Created
September 6, 2025 05:45
-
-
Save davidberard98/34e1943dd9c057b1a9e186278ae8fc55 to your computer and use it in GitHub Desktop.
This file contains hidden or bidirectional Unicode text that may be interpreted or compiled differently than what appears below. To review, open the file in an editor that reveals hidden Unicode characters.
Learn more about bidirectional Unicode characters
| # only skipped commits left to test | |
| # possible first bad commit: [f0975f9d02f6d8b69146aea84b8ee7c6e81c4ed5] [AMD][Build] Fix build issue with AMD lld (#7608) | |
| # possible first bad commit: [620c59165a1452ddd5dd685054b84485a35dc92e] [AMD][NFC] Group scheduling functions in StreamPipeliner (#7607) | |
| # possible first bad commit: [16b25e1620e87f02f7d16ff5b9da2d425c6e99a6] [BACKEND] combineRedundantWaitOps should not combine across loops/branches (#7593) | |
| # possible first bad commit: [fdd694d48a2b05c49bd9658fbedc3204d04654e1] [Triton][Gluon] Add `map_elementwise` (#7564) | |
| # possible first bad commit: [03cdcdb38a4812f8c371ff7de433e6c7b605b1ef] [KERNELS] Fix `bench_mlp.py` for AMD (#7600) | |
| # possible first bad commit: [ef72c317570de66807d83648bde1bff64e2898f8] [Tests] Improve regex for test_compile_only_dot (#7602) | |
| # possible first bad commit: [a7a89c7c9262ed3d761ea771a907d53f4caba92a] [FRONTEND] Refactor unsplat to use new op (#7586) | |
| # possible first bad commit: [6415039bf96145fdabca0ca0ac5f05b9d42cf45f] [AMD] Support 4x64 and 64x4 MFMA layout for dot (#7576) | |
| # possible first bad commit: [8a5862dceae0d4a12b38158b4abd5fcbb7209fa2] [AMD] Enable lowerLocalLdSt for AMD path (#7355) | |
| # possible first bad commit: [75eed88ab167eaf1d4df7897cae0ce559f2e6951] [KERNELS] fixup leftover experiment code (#7599) | |
| # possible first bad commit: [d0642aab963a6e435b11eaa6dd184dc9208fb1df] Revert "[BACKEND] Hoist tmem alloc outside of if (#7568)" (#7597) | |
| # possible first bad commit: [e6aa86c9e9e232aae80548310dbf7a6a112be400] [BACKEND] Fix fp16 to fp32 conversion (#7585) | |
| # possible first bad commit: [ca0fe1b41dc4adfcfcefa23858e9a040d1a4a516] [KERNELS] Skip test_upcast_mxfp4_to_bf16 on AMD (#7595) | |
| # possible first bad commit: [de846c0b0858d06e1fc7783b1ea13ecb6e2611fa] [KERNELS] added missing mxfp4 tests (#7591) | |
| # possible first bad commit: [fde96e8f1703ab9a6410c5c6b0ff3a6be64b3a55] Remove redundant calls to inspect.getsource (#7588) | |
| # possible first bad commit: [981b0bb0c66b28d76a71141b837ecff594207753] [KERNELS] improved twiddling/swizzling for H100 simulated mxfp4 (h/t @lezcano) (#7587) | |
| # possible first bad commit: [63f34320e389d95128ab9cff5f356d2096dfba7f] [WS] Update RewritePartitionDepdencies to insert arefs (#7561) | |
| # possible first bad commit: [6a6ed52e1781d38fad487da39443604273761bb9] [BACKEND] Generalize `getShapePerCTA` (#7580) | |
| # possible first bad commit: [167ed2866c56c0b5e168d7f02dddec51c8ff4c5b] Cache NvidiaTool.from_path (#7569) | |
| # possible first bad commit: [d5de496dcd0fbefa0b785efd708d43b577fbf3d5] [KERNELS] Remove incorrect check (#7583) | |
| # possible first bad commit: [7708b9c12a9848a0be33e1c21c7c7bdd23b207d4] [mxfp/easy] handle an empty tensor in downcast_to_mxfp (#7579) | |
| # possible first bad commit: [9660a0d479774e40f451c682b6bb9cf32330dfcc] [CI] Reduce MacOS CI build time (#7559) | |
| # possible first bad commit: [13b3a75de6f418826254e3341cd2095a3347ee59] [KERNELS] Fix swizzling contiguity (#7582) | |
| # possible first bad commit: [984b694dc2916ee4f8cd18d3a28d1d8da14e076d] [mxfp/easy] add MXFP_BLOCK_SIZE constant (#7567) | |
| # possible first bad commit: [67f647af21b6ec85bbe561a6591ee5c28cffe509] [IR] Add type checks for `atomic_cas` (#7578) | |
| # possible first bad commit: [948ba8f4fd171588b584f51109f04692b77d5c27] [DOC] Update install instructions in docs (#7572) | |
| # possible first bad commit: [db7170e18a3529bdaf44dcdeecb56eb36d375c54] [Backend] Follow-up refactor of `getWarpLayoutConvertDecomposition` (#7571) | |
| # possible first bad commit: [1e0a371139142b04c60b553a1324907c00ea3fbc] [AMD] Restrict merging async_wait in StreamPipeliner (#7577) | |
| # possible first bad commit: [4048f318a4bb4c7da693e56cd8c9a4d5a39e4e39] [BACKEND] Add LocalLoadTrait to group local load-like ops (#7511) | |
| # possible first bad commit: [3854ae83b65a4234c0c4b6b39054b5ea3e3a4b73] [Backend] Improve warp-local layout conversion algo using shuffles (#7558) | |
| # possible first bad commit: [570f24d016702cbfe1179beae3eb03d24e9c6b40] [AMD] Improve register usage in Float8 conversions (#7527) | |
| # possible first bad commit: [96e91d49f1fa0760594ea65699920dcfa04ed206] [BACKEND] Hoist tmem alloc outside of if (#7568) | |
| # possible first bad commit: [7affb3b232166fcac3af9f8295df7bede6b6b35d] [BACKEND] Ignore bad bytes in ldconfig -p (#7566) | |
| # possible first bad commit: [dba750e4d6786f077ec6f064153482ebaab22019] [AMD] Support async load in ping-pong pass (#7458) | |
| # possible first bad commit: [ea82c9f485fa5283fee4abc0e62df69f2380ce15] [Gluon] Clarify o_init meaning in attention tutorial (#7563) | |
| # possible first bad commit: [3bd3e320d901aa7de663795e7723095334b3c463] [AMD] Use lld library API for linking (#7548) |
Sign up for free
to join this conversation on GitHub.
Already have an account?
Sign in to comment