main

Ccandyhongfix: sync with upstream commit 523a1b2

03c21c85创建于 1月29日历史提交

文件	最后提交记录	最后更新时间
accelerate-amd-matmul-chain-dot.mlir	[AMD] NFC: Drop version minor for AMD MFMA layout (#7285)	11 个月前
accelerate-amd-matmul-fma.mlir	[AMD][NFC] use double quote instead of single quote (#6311)	1 年前
accelerate-amd-matmul-mfma-gfx950.mlir	[AMD] Improve scaled dot with (b)f16 types on GFX950 (#7693)	10 个月前
accelerate-amd-matmul-mfma.mlir	[AMD] Support 4x64 and 64x4 MFMA layout for dot (#7576)	10 个月前
accelerate-amd-matmul-wmma-gen1.mlir	[AMD] Disable dot wmma configurations that use f16 as accumulator (#6460)	1 年前
accelerate-amd-matmul-wmma-gen2.mlir	[AMD][gfx12] Enable dot for f8 operands (#6814)	1 年前
amd-block-pingpong-chained-dots.mlir	[IR] Improve memdesc_index printing to make it obvious what operand is the index (#7876)	9 个月前
amd-block-pingpong.mlir	[IR] Improve memdesc_index printing to make it obvious what operand is the index (#7876)	9 个月前
amd-canonicalize-extract-slice.mlir	[AMD] Added a canonicalizer to ConcatOp (#7273)	11 个月前
amd-canonicalize-pointers-dont-run-mlir-canonicalizer.mlir	[AMD] Fix multi-yield scf.if during pointer canonicalization (#6276)	1 年前
amd-canonicalize-pointers.mlir	[Dialect] Actually enable TMEM layout check and fix all the tests (#7723)	10 个月前
amd-coalesce-async-copy.mlir	[AMD] Reject async global to local ops for less than 4 bytes (#6734)	1 年前
amd-concat-op.mlir	[AMD] Support register broadcast in slice/concat ops (#7407)	10 个月前
amd-conditional-barrier.mlir	[AMD][NFC] use double quote instead of single quote (#6311)	1 年前
amd-convert-buffer-ops-range-analysis.mlir	[Dialect] Actually enable TMEM layout check and fix all the tests (#7723)	10 个月前
amd-convert-buffer-ops.mlir	[AMD] Fix buffer op mask operand removal (#7963)	9 个月前
amd-extractslice-op.mlir	[AMD] Support register broadcast in slice/concat ops (#7407)	10 个月前
amd-fold-true-cmpi.mlir	[IR] Improve memdesc_index printing to make it obvious what operand is the index (#7876)	9 个月前
amd-hoist-cvtToDotOp.mlir	[AMD] NFC: Drop version minor for AMD MFMA layout (#7285)	11 个月前
amd-optimize-dot-operands.mlir	[AMD] Use single LDS for both transposed and non-transposed access (#7813)	9 个月前
amd-optimize-epilogue.mlir	[Dialect] Actually enable TMEM layout check and fix all the tests (#7723)	10 个月前
amd-pipeline-chained-dots.mlir	[AMD] Add ChainedDotSchedule to StreamPipeliner (#7601)	10 个月前
amd-range-analysis.mlir	[IR] Improve memdesc_index printing to make it obvious what operand is the index (#7876)	9 个月前
amd-reorder-instructions.mlir	[IR] Improve memdesc_index printing to make it obvious what operand is the index (#7876)	9 个月前
amd-schedule-hint.mlir	[AMD] NFC: Drop version minor for AMD MFMA layout (#7285)	11 个月前
amd-stream-loop-assume.mlir	[LAYOUTS] Enable generic swizzling on AMD (#7225)	11 个月前
amd-stream-prefetch.mlir	[LAYOUTS] Enable generic swizzling on AMD (#7225)	11 个月前
amd-update-async-wait-count.mlir	[Gluon] Actually enable scf+cf+arith canonicalizers (#7775)	9 个月前
in-thread-transpose.mlir	[IR] Improve memdesc_index printing to make it obvious what operand is the index (#7876)	9 个月前
invalid.mlir	[LAYOUTS] Implement toLinearLayout for TensorMemoryEncodingAttr (#7748)	10 个月前
mfma-double-rate.mlir	[AMD] Support async load in ping-pong pass (#7458)	10 个月前
mfma-xf32.mlir	[AMD] NFC: Drop version minor for AMD MFMA layout (#7285)	11 个月前
optimize-lds-usage.mlir	[AMD] Prefer swizzle layout in optimize-lds-usage pass (#7750)	9 个月前
sink-setprio-mfma.mlir	[AMD] NFC: Drop version minor for AMD MFMA layout (#7285)	11 个月前