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

• 7f90497efc memplan: fix large buffers planning (#11876) nimlgenandGitHub 2025-08-27 22:04:15 +03:00
• e4afdf9ea1 improve DEBUG=2 string with TB/s and TFLOPS [pr] (#11875) George HotzandGitHub 2025-08-27 11:42:41 -07:00
• e9789d8a70 Add mxfp4 support (#11873) Jordan ChalupkaandGitHub 2025-08-27 13:56:56 -04:00
• 884eb53e89 tracing: fix types (#11871) qazalandGitHub 2025-08-27 15:50:43 +03:00
• d39365809a add ctx to z3_renderer arg (#11867) Sieds LyklesandGitHub 2025-08-27 03:38:15 +02:00
• 24c00a4061 darken hex on viz (#11865) George HotzandGitHub 2025-08-26 15:57:50 -07:00
• f38e4af226 viz: add custom zoom filter (#11861) qazalandGitHub 2025-08-27 01:30:29 +03:00
• 62df6c39af amd: correct handling of relocations (#11863) nimlgenandGitHub 2025-08-27 01:26:45 +03:00
• d261458ecd add colors to range (#11860) George HotzandGitHub 2025-08-26 14:32:12 -07:00
• 7dfc7e4abc uops_to_z3 helper(#11859) Sieds LyklesandGitHub 2025-08-26 22:58:05 +02:00
• 1bbb578afd named expression for POW and MAX gradient (#11858) chenyuandGitHub 2025-08-26 16:03:03 -04:00
• 7028cb4167 clean up TestBitcastConstFolding (#11856) chenyuandGitHub 2025-08-26 15:26:47 -04:00
• d4154e0349 split devectorizing of buf/index (#11855) George HotzandGitHub 2025-08-26 12:05:48 -07:00
• b268755d51 small changes from postopt (#11854) George HotzandGitHub 2025-08-26 11:56:16 -07:00
• a3aeef45cc associative variation of where branch-merging (#11851) Sieds LyklesandGitHub 2025-08-26 19:27:05 +02:00
• aabe7756be fix type in fold_bitcast [pr] (#11853) chenyuandGitHub 2025-08-26 13:22:30 -04:00
• 4785cd959a [TYPED=1] cvar should allow dtype as a tuple (#11770) Jordan ChalupkaandGitHub 2025-08-26 12:49:51 -04:00
• b111076301 viz: fixup click on overlay rect (#11850) qazalandGitHub 2025-08-26 19:25:42 +03:00
• 1dd613cb89 test float_to_bf16 round-to-even behavior (#11849) 3 people 2025-08-27 00:16:10 +08:00
• 409399c609 fix nan in float_to_bf16 (#11843) 3 people 2025-08-26 23:42:25 +08:00
• 43d5d66d34 viz: add UOp ports to edges (#11847) qazalandGitHub 2025-08-26 18:31:52 +03:00
• f28f613f85 improved float_to_bf16 (#11848) chenyuandGitHub 2025-08-26 11:14:06 -04:00
• afe14ccbfa amd: aql default when several xccs (#11832) nimlgenandGitHub 2025-08-26 15:16:36 +03:00
• 3674c0754e viz: small uop click changes (#11846) qazalandGitHub 2025-08-26 14:56:13 +03:00
• f2a3c27372 viz: g.edges() once (#11845) qazalandGitHub 2025-08-26 13:29:59 +03:00
• b0df3e62a8 viz: light up srcs and paths on UOp click (#11844) qazalandGitHub 2025-08-26 09:03:09 +03:00
• 6236749867 viz: move rect styles to classes (#11842) qazalandGitHub 2025-08-26 07:55:34 +03:00
• 81ffa07439 viz: pass through nodes without a link (#11841) qazalandGitHub 2025-08-26 07:00:43 +03:00
• 265d287615 add decomp for !x&!y -> !(x|y) (#11836) Sieds LyklesandGitHub 2025-08-26 05:21:06 +02:00
• 337e979a59 call dtypes.as_const in Tensor(list) (#11840) chenyuandGitHub 2025-08-25 22:08:26 -04:00
• 215818379b new (post) group for reduce (#11837) George HotzandGitHub 2025-08-25 18:03:00 -07:00
• ac3449b0c8 truncate_fp16 cleanup (#11838) chenyuandGitHub 2025-08-25 19:03:41 -04:00
• e146418f65 hotfix: profiler content-type is application/octet-stream (#11831) qazalandGitHub 2025-08-25 15:56:42 +03:00
