.60
GenDot / DefOpt / cpu / PreRev 0.000019253 s 0.00001183655995191657 s 1.63
GenDot / DefOpt / cpu / PostRev 0.000019202 s 0.000011791820052167168 s 1.63
GenDot / DefOpt / cpu / BothRev 0.00001932 s 0.000012289839978620876 s 1.57
GenDot / IDefOpt / cpu / PreRev 0.000018827 s 0.000011853979976876872 s 1.59
GenDot / IDefOpt / cpu / PostRev 0.000019058 s 0.000011159560017404146 s 1.71
GenDot / IDefOpt / cpu / BothRev 0.000018982 s 0.000012050519981130492 s 1.58
hlo_ffi / JaXPipe / cuda / Primal 0.000001984 s 0.000001984 s 1
hlo_ffi / Jax / cuda / Primal 0.000001984 s 0.000001983 s 1.00
hlo_ffi / HLOOpt / cuda / Primal 0.000001983 s 0.000001984 s 1.00
hlo_ffi / PartOpt / cuda / Primal 0.000001984 s 0.000001984 s 1
hlo_ffi / IPartOpt / cuda / Primal 0.000001984 s 0.000001983 s 1.00
hlo_ffi / DefOpt / cuda / Primal 0.000001984 s 0.000001983 s 1.00
hlo_ffi / IDefOpt / cuda / Primal 0.000001984 s 0.000001983 s 1.00
hlo_ffi / JaXPipe / cuda / Forward 0.000002049 s 0.00000208 s 0.99
hlo_ffi / Jax / cuda / Forward 0.00000208 s 0.00000208 s 1
hlo_ffi / HLOOpt / cuda / Forward 0.00000208 s 0.00000208 s 1
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 0.000002048 s 0.000002047 s 1.00
hlo_ffi / JaXPipe / cuda / PostRev 0.000002047 s 0.000002048 s 1.00
hlo_ffi / JaXPipe / cuda / BothRev 0.000002048 s 0.000002047 s 1.00
hlo_ffi / Jax / cuda / BothRev 0.000002047 s 0.000002048 s 1.00
hlo_ffi / HLOOpt / cuda / PreRev 0.000002047 s 0.000002047 s 1
hlo_ffi / HLOOpt / cuda / PostRev 0.000002048 s 0.000002047 s 1.00
hlo_ffi / HLOOpt / cuda / BothRev 0.000002048 s 0.000002048 s 1
hlo_ffi / PartOpt / cuda / PreRev 0.000002047 s 0.000002048 s 1.00
hlo_ffi / PartOpt / cuda / PostRev 0.000002047 s 0.000002048 s 1.00
hlo_ffi / PartOpt / cuda / BothRev 0.000002048 s 0.000002048 s 1
hlo_ffi / IPartOpt / cuda / PreRev 0.000002047 s 0.000002047 s 1
hlo_ffi / IPartOpt / cuda / PostRev 0.000002048 s 0.000002047 s 1.00
hlo_ffi / IPartOpt / cuda / BothRev 0.000002048 s 0.000002048 s 1
hlo_ffi / DefOpt / cuda / PreRev 0.000002048 s 0.000002048 s 1
hlo_ffi / DefOpt / cuda / PostRev 0.000002048 s 0.000002048 s 1
hlo_ffi / DefOpt / cuda / BothRev 0.000002048 s 0.000002047 s 1.00
hlo_ffi / IDefOpt / cuda / PreRev 0.000002048 s 0.000002048 s 1
hlo_ffi / IDefOpt / cuda / PostRev 0.000002047 s 0.000002048 s 1.00
hlo_ffi / IDefOpt / cuda / BothRev 0.000002048 s 0.000002047 s 1.00
hlo_ffi / JaXPipe / tpu / Primal 9.28575e-7 s 9.315e-7 s 1.00
hlo_ffi / Jax / tpu / Primal 9.5065e-7 s 9.547e-7 s 1.00
hlo_ffi / HLOOpt / tpu / Primal 9.06675e-7 s 9.0845e-7 s 1.00
hlo_ffi / PartOpt / tpu / Primal 9.527e-7 s 9.568749999999998e-7 s 1.00
hlo_ffi / IPartOpt / tpu / Primal 9.06675e-7 s 9.10425e-7 s 1.00
hlo_ffi / DefOpt / tpu / Primal 9.5065e-7 s 9.53225e-7 s 1.00
hlo_ffi / IDefOpt / tpu / Primal 9.08825e-7 s 9.049e-7 s 1.00
hlo_ffi / JaXPipe / tpu / Forward 9.4945e-7 s 9.495e-7 s 1.00
hlo_ffi / Jax / tpu / Forward 9.81775e-7 s 9.81775e-7 s 1
hlo_ffi / HLOOpt / tpu / Forward 9.74675e-7 s 9.736e-7 s 1.00
hlo_ffi / PartOpt / tpu / Forward 9.34625e-7 s 9.34375e-7 s 1.00
hlo_ffi / IPartOpt / tpu / Forward 9.73775e-7 s 9.73775e-7 s 1
hlo_ffi / DefOpt / tpu / Forward 9.3455e-7 s 9.3415e-7 s 1.00
hlo_ffi / IDefOpt / tpu / Forward 9.74625e-7 s 9.742e-7 s 1.00
hlo_ffi / JaXPipe / tpu / PreRev 9.37825e-7 s 9.377e-7 s 1.00
hlo_ffi / JaXPipe / tpu / PostRev 9.6525e-7 s 9.6535e-7 s 1.00
hlo_ffi / JaXPipe / tpu / BothRev 9.625749999999998e-7 s 9.6205e-7 s 1.00
hlo_ffi / Jax / tpu / BothRev 9.651e-7 s 9.65125e-7 s 1.00
hlo_ffi / HLOOpt / tpu / PreRev 9.62125e-7 s 9.628e-7 s 1.00
hlo_ffi / HLOOpt / tpu / PostRev 9.6505e-7 s 9.64675e-7 s 1.00
hlo_ffi / HLOOpt / tpu / BothRev 9.62125e-7 s 9.61425e-7 s 1.00
hlo_ffi / PartOpt / tpu / PreRev 9.65175e-7 s 9.64875e-7 s 1.00
hlo_ffi / PartOpt / tpu / PostRev 9.62225e-7 s 9.616e-7 s 1.00
hlo_ffi / PartOpt / tpu / BothRev 9.65225e-7 s 9.650749999999998e-7 s 1.00
hlo_ffi / IPartOpt / tpu / PreRev 9.61875e-7 s 9.62125e-7 s 1.00
hlo_ffi / IPartOpt / tpu / PostRev 9.64625e-7 s 9.6515e-7 s 1.00
hlo_ffi / IPartOpt / tpu / BothRev 9.618e-7 s 9.62425e-7 s 1.00
hlo_ffi / DefOpt / tpu / PreRev 9.651e-7 s 9.64625e-7 s 1.00
hlo_ffi / DefOpt / tpu / PostRev 9.62475e-7 s 9.62125e-7 s 1.00
hlo_ffi / DefOpt / tpu / BothRev 9.64675e-7 s 9.64575e-7 s 1.00
hlo_ffi / IDefOpt / tpu / PreRev 9.62225e-7 s 9.61425e-7 s 1.00
hlo_ffi / IDefOpt / tpu / PostRev 9.649e-7 s 9.64925e-7 s 1.00
hlo_ffi / IDefOpt / tpu / BothRev 9.625250000000002e-7 s 9.62175e-7 s 1.00
hlo_ffi / JaXPipe / cpu / Primal 0.000017303 s 0.00001101443993320572 s 1.57
hlo_ffi / Jax / cpu / Primal 0.000017049999999999998 s 0.00001038607993905316 s 1.64
hlo_ffi / HLOOpt / cpu / Primal 0.000016889 s 0.00000995977994534769 s 1.70
hlo_ffi / PartOpt / cpu / Primal 0.000017099 s 0.000010140960057469783 s 1.69
hlo_ffi / IPartOpt / cpu / Primal 0.000017047 s 0.000010418199926789383 s 1.64
hlo_ffi / DefOpt / cpu / Primal 0.000017088 s 0.000009951919982995604 s 1.72
hlo_ffi / IDefOpt / cpu / Primal 0.000017173 s 0.000009953820026566972 s 1.73
hlo_ffi / JaXPipe / cpu / Forward 0.000024609 s 0.000014495539962808836 s 1.70
hlo_ffi / Jax / cpu / Forward 0.000023795 s 0.00001425110014679376 s 1.67
hlo_ffi / HLOOpt / cpu / Forward 0.000024204 s 0.000014269539951783372 s 1.70
hlo_ffi / PartOpt / cpu / Forward 0.000023764 s 0.000014493419930659002 s 1.64
hlo_ffi / IPartOpt / cpu / Forward 0.000023974 s 0.000014325679949251936 s 1.67
hlo_ffi / DefOpt / cpu / Forward 0.000023912 s 0.000014539819985657232 s 1.64
hlo_ffi / IDefOpt / cpu / Forward 0.000023515 s 0.000014344060036819427 s 1.64
hlo_ffi / JaXPipe / cpu / PreRev 0.00002426 s 0.00001870272000815021 s 1.30
hlo_ffi / JaXPipe / cpu / PostRev 0.000023166 s 0.000014635620009357807 s 1.58
hlo_ffi / JaXPipe / cpu / BothRev 0.000023496 s 0.000014097020048211562 s 1.67
hlo_ffi / Jax / cpu / BothRev 0.000023183 s 0.000014634620056312997 s 1.58
hlo_ffi / HLOOpt / cpu / PreRev 0.000023997 s 0.000014630739970016294 s 1.64
hlo_ffi / HLOOpt / cpu / PostRev 0.0000235 s 0.000018660599998838734 s 1.26
hlo_ffi / HLOOpt / cpu / BothRev 0.000023225 s 0.000014310059978015488 s 1.62
