Commit Graph

Select branches

Hide Pull Requests

geohot/updates

gh-pages

master

remove_anchors

#1

#2

#3

v0.10.0

v0.10.1

v0.10.2

v0.10.3

v0.11.0

v0.12.0

v0.13.0

v0.4.0

v0.5.0

v0.6.0

v0.7.0

v0.8.0

v0.9.0

v0.9.1

v0.9.2

Mono Color

• 5358b0904b update uop_given_valid if a node becomes const (#9604) chenyuandGitHub 2025-03-27 14:57:46 -04:00
• a187dfd3df bert BEAM_UOPS_MAX 3000->4000 (#9603) chenyuandGitHub 2025-03-27 11:58:47 -04:00
• 088a677e25 rescale to fit viz graph [pr] (#9599) qazalandGitHub 2025-03-27 23:33:51 +08:00
• 3737821b9e prepare for clang graph (#9600) nimlgenandGitHub 2025-03-27 20:09:37 +07:00
• bf94924d5a fix viz with nested graph_rewrite (#9595) qazalandGitHub 2025-03-27 13:14:28 +08:00
• c011751b41 statically define viz arrow heads (#9594) qazalandGitHub 2025-03-27 12:22:04 +08:00
• 0877497bad hotfix: use captured uops in viz render [pr] (#9593) qazalandGitHub 2025-03-27 11:52:12 +08:00
• e5ff7b23d7 refactor to @track_matches + add failing test_nested_rewrite (#9592) qazalandGitHub 2025-03-27 11:11:56 +08:00
• 62888614f6 lower bert eval bs to 24 (#9590) chenyuandGitHub 2025-03-26 21:25:23 -04:00
• dc9da1d917 memplan into one buffer (#9526) nimlgenandGitHub 2025-03-27 01:46:50 +07:00
• 8b717c345c cache viz worker at launch (#9589) qazalandGitHub 2025-03-27 01:10:02 +08:00
• d62ced8981 symbolic -> symbolic_flat (#9588) George HotzandGitHub 2025-03-26 23:34:43 +08:00
• 8aaa5e1ec5 generate the individual indexes (#9587) George HotzandGitHub 2025-03-26 22:32:06 +08:00
• 5c6cd884e3 multiple simplifies is faster [pr] (#9586) George HotzandGitHub 2025-03-26 21:42:52 +08:00
• 1e6e75e39a little changes from dsp branch (#9582) George HotzandGitHub 2025-03-26 20:01:21 +08:00
• e88a640ca5 fix _access_resources for offset buffers (#9580) nimlgenandGitHub 2025-03-26 18:42:43 +07:00
• 7b865ed03d use tuple in isinstance for type checking (#9583) AndreyandGitHub 2025-03-26 08:36:48 -03:00
• 9115ce8860 linearizer fixups from DSP branch (#9581) George HotzandGitHub 2025-03-26 18:28:15 +08:00
• e799df537e prep viz UI cleanup for grid scales (#9579) qazalandGitHub 2025-03-26 17:48:15 +08:00
• ccbcdca473 add memplanner tests (#9577) nimlgenandGitHub 2025-03-26 10:59:39 +07:00
• c03dadfcb9 add TORCHVIZ=1 to beautiful_mnist_torch (#9576) qazalandGitHub 2025-03-26 11:17:08 +08:00
• 93bcb974c5 select torch device in examples/beautiful_mnist_torch.py (#9575) qazalandGitHub 2025-03-26 11:01:25 +08:00
• 2c32126fc8 am: AMRegister refactor (#9572) uuuvnandGitHub 2025-03-25 22:52:40 +05:00
• cddd750d68 add a failed test case for jit/nojit rand [pr] (#9574) chenyuandGitHub 2025-03-25 13:32:44 -04:00
• 4cf2b68ca8 am_smi: fix init for newer versions (#9559) nimlgenandGitHub 2025-03-25 23:48:05 +07:00
• a6a5c0aec5 add NULL=1 backend (#9573) qazalandGitHub 2025-03-25 22:20:52 +08:00
• b60d9976b4 better yaxis formatting in viz memory graph (#9570) qazalandGitHub 2025-03-25 16:50:22 +08:00
• faf3b5b245 display kernel metadata in memory viz (#9569) qazalandGitHub 2025-03-25 13:14:54 +08:00
• 52301fe68e move Buffer refcount increment out of schedule.py (#9564) qazalandGitHub 2025-03-25 12:08:27 +08:00
• 262f5a2bd3 hotfix: replace link in viz/readme (#9568) qazalandGitHub 2025-03-25 10:24:49 +08:00
• 6427272bf6 minor update to rand [pr] (#9566) chenyuandGitHub 2025-03-24 18:49:50 -04:00