• a1f6823060 viz: memory layout in client side (#11830) qazalandGitHub 2025-08-25 14:49:33 +03:00
• a6dbb09058 changes for postrange (#11828) George HotzandGitHub 2025-08-24 17:37:07 -07:00
• 27701ef823 add locals support to rangeify (#11826) George HotzandGitHub 2025-08-24 14:03:12 -07:00
• a286a1a6f7 Fast idiv try removing factors of two before cast (#11824) Sieds LyklesandGitHub 2025-08-24 20:04:25 +02:00
• a03b930339 hotfix: green v2 in docs geohot 2025-08-24 10:25:14 -07:00
• 6540bb32a6 move into codegen late [pr] (#11823) George HotzandGitHub 2025-08-24 10:23:25 -07:00
• bba088ef11 amd aql queue (#11708) nimlgenandGitHub 2025-08-24 19:53:00 +03:00
• 1fa09d9ede BLOCK_REORDER is context var, heuristic cleanups [pr] (#11819) George HotzandGitHub 2025-08-24 09:41:34 -07:00
• 8b18cc2a94 viz memory layout cleanup (#11820) qazalandGitHub 2025-08-24 19:37:31 +03:00
• dd69114573 Revert "Better div nesting (#11811)" (#11818) Sieds LyklesandGitHub 2025-08-24 18:11:24 +02:00
• e19f901330 amd: rptr/wptr in create_queue (#11817) nimlgenandGitHub 2025-08-24 18:03:45 +03:00
• d71444857e amd: apply relocs for kernel_code_entry_byte_offset for AMD_LLVM (#11816) nimlgenandGitHub 2025-08-24 17:48:40 +03:00
• 44bc7dc73d remove KernelInfo from GROUP_REDUCE (#11814) George HotzandGitHub 2025-08-23 19:55:41 -07:00
• 229adfb7c3 Revert "remove KernelInfo from gpudims (#11809)" (#11813) George HotzandGitHub 2025-08-23 19:37:10 -07:00
• 952f729b07 Better div nesting (#11811) Sieds LyklesandGitHub 2025-08-24 04:17:40 +02:00
• e652062f92 tweak divmod_folding condition (#11810) Sieds LyklesandGitHub 2025-08-24 02:59:02 +02:00
• 846753f343 remove KernelInfo from gpudims (#11809) George HotzandGitHub 2025-08-23 16:32:45 -07:00
• 07d4ed7e4c one more symbolic add variation (#11807) Sieds LyklesandGitHub 2025-08-24 01:15:04 +02:00
• 759ebea4eb viz: reflect timeline API boundary in names (#11808) qazalandGitHub 2025-08-24 02:12:12 +03:00
• 132f09fab7 global/locals from AxisType in range (#11806) George HotzandGitHub 2025-08-23 15:49:17 -07:00
• 0d86288bd7 viz: calculate timeline fixed points in client side (#11805) qazalandGitHub 2025-08-24 01:44:40 +03:00
• a75da49951 use AxisType for UPCAST/UNROLL (#11800) George HotzandGitHub 2025-08-23 14:44:48 -07:00
• 2407fecdae viz bytepack format (#11792) qazalandGitHub 2025-08-23 23:50:21 +03:00
• b12d1d866c count bytes per kernel in test_viz (#11801) qazalandGitHub 2025-08-23 23:35:27 +03:00
• 6a50ab6b87 adjust idiv min_max (#11802) Sieds LyklesandGitHub 2025-08-23 22:25:51 +02:00
• 9d4cccd0f9 test_dtype_alu cleanups (#11799) chenyuandGitHub 2025-08-23 15:11:17 -04:00
• aefabaf774 add AxisType to range (#11798) George HotzandGitHub 2025-08-23 11:15:00 -07:00
• b975830424 add profile loader helper in test_viz (#11797) qazalandGitHub 2025-08-23 19:20:29 +03:00
• 7123df3928 Use Tensor.logaddexp to implement Tensor.softplus (#11796) chenyuandGitHub 2025-08-23 11:52:29 -04:00
• aaea6b97ad viz memory: compute nbytes (#11795) qazalandGitHub 2025-08-23 17:34:07 +03:00
• 58653b5eae viz: store memory scale (#11794) qazalandGitHub 2025-08-23 16:19:44 +03:00
• fb8ee02424 Tensor.logaddexp (#11793) chenyuandGitHub 2025-08-23 09:15:00 -04:00
• 5a6817d5f8 Fix z3 rendering of floats in indexing (#11740) Sieds LyklesandGitHub 2025-08-23 05:56:19 +02:00