hlo_ffi / PartOpt / cpu / PreRev 0.000023511 s 0.000014527380026265746 s 1.62
hlo_ffi / PartOpt / cpu / PostRev 0.000023395 s 0.000014577159945474704 s 1.60
hlo_ffi / PartOpt / cpu / BothRev 0.000023626 s 0.00001433232004274032 s 1.65
hlo_ffi / IPartOpt / cpu / PreRev 0.000023608 s 0.00001449733992558322 s 1.63
hlo_ffi / IPartOpt / cpu / PostRev 0.000023435 s 0.000014475579864665634 s 1.62
hlo_ffi / IPartOpt / cpu / BothRev 0.000023071 s 0.00001461617996028508 s 1.58
hlo_ffi / DefOpt / cpu / PreRev 0.000023847 s 0.000014917540011083474 s 1.60
hlo_ffi / DefOpt / cpu / PostRev 0.000023922 s 0.000014270540013967547 s 1.68
hlo_ffi / DefOpt / cpu / BothRev 0.000023143 s 0.000014421520027099178 s 1.60
hlo_ffi / IDefOpt / cpu / PreRev 0.000023709 s 0.000014733080060977954 s 1.61
hlo_ffi / IDefOpt / cpu / PostRev 0.000023419 s 0.000014166279979690444 s 1.65
hlo_ffi / IDefOpt / cpu / BothRev 0.000023752 s 0.000014473339997493894 s 1.64
jaxmd20 / JaXPipe / cuda / Primal 0.0014578249999999 s 0.001461953 s 1.00
jaxmd20 / Jax / cuda / Primal 0.001506528 s 0.001513856 s 1.00
jaxmd20 / HLOOpt / cuda / Primal 0.00133968 s 0.0013292479999999 s 1.01
jaxmd20 / PartOpt / cuda / Primal 0.0013274559999999 s 0.001310145 s 1.01
jaxmd20 / IPartOpt / cuda / Primal 0.001320415 s 0.001310016 s 1.01
jaxmd20 / DefOpt / cuda / Primal 0.000913184 s 0.000921888 s 0.99
jaxmd20 / IDefOpt / cuda / Primal 0.000944736 s 0.000957184 s 0.99
jaxmd20 / JaXPipe / cuda / Forward 0.001560065 s 0.001541601 s 1.01
jaxmd20 / Jax / cuda / Forward 0.001797631 s 0.0017879039999999 s 1.01
jaxmd20 / HLOOpt / cuda / Forward 0.001667296 s 0.001628385 s 1.02
jaxmd20 / PartOpt / cuda / Forward 0.0016425 s 0.001642914 s 1.00
jaxmd20 / IPartOpt / cuda / Forward 0.00161024 s 0.0016282879999999 s 0.99
jaxmd20 / DefOpt / cuda / Forward 0.001635167 s 0.0016393599999999 s 1.00
jaxmd20 / IDefOpt / cuda / Forward 0.001616801 s 0.001614976 s 1.00
jaxmd20 / JaXPipe / cuda / PreRev 0.002667714 s 0.002704768 s 0.99
jaxmd20 / JaXPipe / cuda / PostRev 0.005376929 s 0.005357793 s 1.00
jaxmd20 / JaXPipe / cuda / BothRev 0.0027833999999999 s 0.002702431 s 1.03
jaxmd20 / Jax / cuda / BothRev 0.005346277 s 0.005367966 s 1.00
jaxmd20 / HLOOpt / cuda / PreRev 0.00275184 s 0.002775808 s 0.99
jaxmd20 / HLOOpt / cuda / PostRev 0.00532515 s 0.005392673 s 0.99
jaxmd20 / HLOOpt / cuda / BothRev 0.002715555 s 0.002727138 s 1.00
jaxmd20 / PartOpt / cuda / PreRev 0.00281008 s 0.002822815 s 1.00
jaxmd20 / PartOpt / cuda / PostRev 0.0054553599999999 s 0.0054714609999999 s 1.00
jaxmd20 / PartOpt / cuda / BothRev 0.002756899 s 0.002776291 s 0.99
jaxmd20 / IPartOpt / cuda / PreRev 0.002830081 s 0.002825184 s 1.00
jaxmd20 / IPartOpt / cuda / PostRev 0.005462531 s 0.005470689 s 1.00
jaxmd20 / IPartOpt / cuda / BothRev 0.002756481 s 0.002770497 s 0.99
jaxmd20 / DefOpt / cuda / PreRev 0.00282221 s 0.002826978 s 1.00
jaxmd20 / DefOpt / cuda / PostRev 0.002792354 s 0.002780129 s 1.00
jaxmd20 / DefOpt / cuda / BothRev 0.002755296 s 0.002751842 s 1.00
jaxmd20 / IDefOpt / cuda / PreRev 0.0028104 s 0.002833476 s 0.99
jaxmd20 / IDefOpt / cuda / PostRev 0.002306015 s 0.002308319 s 1.00
jaxmd20 / IDefOpt / cuda / BothRev 0.002864993 s 0.002736737 s 1.05
jaxmd20 / JaXPipe / tpu / Primal 0.009286555 s 0.00926679 s 1.00
jaxmd20 / Jax / tpu / Primal 0.0092696137499999 s 0.009271479375 s 1.00
jaxmd20 / HLOOpt / tpu / Primal 0.0092450181249999 s 0.009153965 s 1.01
jaxmd20 / PartOpt / tpu / Primal 0.0091991975 s 0.009208060625 s 1.00
jaxmd20 / IPartOpt / tpu / Primal 0.0091993399999999 s 0.00920467375 s 1.00
jaxmd20 / DefOpt / tpu / Primal 0.008796989375 s 0.008802515 s 1.00
jaxmd20 / IDefOpt / tpu / Primal 0.008700651875 s 0.008702111875 s 1.00
jaxmd20 / JaXPipe / tpu / Forward 0.017415965 s 0.017405166875 s 1.00
jaxmd20 / Jax / tpu / Forward 0.018733326875 s 0.018732650625 s 1.00
jaxmd20 / HLOOpt / tpu / Forward 0.01739920375 s 0.0174011031249999 s 1.00
jaxmd20 / PartOpt / tpu / Forward 0.01741451375 s 0.017417733125 s 1.00
jaxmd20 / IPartOpt / tpu / Forward 0.017409101875 s 0.01741585875 s 1.00
jaxmd20 / DefOpt / tpu / Forward 0.017413069375 s 0.01742094875 s 1.00
jaxmd20 / IDefOpt / tpu / Forward 0.017412073125 s 0.017415853125 s 1.00
jaxmd20 / JaXPipe / tpu / PreRev 0.02544395125 s 0.025454308125 s 1.00
jaxmd20 / JaXPipe / tpu / PostRev 0.02187623375 s 0.021875835625 s 1.00
jaxmd20 / JaXPipe / tpu / BothRev 0.02546934 s 0.025454168125 s 1.00
jaxmd20 / Jax / tpu / BothRev 0.0218733775 s 0.02187132625 s 1.00
jaxmd20 / HLOOpt / tpu / PreRev 0.025586291875 s 0.025574765625 s 1.00
jaxmd20 / HLOOpt / tpu / PostRev 0.0210028875 s 0.020701680625 s 1.01
jaxmd20 / HLOOpt / tpu / BothRev 0.0256876681249999 s 0.02567604875 s 1.00
jaxmd20 / PartOpt / tpu / PreRev 0.0254885425 s 0.0254586274999999 s 1.00
jaxmd20 / PartOpt / tpu / PostRev 0.021538471875 s 0.021507803125 s 1.00
jaxmd20 / PartOpt / tpu / BothRev 0.025570053125 s 0.025554400625 s 1.00
jaxmd20 / IPartOpt / tpu / PreRev 0.02547263125 s 0.025461835 s 1.00
jaxmd20 / IPartOpt / tpu / PostRev 0.021521653125 s 0.02151244375 s 1.00
jaxmd20 / IPartOpt / tpu / BothRev 0.025569895625 s 0.025553270625 s 1.00
jaxmd20 / DefOpt / tpu / PreRev 0.02548715875 s 0.02545994625 s 1.00
jaxmd20 / DefOpt / tpu / PostRev 0.0188232662499999 s 0.018819655 s 1.00
jaxmd20 / DefOpt / tpu / BothRev 0.025572689375 s 0.025548431875 s 1.00
jaxmd20 / IDefOpt / tpu / PreRev 0.025472184375 s 0.025461729375 s 1.00
jaxmd20 / IDefOpt / tpu / PostRev 0.0183257174999999 s 0.018307651875 s 1.00
jaxmd20 / IDefOpt / tpu / BothRev 0.0255687 s 0.02554862875 s 1.00
jaxmd40 / JaXPipe / cpu / Primal 0.072995583 s 0.0684312 s 1.07
jaxmd40 / Jax / cpu / Primal 0.077871358 s 0.073527102 s 1.06
jaxmd40 / HLOOpt / cpu / Primal 0.0919285629999999 s 0.0889373889999999 s 1.03
jaxmd40 / PartOpt / cpu / Primal 0.074208656 s 0.074375692 s 1.00
jaxmd40 / IPartOpt / cpu / Primal 0.064872874 s 0.072943314 s 0.89
jaxmd40 / DefOpt / cpu / Primal 0.096048702 s 0.0914378459999999 s 1.05
jaxmd40 / IDefOpt / cpu / Primal 0.0894831 s 0.092339397 s 0.97