• b0e070e737 remove MOCKGPU workaround in rand (#9565) chenyuandGitHub 2025-03-24 17:49:45 -04:00
• d7c754ce49 failing test for UOp buffer ref count (#9563) qazalandGitHub 2025-03-25 00:10:48 +08:00
• f90001e1a6 amd llvm render (no_comgr prereq) (#9543) 3 people 2025-03-24 22:50:51 +08:00
• 4f5e03bd60 better fix inplace detach (#9557) Priyank PatelandGitHub 2025-03-24 07:50:28 -07:00
• 1c40873962 show buffer info in memory viz (#9562) qazalandGitHub 2025-03-24 22:12:30 +08:00
• efaee75656 start viz of memory usage (#9561) qazalandGitHub 2025-03-24 19:05:35 +08:00
• 1cfe6d02fe refactor uop_to_json to return a dict [pr] (#9560) qazalandGitHub 2025-03-24 16:38:17 +08:00
• edf9e1bf8d am: move out soc21 to a sep module (#9551) nimlgenandGitHub 2025-03-24 14:17:42 +07:00
• 74d98eafb8 add onnx frontend stub [pr] (#9558) George HotzandGitHub 2025-03-24 12:24:34 +08:00
• de7d6cec3a hotfix: DEBUG 5 prints the ast geohot 2025-03-24 11:43:11 +08:00
• ba41076e94 update embedding test to not use dtypes.long [pr] (#9556) chenyuandGitHub 2025-03-23 21:33:38 -04:00
• c965f4c20b update bert config (#9555) chenyuandGitHub 2025-03-23 16:14:41 -04:00
• d734e24c01 minor WEBGPU_PATH cleanup [pr] (#9552) chenyuandGitHub 2025-03-23 09:10:02 -04:00
• 7ce7fe0574 Refactor webgpu_dawn lib finding (#9547) Ahmed HarmoucheandGitHub 2025-03-23 13:23:29 +01:00
• c631c72f22 HCQ: Increment timeline signal before submitting (#9550) uuuvnandGitHub 2025-03-23 16:30:38 +05:00
• d5667419af am: move out pte creation logic (#9548) nimlgenandGitHub 2025-03-23 18:29:10 +07:00
• 309afa20b7 add Tensor.max_unpool2d (#9518) geohotstanandGitHub 2025-03-23 00:11:33 +08:00
• bdd44d4255 Fix DSP transcendentals (#9542) quortusandGitHub 2025-03-22 04:08:18 +01:00
• eddafb84e5 Bugfix for TC=3 (#9464) Ignacio SicaandGitHub 2025-03-21 16:43:42 -07:00
• 6da78164f9 assert Kernel ast.op to be Ops.SINK [pr] (#9539) chenyuandGitHub 2025-03-21 18:09:44 -04:00
• c33679c47b increase size in test_multinomial_counterexample (#9540) chenyuandGitHub 2025-03-21 17:46:52 -04:00
• 1a1087e3a0 cleanups on losses and dataset tests (#9538) Francis LataandGitHub 2025-03-21 17:03:18 -04:00
• 8cbe4009fc RetinaNet losses (#9536) Francis LataandGitHub 2025-03-21 15:52:54 -04:00
• e6389184c5 update comment for retinanet dataloader implementations (#9534) 3 people 2025-03-21 15:07:45 -04:00
• ee3d313b34 Revert "update ruff to 0.11.2 (#9531)" (#9535) chenyuandGitHub 2025-03-21 14:52:25 -04:00
• b46b8ee15e add a flag to log when beam surpassed max limit [pr] (#9533) chenyuandGitHub 2025-03-21 13:37:02 -04:00
• eb95825eea RetinaNet dataloader (#9442) Francis LataandGitHub 2025-03-21 13:36:41 -04:00
• 58206fa8a9 add amd llvm compiler (#9519) 4 people 2025-03-21 23:13:27 +08:00
• d8d65e2747 update ruff to 0.11.2 (#9531) chenyuandGitHub 2025-03-21 10:32:59 -04:00
• ee3ed73ed1 add reorder_view matcher to scheduler [pr] (#9528) qazalandGitHub 2025-03-21 17:46:20 +08:00
• 8e555c586c switch quantization to unsigned/unsigned + add Ops.REDUCE (#9527) George HotzandGitHub 2025-03-21 17:02:37 +08:00
• a35b0a88bf am: just rename and reorder ip init funcs (#9504) nimlgenandGitHub 2025-03-21 15:57:32 +08:00
• 8a131ab271 am: allow allocations as small as a page (#9523) nimlgenandGitHub 2025-03-21 15:53:32 +08:00
• 3ad3ac4d1e Change dtypes.int to dtypes.ints (#9517) Sieds LyklesandGitHub 2025-03-20 22:24:26 +01:00
• b9fab9b914 pin ruff to 0.11.0 in CI (#9520) chenyuandGitHub 2025-03-20 13:12:50 -04:00