• 4267c45db3 non-supported dtype in transcendental (#11754) chenyuandGitHub 2025-08-22 23:13:45 -04:00
• e39b25cd36 upcast float exp to at least float32 (#11758) chenyuandGitHub 2025-08-22 20:16:34 -04:00
• b057a90d49 memory: rename is_huge_page -> is_page (#11786) nimlgenandGitHub 2025-08-22 20:08:58 +03:00
• 38f0fa7bde viz: only send trace duration (#11789) qazalandGitHub 2025-08-22 20:00:48 +03:00
• 1c81ec9248 viz: rename to start/end timestamp (#11788) qazalandGitHub 2025-08-22 19:47:49 +03:00
• 9ff03680ba viz: store relative timestamps (#11787) qazalandGitHub 2025-08-22 19:30:21 +03:00
• 698392334f system: message for eaccess as well (#11785) nimlgenandGitHub 2025-08-22 18:21:32 +03:00
• 1e679bd789 fix max_unpool2d inf (#11784) geohotstanandGitHub 2025-08-22 20:31:24 +08:00
• 9832599c9e test_vmap + permute isn't a sint (#11783) George HotzandGitHub 2025-08-21 22:39:35 -07:00
• bb8de51e5f remove unused early cleanups + contig w range [pr] (#11780) George HotzandGitHub 2025-08-21 20:04:45 -07:00
• 91a4de4ca7 fix getitem with inf in tensor (#11781) chenyuandGitHub 2025-08-21 21:55:32 -04:00
• 66e9d54eed RANGEIFY=2 is partial contig (#11777) George HotzandGitHub 2025-08-21 16:53:58 -07:00
• 8de6db15ac exclude .git from ruff (#11773) Jordan ChalupkaandGitHub 2025-08-21 18:37:50 -04:00
• 5954a0975f fix some assigns on rangeify (#11774) George HotzandGitHub 2025-08-21 15:15:54 -07:00
• 2e0eb88549 viz: add metadata to UOp tracing (#11772) qazalandGitHub 2025-08-22 00:18:45 +03:00
• d6f9606e93 small cleanups to rangeify (#11769) George HotzandGitHub 2025-08-21 11:15:09 -07:00
• bd4a9473b0 Multihost exception handling (#11729) 3 people 2025-08-21 17:51:49 +00:00
• a2c7b807e0 don't bufferize 0s (#11766) George HotzandGitHub 2025-08-21 10:10:56 -07:00
• 9eff7cd1d8 am: support 64bit discovery (#11768) nimlgenandGitHub 2025-08-21 18:28:13 +03:00
• 56cd47a159 fix amd llvm bf16 tc (#11713) 4 people 2025-08-21 21:33:28 +08:00
• a044648111 rangeify load cleanups + multi support (#11765) George HotzandGitHub 2025-08-20 20:55:49 -07:00
• 9f94c25a25 fix symbolic usage. use shrink, not reshape (#11762) George HotzandGitHub 2025-08-20 18:35:42 -07:00
• 5276fbc9c5 fix gather with inf values (#11760) chenyuandGitHub 2025-08-20 20:35:40 -04:00
• b979162c5d llama3 eval train (#11706) wozeparrotandGitHub 2025-08-20 19:56:35 -04:00
• dbd3b67657 clamp GRAD_CLIP_NORM in llama (#11761) chenyuandGitHub 2025-08-20 19:55:50 -04:00
• 9635592141 ** rangeify, try 3 (#11683) George HotzandGitHub 2025-08-20 14:22:44 -07:00
• d7553721d1 clean up test_dtype_alu (#11757) chenyuandGitHub 2025-08-20 14:36:18 -04:00
• 5f08a3e928 hotfix: cast half to float in Tensor.tolist (#11755) chenyuandGitHub 2025-08-20 12:18:35 -04:00
• de4cb722a4 viz: add metadata and var_vals tracing (#11753) qazalandGitHub 2025-08-20 18:39:51 +03:00
• 6589c9e643 hcq: better errors for ifaces (#11751) nimlgenandGitHub 2025-08-20 17:50:51 +03:00
• be7b0b6970 TRANSCENDENTAL_SUPPORTED_DTYPES->TRANSCENDENTAL_DTYPES (#11752) chenyuandGitHub 2025-08-20 10:29:36 -04:00
• 220a2a88d7 a*(1/b) -> a/b on LLVM, CPU (#11743) ttomsaandGitHub 2025-08-20 14:35:10 +01:00
• 12ab3f8b06 correct row_count in process replay (#11748) George HotzandGitHub 2025-08-19 22:21:07 -07:00
• 8af8808c61 cleanup tests, bump caches (#11746) George HotzandGitHub 2025-08-19 21:21:07 -07:00