jaxmd40 / JaXPipe / cpu / Forward 0.170373676 s 0.1712232119999999 s 1.00
jaxmd40 / Jax / cpu / Forward 0.086743587 s 0.0903042959999999 s 0.96
jaxmd40 / HLOOpt / cpu / Forward 0.173024437 s 0.1781862639999999 s 0.97
jaxmd40 / PartOpt / cpu / Forward 0.163292465 s 0.164765115 s 0.99
jaxmd40 / IPartOpt / cpu / Forward 0.162668462 s 0.175371687 s 0.93
jaxmd40 / DefOpt / cpu / Forward 0.167398347 s 0.166921464 s 1.00
jaxmd40 / IDefOpt / cpu / Forward 0.161209102 s 0.166643084 s 0.97
jaxmd40 / JaXPipe / cpu / PreRev 0.243966287 s 0.225262923 s 1.08
jaxmd40 / JaXPipe / cpu / PostRev 0.144313544 s 0.142715861 s 1.01
jaxmd40 / JaXPipe / cpu / BothRev 0.236825549 s 0.221521487 s 1.07
jaxmd40 / Jax / cpu / BothRev 0.134210464 s 0.141799249 s 0.95
jaxmd40 / HLOOpt / cpu / PreRev 0.219472778 s 0.220571377 s 1.00
jaxmd40 / HLOOpt / cpu / PostRev 0.179709087 s 0.1786106069999999 s 1.01
jaxmd40 / HLOOpt / cpu / BothRev 0.2551370169999999 s 0.252948686 s 1.01
jaxmd40 / PartOpt / cpu / PreRev 0.237557499 s 0.227161047 s 1.05
jaxmd40 / PartOpt / cpu / PostRev 0.124745071 s 0.132065355 s 0.94
jaxmd40 / PartOpt / cpu / BothRev 0.25122594 s 0.239801865 s 1.05
jaxmd40 / IPartOpt / cpu / PreRev 0.223053461 s 0.216225257 s 1.03
jaxmd40 / IPartOpt / cpu / PostRev 0.133288232 s 0.131393485 s 1.01
jaxmd40 / IPartOpt / cpu / BothRev 0.247687632 s 0.2527059299999999 s 0.98
jaxmd40 / DefOpt / cpu / PreRev 0.2257147099999999 s 0.224775399 s 1.00
jaxmd40 / DefOpt / cpu / PostRev 0.167109407 s 0.171236742 s 0.98
jaxmd40 / DefOpt / cpu / BothRev 0.252800479 s 0.2603366589999999 s 0.97
jaxmd40 / IDefOpt / cpu / PreRev 0.2208882439999999 s 0.21852543 s 1.01
jaxmd40 / IDefOpt / cpu / PostRev 0.170210224 s 0.175843656 s 0.97
jaxmd40 / IDefOpt / cpu / BothRev 0.2327637 s 0.242549485 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 / JaXPipe / cuda / Primal 0.000279488 s 0.000281344 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 / cuda / Primal 0.000279168 s 0.000280544 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 0.0002863999999999 s 0.000288127 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 / cuda / Primal 0.00027888 s 0.0002802879999999 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 / cuda / Primal 0.0002790079999999 s 0.0002807669999999 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 / DefOpt / cuda / Primal 0.00028608 s 0.000287712 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 / IDefOpt / cuda / Primal 0.000288352 s 0.000288673 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 0.000557728 s 0.0005579199999999 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 0.000538879 s 0.0005402559999999 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 0.0005574719999999 s 0.000558048 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 0.000557056 s 0.000556992 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 0.000558368 s 0.00055888 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 0.000557728 s 0.000557985 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 0.000558208 s 0.000557952 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 0.001024416 s 0.0010287999999999 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 0.000984864 s 0.000987616 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 0.001026175 s 0.001027104 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 0.000988768 s 0.000989728 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 0.001012865 s 0.001014016 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 0.001037856 s 0.001041407 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 0.001011616 s 0.00101472 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 0.00102784 s 0.00103056 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 0.000976064 s 0.000978272 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 0.001028355 s 0.001029024 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 0.001026624 s 0.001029248 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 0.000976225 s 0.00097744 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 0.00102528 s 0.001028896 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 0.001022081 s 0.001025217 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 0.00096064 s 0.000963104 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 0.001024672 s 0.001026944 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 0.001021986 s 0.001022464 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 0.001020897 s 0.001023265 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 0.001022305 s 0.001024065 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 0.000130708 s 0.0001308955 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 / Primal 0.00012383425 s 0.00012357475 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 / Primal 0.0001600842499999 s 0.00015994075 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 / Primal 0.0001313199999999 s 0.0001310315 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 / Primal 0.00013863775 s 0.0001380562499999 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 / Primal 0.00014560525 s 0.000145431 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 / Primal 0.0001580635 s 0.0001580635 s 1