• 3c5161b4cb add validation of the bounds of Ops.INDEX (#9503) 3 people 2025-03-20 12:15:55 +08:00
• 0b20f91ce7 remove move_mask from the devectorizer (#9511) qazalandGitHub 2025-03-20 11:53:12 +08:00
• 9302738263 hotfix: more consistent wgsl.py spacing + cleanups [pr] (#9515) qazalandGitHub 2025-03-20 11:07:15 +08:00
• 68053d0510 dsp stuff / sniff ioctls from snpe (#9490) 3 people 2025-03-20 10:38:23 +08:00
• 2223b93338 add UPat.or_casted [pr] (#9513) qazalandGitHub 2025-03-20 10:08:32 +08:00
• 1839e8c9b3 place masks in INDEX for TestGatedStoreRewrite [pr] (#9512) qazalandGitHub 2025-03-20 09:46:53 +08:00
• bd731a8624 AMDCompiler refactor (no_comgr prereq) (#9497) 3 people 2025-03-20 09:44:07 +08:00
• 8c0d0a122c Add return_indices to max_pool (#9506) geohotstanandGitHub 2025-03-20 03:25:37 +08:00
• 189f62d44f add rounding to tqdm unit scale (#9507) chenyuandGitHub 2025-03-19 12:08:46 -04:00
• a5c971ff3a am: prereqs for rdna4 1/n (#9495) nimlgenandGitHub 2025-03-19 17:14:57 +08:00
• 1e5d9ad8f7 extra/gemm/max_matmul: start of custom kernels for GEMM (#6926) Francis LamandGitHub 2025-03-19 00:04:57 -07:00
• 865f23dd7b olmoe memory usage cleanups geohot 2025-03-19 12:28:18 +08:00
• 2c87a22cf2 fix prg size calculation when there are adjacent mapped ranges (#9498) 3 people 2025-03-19 11:55:03 +08:00
• 1d71436e6a use libllvm19 in ci (#9494) 3 people 2025-03-19 11:53:32 +08:00
• a95b489a55 nanoGPT train works with tiny torch backend (#9283) 3 people 2025-03-19 11:51:02 +08:00
• f8976dd2eb enable more webgpu tests (#9502) chenyuandGitHub 2025-03-18 23:03:54 -04:00
• ae688e4103 simple failing test for scheduling parallel reduce [pr] (#9501) qazalandGitHub 2025-03-19 10:52:13 +08:00
• e4dad99145 nn.state docs cleanup (#8332) 4 people 2025-03-18 22:16:40 +01:00
• 1ea4876dfa olmoe touchups (#9499) chenyuandGitHub 2025-03-18 15:25:45 -04:00
• f7506c6c25 JIT OLMoE (#9396) geohotstanandGitHub 2025-03-19 02:49:02 +08:00
• 5c56cac0a0 MI300 mfma support (#9417) Ignacio SicaandGitHub 2025-03-18 14:33:30 -03:00
• 5500887eed improve reproducibility of WebGPU CI puppeteer test (#9496) hoovedandGitHub 2025-03-18 09:27:38 -04:00
• cde4fd3be3 do not view_left assign + elementwise sources always have a shape [pr] (#9491) qazalandGitHub 2025-03-18 17:42:51 +08:00
• 117b7a16ef VALIDATE_WITH_CPU [pr] (#9488) George HotzandGitHub 2025-03-18 15:15:04 +08:00
• 935cd01f56 simple failing test for graph_rewrite children [pr] (#9489) qazalandGitHub 2025-03-18 13:07:21 +08:00
• d20494e6d7 move buffer logic to Buffer [pr] (#9487) George HotzandGitHub 2025-03-18 11:21:21 +08:00
• 3be228182f unbind Tensor variables last [pr] (#9486) qazalandGitHub 2025-03-18 09:52:01 +08:00
• b44f9c409a reorder do_realize [pr] (#9485) qazalandGitHub 2025-03-18 09:30:10 +08:00
• a82c9332d3 am: rename soc21 to soc (#9482) nimlgenandGitHub 2025-03-18 08:54:26 +08:00
• b100fc0b20 split the rule that uses context in scheduler simplifier [pr] (#9484) qazalandGitHub 2025-03-18 08:12:26 +08:00
• 5e58f4b65b Tiny backend test_ops fix part 3 (#9483) Anish UmaleandGitHub 2025-03-18 03:31:51 +05:30
• 9fcef4d009 add masked_select to tensor.py (#9468) 3 people 2025-03-17 13:05:36 -07:00
• 4f8eac59ea failed test case for threefry (#9469) chenyuandGitHub 2025-03-17 14:52:10 -04:00
• 6dd8e5ba7c refactor llvm compiler (#9403) 4 people 2025-03-18 00:13:49 +08:00