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 0.0002137719999999 s 0.0002134492499999 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 0.00026212075 s 0.000262756 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 0.00022022475 s 0.00022040825 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 / Forward 0.0002150175 s 0.0002148825 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 / Forward 0.000215695 s 0.00021563875 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 / Forward 0.000217985 s 0.00021784225 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 0.000216062 s 0.00021554975 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 / PreRev 0.00035520125 s 0.00035536275 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 0.000256249 s 0.00025649775 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 0.0003552665 s 0.00035530275 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 / BothRev 0.000256134 s 0.00025637775 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 0.0003551575 s 0.0003550855 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 0.00029047175 s 0.0002903284999999 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 0.00035539675 s 0.0003555837499999 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 0.0003548515 s 0.00035495925 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 0.0002717762499999 s 0.0002719545 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 / BothRev 0.0003546925 s 0.00035480775 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 0.0003553074999999 s 0.0003554125 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 0.00027166775 s 0.000271823 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 0.000355185 s 0.0003553045 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 0.0003573155 s 0.000357019 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 0.00028388325 s 0.00028346925 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 0.00035686225 s 0.0003572665 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 0.00035776925 s 0.000357612 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 0.00030046675 s 0.0003007119999999 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 0.00035785775 s 0.00035751075 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 0.002014537 s 0.000942452199888 s 2.14
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 0.001874894 s 0.0009337108000181 s 2.01
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 0.002008521 s 0.0010303778000888 s 1.95
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 0.001761161 s 0.0009287074000894 s 1.90
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 0.001543682 s 0.0009517532000245 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 / DefOpt / cpu / Primal 0.0017202679999999 s 0.0009927374001563 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 / IDefOpt / cpu / Primal 0.001921045 s 0.0009864354000455 s 1.95
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 0.004989449 s 0.0024338845998499 s 2.05
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 0.0052621149999999 s 0.0024766360000285 s 2.12
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 0.005029493 s 0.0023427183999956 s 2.15
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 0.005188018 s 0.002410401599991 s 2.15
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 0.005162447 s 0.0022836635998828 s 2.26
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 0.005029522 s 0.0023569746001157 s 2.13
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 0.004948357 s 0.0023437149997334 s 2.11
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 0.0094842289999999 s 0.0051605818000098 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 / JaXPipe / cpu / PostRev 0.009833693 s 0.0058353222000732 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 / BothRev 0.007870036 s 0.0058945830003722 s 1.34
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 0.008958305 s 0.0056922598001619 s 1.57
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 0.008127435 s 0.0058994788001655 s 1.38
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 0.008445199 s 0.0056784344000334 s 1.49
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 0.008108179 s 0.006528671200067 s 1.24
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 0.00774212 s 0.0035868383998604 s 2.16
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 0.008358503 s 0.0063951135998649 s 1.31
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 0.009540762 s 0.0056096246000379 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 / IPartOpt / cpu / PreRev 0.007322973 s 0.0056195833998572 s 1.30
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 0.009188155 s 0.0056347833999097 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 / BothRev 0.0091648169999999 s 0.005715845800114 s 1.60
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 0.008688635 s 0.0036831345998507 s 2.36
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 0.006714003 s 0.0055874166000648 s 1.20
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 0.008662615 s 0.0065726325998184 s 1.32
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 0.008276221 s 0.0036468901998887 s 2.27
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 0.007799209 s 0.0057078440000623 s 1.37
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 0.008331845 s 0.0058645370001613 s 1.42
neuralgcm_v1/deterministic_2_8_deg_inner_steps_24_outer_steps_4 / JaXPipe / cuda / Primal 1.702005693 s 1.702956951 s 1.00
neuralgcm_v1/deterministic_2_8_deg_inner_steps_24_outer_steps_4 / Jax / cuda / Primal 1.704635883 s 1.704677063 s 1.00
neuralgcm_v1/deterministic_2_8_deg_inner_steps_24_outer_steps_4 / HLOOpt / cuda / Primal 1.715270902 s 1.7155073010000002 s 1.00
neuralgcm_v1/deterministic_2_8_deg_inner_steps_24_outer_steps_4 / PartOpt / cuda / Primal 1.697321589 s 1.69709385 s 1.00
neuralgcm_v1/deterministic_2_8_deg_inner_steps_24_outer_steps_4 / IPartOpt / cuda / Primal 1.695032195 s 1.694971321 s 1.00
neuralgcm_v1/deterministic_2_8_deg_inner_steps_24_outer_steps_4 / DefOpt / cuda / Primal 1.666882187 s 1.665851588 s 1.00
neuralgcm_v1/deterministic_2_8_deg_inner_steps_24_outer_steps_4 / IDefOpt / cuda / Primal 1.910868471 s 1.912409822 s 1.00
neuralgcm_v1/deterministic_2_8_deg_inner_steps_24_outer_steps_4 / JaXPipe / tpu / Primal 3.038949814375 s 3.038985894375 s 1.00
neuralgcm_v1/deterministic_2_8_deg_inner_steps_24_outer_steps_4 / Jax / tpu / Primal 3.03955622 s 3.03952863125 s 1.00
neuralgcm_v1/deterministic_2_8_deg_inner_steps_24_outer_steps_4 / HLOOpt / tpu / Primal 3.121863075625 s 3.121900450625 s 1.00
neuralgcm_v1/deterministic_2_8_deg_inner_steps_24_outer_steps_4 / PartOpt / tpu / Primal 3.06030164 s 3.060244081875 s 1.00
neuralgcm_v1/deterministic_2_8_deg_inner_steps_24_outer_steps_4 / IPartOpt / tpu / Primal 3.060495848125 s 3.06051306875 s 1.00
neuralgcm_v1/deterministic_2_8_deg_inner_steps_24_outer_steps_4 / DefOpt / tpu / Primal 2.10247015875 s 2.1025194700000003 s 1.00
neuralgcm_v1/deterministic_2_8_deg_inner_steps_24_outer_steps_4 / IDefOpt / tpu / Primal 4.35654923375 s 4.35645801625 s 1.00
neuralgcm_v1/deterministic_2_8_deg_inner_steps_2_outer_steps_2 / JaXPipe / cpu / Primal 6.317055561 s 6.19071066 s 1.02
neuralgcm_v1/deterministic_2_8_deg_inner_steps_2_outer_steps_2 / Jax / cpu / Primal 6.172251312999999 s 6.268450071 s 0.98
neuralgcm_v1/deterministic_2_8_deg_inner_steps_2_outer_steps_2 / HLOOpt / cpu / Primal 6.254881232 s 6.097149982 s 1.03
neuralgcm_v1/deterministic_2_8_deg_inner_steps_2_outer_steps_2 / PartOpt / cpu / Primal 6.167481484 s 6.316660036999999 s 0.98
neuralgcm_v1/deterministic_2_8_deg_inner_steps_2_outer_steps_2 / IPartOpt / cpu / Primal 6.30376062 s 6.199206968 s 1.02
neuralgcm_v1/deterministic_2_8_deg_inner_steps_2_outer_steps_2 / DefOpt / cpu / Primal 2.5054433090000003 s 2.463648363 s 1.02
neuralgcm_v1/deterministic_2_8_deg_inner_steps_2_outer_steps_2 / IDefOpt / cpu / Primal 6.540719687 s 6.656476705 s 0.98
scatter_sum / JaXPipe / cuda / Primal 0.000009856 s 0.000010048 s 0.98
scatter_sum / Jax / cuda / Primal 0.00001008 s 0.000009888 s 1.02
scatter_sum / HLOOpt / cuda / Primal 0.000009856 s 0.000009952 s 0.99
scatter_sum / PartOpt / cuda / Primal 0.00001008 s 0.000009888 s 1.02
scatter_sum / IPartOpt / cuda / Primal 0.00001104 s 0.000010496 s 1.05
scatter_sum / DefOpt / cuda / Primal 0.000011424 s 0.000010464 s 1.09
scatter_sum / IDefOpt / cuda / Primal 0.000009888 s 0.000010112 s 0.98
scatter_sum / JaXPipe / cuda / Forward 0.000017088 s 0.000017503999999999997 s 0.98
scatter_sum / Jax / cuda / Forward 0.000017601 s 0.000017536 s 1.00
scatter_sum / HLOOpt / cuda / Forward 0.000017184 s 0.000017344 s 0.99
scatter_sum / PartOpt / cuda / Forward 0.000017247999999999998 s 0.000017472 s 0.99
scatter_sum / IPartOpt / cuda / Forward 0.000017631 s 0.000017568000000000002 s 1.00
scatter_sum / DefOpt / cuda / Forward 0.00001728 s 0.00001744 s 0.99
scatter_sum / IDefOpt / cuda / Forward 0.00001712 s 0.000017344 s 0.99
scatter_sum / JaXPipe / cuda / PreRev 0.000017056 s 0.000017568000000000002 s 0.97
scatter_sum / JaXPipe / cuda / PostRev 0.00001664 s 0.00001728 s 0.96
scatter_sum / JaXPipe / cuda / BothRev 0.000016768000000000003 s 0.0000184 s 0.91
scatter_sum / Jax / cuda / BothRev 0.000017056 s 0.000017247 s 0.99
scatter_sum / HLOOpt / cuda / PreRev 0.00001728 s 0.000017375999999999998 s 0.99
scatter_sum / HLOOpt / cuda / PostRev 0.000016896999999999998 s 0.000017216 s 0.98
scatter_sum / HLOOpt / cuda / BothRev 0.000017247999999999998 s 0.000017824 s 0.97
scatter_sum / PartOpt / cuda / PreRev 0.000017472 s 0.000020064 s 0.87
scatter_sum / PartOpt / cuda / PostRev 0.0000168 s 0.000017184 s 0.98
scatter_sum / PartOpt / cuda / BothRev 0.000017568000000000002 s 0.000017056 s 1.03
scatter_sum / IPartOpt / cuda / PreRev 0.000017152 s 0.000017726999999999998 s 0.97
scatter_sum / IPartOpt / cuda / PostRev 0.000017696 s 0.000019648 s 0.90
scatter_sum / IPartOpt / cuda / BothRev 0.000017152 s 0.00001744 s 0.98
scatter_sum / DefOpt / cuda / PreRev 0.000017247999999999998 s 0.000017728 s 0.97
scatter_sum / DefOpt / cuda / PostRev 0.00001744 s 0.000017375999999999998 s 1.00
scatter_sum / DefOpt / cuda / BothRev 0.000017216 s 0.000017471 s 0.99
scatter_sum / IDefOpt / cuda / PreRev 0.000017728 s 0.000017728 s 1
scatter_sum / IDefOpt / cuda / PostRev 0.000016736 s 0.00001712 s 0.98
scatter_sum / IDefOpt / cuda / BothRev 0.000017119 s 0.000017472 s 0.98
scatter_sum / JaXPipe / tpu / Primal 0.000001343625 s 0.000001343 s 1.00
scatter_sum / Jax / tpu / Primal 0.0000014043 s 0.0000014042750000000002 s 1.00
scatter_sum / HLOOpt / tpu / Primal 0.000001343475 s 0.00000134435 s 1.00
scatter_sum / PartOpt / tpu / Primal 0.000001404025 s 0.0000014044 s 1.00
scatter_sum / IPartOpt / tpu / Primal 0.000001343275 s 0.0000013437 s 1.00
scatter_sum / DefOpt / tpu / Primal 0.00000140405 s 0.000001404675 s 1.00
scatter_sum / IDefOpt / tpu / Primal 0.00000134325 s 0.00000134335 s 1.00
scatter_sum / JaXPipe / tpu / Forward 0.0000027019500000000004 s 0.000002697975 s 1.00
scatter_sum / Jax / tpu / Forward 0.0000027221 s 0.000002713525 s 1.00
scatter_sum / HLOOpt / tpu / Forward 0.000002707325 s 0.0000027010000000000005 s 1.00
scatter_sum / PartOpt / tpu / Forward 0.0000026926 s 0.000002683925 s 1.00
scatter_sum / IPartOpt / tpu / Forward 0.000002711075 s 0.000002698175 s 1.00
scatter_sum / DefOpt / tpu / Forward 0.00000268695 s 0.0000026904 s 1.00
scatter_sum / IDefOpt / tpu / Forward 0.000002711 s 0.00000270295 s 1.00
scatter_sum / JaXPipe / tpu / PreRev 0.0000026928000000000003 s 0.00000268745 s 1.00
scatter_sum / JaXPipe / tpu / PostRev 0.0000026872750000000005 s 0.000002688775 s 1.00
scatter_sum / JaXPipe / tpu / BothRev 0.000002698725 s 0.0000026938250000000004 s 1.00
scatter_sum / Jax / tpu / BothRev 0.000002741475 s 0.00000274215 s 1.00
scatter_sum / HLOOpt / tpu / PreRev 0.000002698775 s 0.000002695575 s 1.00
scatter_sum / HLOOpt / tpu / PostRev 0.0000027431250000000005 s 0.000002761025 s 0.99
scatter_sum / HLOOpt / tpu / BothRev 0.000002695275 s 0.0000026945750000000004 s 1.00
scatter_sum / PartOpt / tpu / PreRev 0.0000027447500000000003 s 0.0000027512 s 1.00
scatter_sum / PartOpt / tpu / PostRev 0.000002695475 s 0.000002691725 s 1.00
scatter_sum / PartOpt / tpu / BothRev 0.0000027447749999999995 s 0.0000027453 s 1.00
scatter_sum / IPartOpt / tpu / PreRev 0.0000026996 s 0.0000026937 s 1.00
scatter_sum / IPartOpt / tpu / PostRev 0.000002749525 s 0.0000027463500000000004 s 1.00
scatter_sum / IPartOpt / tpu / BothRev 0.000002697975 s 0.0000026924 s 1.00
scatter_sum / DefOpt / tpu / PreRev 0.00000274075 s 0.000002739525 s 1.00
scatter_sum / DefOpt / tpu / PostRev 0.000002711675 s 0.0000026955 s 1.01
scatter_sum / DefOpt / tpu / BothRev 0.000002739975 s 0.00000273795 s 1.00
scatter_sum / IDefOpt / tpu / PreRev 0.00000270325 s 0.0000026977 s 1.00
scatter_sum / IDefOpt / tpu / PostRev 0.00000274865 s 0.00000273565 s 1.00
scatter_sum / IDefOpt / tpu / BothRev 0.000002696575 s 0.0000027043 s 1.00
scatter_sum / JaXPipe / cpu / Primal 0.000015665 s 0.000007906760001787916 s 1.98
scatter_sum / Jax / cpu / Primal 0.000015958 s 0.000008491600056004245 s 1.88
scatter_sum / HLOOpt / cpu / Primal 0.000015695 s 0.00000850741997055593 s 1.84
scatter_sum / PartOpt / cpu / Primal 0.000015453 s 0.000008762860034039477 s 1.76
scatter_sum / IPartOpt / cpu / Primal 0.000015765999999999998 s 0.000008235420009441441 s 1.91
scatter_sum / DefOpt / cpu / Primal 0.000015765999999999998 s 0.000008393819953198545 s 1.88
scatter_sum / IDefOpt / cpu / Primal 0.000015472 s 0.000008244539985753363 s 1.88
scatter_sum / JaXPipe / cpu / Forward 0.000023147 s 0.000012941920012963236 s 1.79
scatter_sum / Jax / cpu / Forward 0.000022074 s 0.0000128758200116863 s 1.71
scatter_sum / HLOOpt / cpu / Forward 0.000022661 s 0.000013030260015511886 s 1.74
scatter_sum / PartOpt / cpu / Forward 0.000022276 s 0.00001253974001883762 s 1.78
scatter_sum / IPartOpt / cpu / Forward 0.000021786 s 0.000013102699940645834 s 1.66
scatter_sum / DefOpt / cpu / Forward 0.000022616 s 0.000012640539953281405 s 1.79
scatter_sum / IDefOpt / cpu / Forward 0.000022472 s 0.000012317439995968016 s 1.82
scatter_sum / JaXPipe / cpu / PreRev 0.00002341 s 0.00001348465992123238 s 1.74
scatter_sum / JaXPipe / cpu / PostRev 0.000022547 s 0.000012731879978673533 s 1.77
scatter_sum / JaXPipe / cpu / BothRev 0.000022338 s 0.000012610540015884909 s 1.77
scatter_sum / Jax / cpu / BothRev 0.000022252 s 0.000013955280082882384 s 1.59
scatter_sum / HLOOpt / cpu / PreRev 0.000022457 s 0.00001342508003290277 s 1.67
scatter_sum / HLOOpt / cpu / PostRev 0.000023001 s 0.000014285839952208337 s 1.61
scatter_sum / HLOOpt / cpu / BothRev 0.000022686000000000003 s 0.000012689840004895814 s 1.79
scatter_sum / PartOpt / cpu / PreRev 0.000022885 s 0.00001280814005440334 s 1.79
scatter_sum / PartOpt / cpu / PostRev 0.000022865 s 0.000012537939892354188 s 1.82
scatter_sum / PartOpt / cpu / BothRev 0.000022125 s 0.000013116259906382763 s 1.69
scatter_sum / IPartOpt / cpu / PreRev 0.000022604 s 0.000012832160045945784 s 1.76
scatter_sum / IPartOpt / cpu / PostRev 0.000022253 s 0.000012869779984612253 s 1.73
scatter_sum / IPartOpt / cpu / BothRev 0.000022728000000000003 s 0.000012823780089092909 s 1.77
scatter_sum / DefOpt / cpu / PreRev 0.000022877 s 0.000013038839988439576 s 1.75
scatter_sum / DefOpt / cpu / PostRev 0.000021823 s 0.00001323500006037648 s 1.65
scatter_sum / DefOpt / cpu / BothRev 0.000022387 s 0.0000122463000479911 s 1.83
scatter_sum / IDefOpt / cpu / PreRev 0.000022417 s 0.000013496480041794712 s 1.66
scatter_sum / IDefOpt / cpu / PostRev 0.000022295 s 0.000012729499976558144 s 1.75
scatter_sum / IDefOpt / cpu / BothRev 0.000022397 s 0.00001315758005148382 s 1.70
slicing / JaXPipe / cuda / Primal 0.000001888 s 0.000001887 s 1.00
slicing / Jax / cuda / Primal 0.000001887 s 0.000001887 s 1
slicing / HLOOpt / cuda / Primal 0.000001887 s 0.000001888 s 1.00
slicing / PartOpt / cuda / Primal 0.000001888 s 0.000001887 s 1.00
slicing / IPartOpt / cuda / Primal 0.000001887 s 0.000001887 s 1
slicing / DefOpt / cuda / Primal 0.000001887 s 0.000001887 s 1
slicing / IDefOpt / cuda / Primal 0.000001887 s 0.000001888 s 1.00
slicing / JaXPipe / cuda / Forward 0.000010208 s 0.000010048 s 1.02
slicing / Jax / cuda / Forward 0.000010112 s 0.000010049 s 1.01
slicing / HLOOpt / cuda / Forward 0.00001008 s 0.000010272 s 0.98
slicing / PartOpt / cuda / Forward 0.00001056 s 0.000010304 s 1.02
slicing / IPartOpt / cuda / Forward 0.000009856 s 0.000010175 s 0.97
slicing / DefOpt / cuda / Forward 0.000010208 s 0.000009696 s 1.05
slicing / IDefOpt / cuda / Forward 0.000010048 s 0.000009824 s 1.02
slicing / JaXPipe / cuda / PreRev 0.000009984 s 0.000010049 s 0.99
slicing / JaXPipe / cuda / PostRev 0.000009856 s 0.000010272 s 0.96
slicing / JaXPipe / cuda / BothRev 0.000010144 s 0.000009984 s 1.02
slicing / Jax / cuda / BothRev 0.000010143 s 0.000010336 s 0.98
slicing / HLOOpt / cuda / PreRev 0.000009824 s 0.000010176 s 0.97
slicing / HLOOpt / cuda / PostRev 0.000009664 s 0.000010304 s 0.94
slicing / HLOOpt / cuda / BothRev 0.000009857 s 0.00001008 s 0.98
slicing / PartOpt / cuda / PreRev 0.000009536 s 0.000009952 s 0.96
slicing / PartOpt / cuda / PostRev 0.000009984 s 0.000009888 s 1.01
slicing / PartOpt / cuda / BothRev 0.000011072 s 0.000010112 s 1.09
slicing / IPartOpt / cuda / PreRev 0.000009952 s 0.000009984 s 1.00
slicing / IPartOpt / cuda / PostRev 0.000009887 s 0.000009952 s 0.99
slicing / IPartOpt / cuda / BothRev 0.000010144 s 0.000009984 s 1.02
slicing / DefOpt / cuda / PreRev 0.000010208 s 0.000010145 s 1.01
slicing / DefOpt / cuda / PostRev 0.000009632 s 0.000010272 s 0.94
slicing / DefOpt / cuda / BothRev 0.00000992 s 0.000010176 s 0.97
slicing / IDefOpt / cuda / PreRev 0.000010112 s 0.000010272 s 0.98
slicing / IDefOpt / cuda / PostRev 0.000010016 s 0.000010111 s 0.99
slicing / IDefOpt / cuda / BothRev 0.000009984 s 0.00000976 s 1.02
slicing / JaXPipe / tpu / Primal 0.000001018075 s 9.70875e-7 s 1.05
slicing / Jax / tpu / Primal 9.65225e-7 s 9.697e-7 s 1.00
slicing / HLOOpt / tpu / Primal 0.00000102025 s 9.6505e-7 s 1.06
slicing / PartOpt / tpu / Primal 9.704e-7 s 9.71275e-7 s 1.00
slicing / IPartOpt / tpu / Primal 0.000001018 s 9.67175e-7 s 1.05
slicing / DefOpt / tpu / Primal 9.73e-7 s 9.754e-7 s 1.00
slicing / IDefOpt / tpu / Primal 0.000001019225 s 9.7415e-7 s 1.05
slicing / JaXPipe / tpu / Forward 0.00000140195 s 0.000001414775 s 0.99
slicing / Jax / tpu / Forward 0.00000146755 s 0.0000014294000000000002 s 1.03
slicing / HLOOpt / tpu / Forward 0.000001509825 s 0.00000151655 s 1.00
slicing / PartOpt / tpu / Forward 0.000001486675 s 0.0000014375750000000002 s 1.03
slicing / IPartOpt / tpu / Forward 0.00000151645 s 0.0000015182 s 1.00
slicing / DefOpt / tpu / Forward 0.000001495675 s 0.00000144075 s 1.04
slicing / IDefOpt / tpu / Forward 0.0000015095249999999995 s 0.00000152035 s 0.99
slicing / JaXPipe / tpu / PreRev 0.00000257065 s 0.000002381825 s 1.08
slicing / JaXPipe / tpu / PostRev 0.0000025186 s 0.000002513125 s 1.00
slicing / JaXPipe / tpu / BothRev 0.0000025813750000000003 s 0.0000023991 s 1.08
slicing / Jax / tpu / BothRev 0.00000253735 s 0.000002529925 s 1.00
slicing / HLOOpt / tpu / PreRev 0.0000025897 s 0.000002401425 s 1.08
slicing / HLOOpt / tpu / PostRev 0.000002536125 s 0.0000025351750000000003 s 1.00
slicing / HLOOpt / tpu / BothRev 0.0000025919750000000003 s 0.0000023994 s 1.08
slicing / PartOpt / tpu / PreRev 0.000002532725 s 0.000002540825 s 1.00
slicing / PartOpt / tpu / PostRev 0.0000025911 s 0.00000240305 s 1.08
slicing / PartOpt / tpu / BothRev 0.000002556925 s 0.000002537425 s 1.01
slicing / IPartOpt / tpu / PreRev 0.0000025867 s 0.00000240415 s 1.08
slicing / IPartOpt / tpu / PostRev 0.0000025371500000000003 s 0.0000025435000000000003 s 1.00
slicing / IPartOpt / tpu / BothRev 0.0000025869500000000003 s 0.000002398425 s 1.08
slicing / DefOpt / tpu / PreRev 0.000002541525 s 0.0000025363 s 1.00
slicing / DefOpt / tpu / PostRev 0.00000258315 s 0.000002393275 s 1.08
slicing / DefOpt / tpu / BothRev 0.000002540475 s 0.000002544625 s 1.00
slicing / IDefOpt / tpu / PreRev 0.0000025843 s 0.000002406925 s 1.07
slicing / IDefOpt / tpu / PostRev 0.000002545825 s 0.0000025308 s 1.01
slicing / IDefOpt / tpu / BothRev 0.0000025863750000000003 s 0.00000240595 s 1.07
slicing / JaXPipe / cpu / Primal 0.000012766 s 0.000007376100038527511 s 1.73
slicing / Jax / cpu / Primal 0.000012574 s 0.000006616539958486101 s 1.90
slicing / HLOOpt / cpu / Primal 0.000012487 s 0.0000066052799411409065 s 1.89
slicing / PartOpt / cpu / Primal 0.000012425 s 0.000006780179974157363 s 1.83
slicing / IPartOpt / cpu / Primal 0.000012599 s 0.000006698280012642499 s 1.88
slicing / DefOpt / cpu / Primal 0.000012671 s 0.000006589420008822344 s 1.92
slicing / IDefOpt / cpu / Primal 0.000012486 s 0.000006755839985999046 s 1.85
slicing / JaXPipe / cpu / Forward 0.000016743000000000002 s 0.000010073260018543806 s 1.66
slicing / Jax / cpu / Forward 0.000016698999999999997 s 0.000010152760005439631 s 1.64
slicing / HLOOpt / cpu / Forward 0.000016549 s 0.000011093820012320066 s 1.49
slicing / PartOpt / cpu / Forward 0.000016595 s 0.00001048930002070847 s 1.58
slicing / IPartOpt / cpu / Forward 0.000016715000000000002 s 0.000010540679995756364 s 1.59
slicing / DefOpt / cpu / Forward 0.000016714 s 0.000010364099998696475 s 1.61
slicing / IDefOpt / cpu / Forward 0.000016915 s 0.000010913759979302997 s 1.55
slicing / JaXPipe / cpu / PreRev 0.000017919999999999998 s 0.000010723480008891784 s 1.67
slicing / JaXPipe / cpu / PostRev 0.000017826000000000002 s 0.00001101179987017531 s 1.62
slicing / JaXPipe / cpu / BothRev 0.000017419 s 0.000011341339886712376 s 1.54
slicing / Jax / cpu / BothRev 0.000017628 s 0.000011364640085957944 s 1.55
slicing / HLOOpt / cpu / PreRev 0.000017368 s 0.00001158138000391773 s 1.50
slicing / HLOOpt / cpu / PostRev 0.000017298 s 0.00001279515990972868 s 1.35
slicing / HLOOpt / cpu / BothRev 0.000017533 s 0.000010748420045274545 s 1.63
slicing / PartOpt / cpu / PreRev 0.000017534000000000002 s 0.00001081095997506054 s 1.62
slicing / PartOpt / cpu / PostRev 0.000017420000000000003 s 0.000010711120103223947 s 1.63
slicing / PartOpt / cpu / BothRev 0.00001754 s 0.000010752960006357173 s 1.63
slicing / IPartOpt / cpu / PreRev 0.000017336 s 0.00001082108001355664 s 1.60
slicing / IPartOpt / cpu / PostRev 0.000017392999999999998 s 0.00001111292003770359 s 1.57
slicing / IPartOpt / cpu / BothRev 0.000016988 s 0.000010380180010542971 s 1.64
slicing / DefOpt / cpu / PreRev 0.000017718000000000002 s 0.000010644059839250986 s 1.66
slicing / DefOpt / cpu / PostRev 0.000017292 s 0.000010868580047826982 s 1.59
slicing / DefOpt / cpu / BothRev 0.000017554 s 0.000010702139952627477 s 1.64
slicing / IDefOpt / cpu / PreRev 0.000017471 s 0.000010620120028761447 s 1.65
slicing / IDefOpt / cpu / PostRev 0.000017171 s 0.000011314000075799414 s 1.52
slicing / IDefOpt / cpu / BothRev 0.000017204 s 0.00001076451997505501 s 1.60
sum / JaXPipe / cuda / Primal 0.000002111 s 0.000002047 s 1.03
sum / Jax / cuda / Primal 0.000002111 s 0.000002047 s 1.03
sum / HLOOpt / cuda / Primal 0.000002111 s 0.000002048 s 1.03
sum / PartOpt / cuda / Primal 0.000002111 s 0.000002048 s 1.03
sum / IPartOpt / cuda / Primal 0.000002111 s 0.000002047 s 1.03
sum / DefOpt / cuda / Primal 0.00000208 s 0.000002048 s 1.02
sum / IDefOpt / cuda / Primal 0.000002111 s 0.000002048 s 1.03
sum / JaXPipe / cuda / Forward 0.000010913 s 0.000010144 s 1.08
sum / Jax / cuda / Forward 0.000010272 s 0.000010176 s 1.01
sum / HLOOpt / cuda / Forward 0.00001056 s 0.000010272 s 1.03
sum / PartOpt / cuda / Forward 0.00001056 s 0.000010113 s 1.04
sum / IPartOpt / cuda / Forward 0.000010528 s 0.000010049 s 1.05
sum / DefOpt / cuda / Forward 0.000010496 s 0.000009984 s 1.05
sum / IDefOpt / cuda / Forward 0.000010368 s 0.000010368 s 1
sum / JaXPipe / cuda / PreRev 0.000009792 s 0.000009728 s 1.01
sum / JaXPipe / cuda / PostRev 0.000010272 s 0.000010048 s 1.02
sum / JaXPipe / cuda / BothRev 0.000010208 s 0.000009984 s 1.02
sum / Jax / cuda / BothRev 0.000009856 s 0.000010208 s 0.97
sum / HLOOpt / cuda / PreRev 0.000009984 s 0.000010144 s 0.98
sum / HLOOpt / cuda / PostRev 0.000009248 s 0.000010016 s 0.92
sum / HLOOpt / cuda / BothRev 0.000009792 s 0.000009825 s 1.00
sum / PartOpt / cuda / PreRev 0.000010208 s 0.000009888 s 1.03
sum / PartOpt / cuda / PostRev 0.000009568 s 0.000010112 s 0.95
sum / PartOpt / cuda / BothRev 0.00000976 s 0.000010144 s 0.96
sum / IPartOpt / cuda / PreRev 0.000009825 s 0.000009984 s 0.98
sum / IPartOpt / cuda / PostRev 0.00000992 s 0.000009888 s 1.00
sum / IPartOpt / cuda / BothRev 0.00000944 s 0.000009952 s 0.95
sum / DefOpt / cuda / PreRev 0.000010336 s 0.000010112 s 1.02
sum / DefOpt / cuda / PostRev 0.000010177 s 0.000009951 s 1.02
sum / DefOpt / cuda / BothRev 0.000009952 s 0.000010272 s 0.97
sum / IDefOpt / cuda / PreRev 0.000009792 s 0.000010112 s 0.97
sum / IDefOpt / cuda / PostRev 0.00001008 s 0.0000096 s 1.05
sum / IDefOpt / cuda / BothRev 0.00000976 s 0.000010304 s 0.95
sum / JaXPipe / tpu / Primal 5.10575e-7 s 5.103250000000001e-7 s 1.00
sum / Jax / tpu / Primal 5.4685e-7 s 5.47075e-7 s 1.00
sum / HLOOpt / tpu / Primal 5.10625e-7 s 5.0995e-7 s 1.00
sum / PartOpt / tpu / Primal 5.46975e-7 s 5.469999999999999e-7 s 1.00
sum / IPartOpt / tpu / Primal 5.1035e-7 s 5.10525e-7 s 1.00
sum / DefOpt / tpu / Primal 5.471999999999999e-7 s 5.47125e-7 s 1.00
sum / IDefOpt / tpu / Primal 5.106499999999999e-7 s 5.106e-7 s 1.00
sum / JaXPipe / tpu / Forward 0.00000155135 s 0.0000015465 s 1.00
sum / Jax / tpu / Forward 0.00000149875 s 0.0000014981 s 1.00
sum / HLOOpt / tpu / Forward 0.0000015379 s 0.0000015301999999999998 s 1.01
sum / PartOpt / tpu / Forward 0.0000015006 s 0.0000014973 s 1.00
sum / IPartOpt / tpu / Forward 0.0000015319250000000002 s 0.000001534625 s 1.00
sum / DefOpt / tpu / Forward 0.000001496875 s 0.00000150475 s 0.99
sum / IDefOpt / tpu / Forward 0.0000015317500000000002 s 0.00000153635 s 1.00
sum / JaXPipe / tpu / PreRev 0.000001054475 s 0.000001002975 s 1.05
sum / JaXPipe / tpu / PostRev 0.000001083475 s 0.0000010413500000000002 s 1.04
sum / JaXPipe / tpu / BothRev 0.0000010579 s 0.0000010012999999999998 s 1.06
sum / Jax / tpu / BothRev 0.0000010977 s 0.0000010361 s 1.06
sum / HLOOpt / tpu / PreRev 0.00000104765 s 0.0000010052 s 1.04
sum / HLOOpt / tpu / PostRev 0.000001094325 s 0.000001038575 s 1.05
sum / HLOOpt / tpu / BothRev 0.000001048425 s 0.0000010065 s 1.04
sum / PartOpt / tpu / PreRev 0.0000010895499999999998 s 0.000001035675 s 1.05
sum / PartOpt / tpu / PostRev 0.0000010536 s 0.000001005275 s 1.05
sum / PartOpt / tpu / BothRev 0.000001090675 s 0.000001038525 s 1.05
sum / IPartOpt / tpu / PreRev 0.00000105495 s 9.99775e-7 s 1.06
sum / IPartOpt / tpu / PostRev 0.000001089725 s 0.0000010369000000000002 s 1.05
sum / IPartOpt / tpu / BothRev 0.000001051325 s 0.000001001675 s 1.05
sum / DefOpt / tpu / PreRev 0.0000010858 s 0.000001037525 s 1.05
sum / DefOpt / tpu / PostRev 0.00000104795 s 0.0000010019 s 1.05
sum / DefOpt / tpu / BothRev 0.000001090275 s 0.000001035175 s 1.05
sum / IDefOpt / tpu / PreRev 0.0000010545 s 0.000001003675 s 1.05
sum / IDefOpt / tpu / PostRev 0.000001088925 s 0.000001035425 s 1.05
sum / IDefOpt / tpu / BothRev 0.0000010465 s 0.0000010136 s 1.03
sum / JaXPipe / cpu / Primal 0.000014626 s 0.000008208279996324564 s 1.78
sum / Jax / cpu / Primal 0.000014473 s 0.000008321959958266234 s 1.74
sum / HLOOpt / cpu / Primal 0.000013809 s 0.000008680300034029642 s 1.59
sum / PartOpt / cpu / Primal 0.00001424 s 0.00000838275993373827 s 1.70
sum / IPartOpt / cpu / Primal 0.000014436 s 0.00000831879990073503 s 1.74
sum / DefOpt / cpu / Primal 0.000014459 s 0.000008195060017897048 s 1.76
sum / IDefOpt / cpu / Primal 0.000014184 s 0.000008042199933697702 s 1.76
sum / JaXPipe / cpu / Forward 0.000019922 s 0.000012547039968922036 s 1.59
sum / Jax / cpu / Forward 0.000019391 s 0.000012595159932971 s 1.54
sum / HLOOpt / cpu / Forward 0.000019642 s 0.000012731680035358297 s 1.54
sum / PartOpt / cpu / Forward 0.00002036 s 0.000012554319928312907 s 1.62
sum / IPartOpt / cpu / Forward 0.000020027 s 0.000012770499943144386 s 1.57
sum / DefOpt / cpu / Forward 0.000019755000000000003 s 0.00001245491997906356 s 1.59
sum / IDefOpt / cpu / Forward 0.000019994 s 0.000012343320031504846 s 1.62
sum / JaXPipe / cpu / PreRev 0.000019048 s 0.000012384859983285424 s 1.54
sum / JaXPipe / cpu / PostRev 0.000018467 s 0.00001178538008389296 s 1.57
sum / JaXPipe / cpu / BothRev 0.000018881 s 0.000012185880059405465 s 1.55
sum / Jax / cpu / BothRev 0.000018662 s 0.00001187664001918165 s 1.57
sum / HLOOpt / cpu / PreRev 0.000018744 s 0.000011666739937936657 s 1.61
sum / HLOOpt / cpu / PostRev 0.000018936000000000003 s 0.000013745919950451937 s 1.38
sum / HLOOpt / cpu / BothRev 0.00001859 s 0.000011835500063170911 s 1.57
sum / PartOpt / cpu / PreRev 0.000018961 s 0.000011959500097873388 s 1.59
sum / PartOpt / cpu / PostRev 0.00001867 s 0.00001190447996123112 s 1.57
sum / PartOpt / cpu / BothRev 0.000018745 s 0.00001176630008558277 s 1.59
sum / IPartOpt / cpu / PreRev 0.000018849 s 0.0000117687799865962 s 1.60
sum / IPartOpt / cpu / PostRev 0.000018692 s 0.000011750659941753838 s 1.59
sum / IPartOpt / cpu / BothRev 0.000018699 s 0.000011083579938713228 s 1.69
sum / DefOpt / cpu / PreRev 0.000018751 s 0.00001203769996209303 s 1.56
sum / DefOpt / cpu / PostRev 0.000018359 s 0.000011714959982782605 s 1.57
sum / DefOpt / cpu / BothRev 0.000018761 s 0.000011712280047504464 s 1.60
sum / IDefOpt / cpu / PreRev 0.000018983 s 0.000011774200029321946 s 1.61
sum / IDefOpt / cpu / PostRev 0.000018407 s 0.000011629780019575264 s 1.58
sum / IDefOpt / cpu / BothRev 0.000018614 s 0.000012202980087749893 s 1.53
value_and_grad / JaXPipe / cuda / Primal 0.000033568 s 0.000033984 s 0.99
value_and_grad / Jax / cuda / Primal 0.00003424 s 0.00003968 s 0.86
value_and_grad / HLOOpt / cuda / Primal 0.00003264 s 0.000034176 s 0.96
value_and_grad / PartOpt / cuda / Primal 0.00003264 s 0.000034049000000000006 s 0.96
value_and_grad / IPartOpt / cuda / Primal 0.00003728 s 0.000034016 s 1.10
value_and_grad / DefOpt / cuda / Primal 0.000033696 s 0.000034016 s 0.99
value_and_grad / IDefOpt / cuda / Primal 0.000034176 s 0.00003424 s 1.00
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 0.00002316 s 0.000015194400020845931 s 1.52
value_and_grad / Jax / cpu / Primal 0.000022256 s 0.000015133479901123791 s 1.47
value_and_grad / HLOOpt / cpu / Primal 0.000022601 s 0.00001517563994639204 s 1.49
value_and_grad / PartOpt / cpu / Primal 0.000022468 s 0.000014475799962383465 s 1.55
value_and_grad / IPartOpt / cpu / Primal 0.000022951 s 0.00001491770000939141 s 1.54
value_and_grad / DefOpt / cpu / Primal 0.000022582 s 0.000014694860037707258 s 1.54
value_and_grad / IDefOpt / cpu / Primal 0.000022694 s 0.000015206619955279168 s 1.49

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

wip

wip

testing

wip

wip

done?

file

remove unneeded files

actually run test

revert isl

dep dialects

missingheader

stray file
@wsmoses wsmoses merged commit 2088acf into main Jan 8, 2026
25 of 27 checks passed
@wsmoses wsmoses deleted the remove-atomics-v2 branch January 8, 2026 12:35
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