[AMD] release/3.2.x AMD perf cherry picks#5191

Merged

antiagainst merged 15 commits intotriton-lang:rc/3.2.xfrom

jataylo:32_amd_cherrypicks_pr

Dec 4, 2024

Commits on Nov 18, 2024

[AMD] Emit vectorized 16-bit float LLVM atomic ops (#4925)

joviliast
authored and
jataylo
committed
[AMD] Restructure ReorderInstructions pass (#4998)

antiagainst
authored and
jataylo
committed
[AMD] Reland sinking the 2nd tt.load after local_load's (#4935)

zhanglx13
authored and
jataylo
committed
[BACKEND] Improve detection of register to register conversion (#4991)

Jokeren
authored and
jataylo
committed
[BACKEND] Replace `isMmaToDotShortcut` with linear layout based logic (#4951)

Jokeren
authored and
jataylo
committed
[AMD] remove redundant LDS bypass checks (#5002)

binarman
authored and
jataylo
committed
[AMD] Support warp-level reduction with DPP (#5019)

knwng
authored and
jataylo
committed
[AMD] Add missing dependency to TritonAMDGPUIR (#5053)

antiagainst
authored and
jataylo
committed
[AMD] Use DPP to accelerate 16-bit floats (#5072)

joviliast
authored and
jataylo
committed

Commits on Dec 3, 2024

Add gfx950 target definitions.

jungpark-mlir
authored and
jataylo
committed

Commits on Dec 4, 2024

[BACKEND] Update LLVM hash (#5040)

peterbell10
authored and
jataylo
committed
[BACKEND] Update llvm to llvm/llvm-project@fa57c7a6a5f594a9e3ae2dbe35… (#5064)

htyu
authored and
jataylo
committed
Update to llvm/llvm-project@bd9145c8c213 (#5180)

antiagainst
authored and
jataylo
committed
[LLVM] Update to llvm-project@86b69c3 (#5242)

peterbell10
authored and
jataylo
committed
[LLVM] Update to llvm/llvm-project@1f20eee6dc36 (#5308)

authored and
jataylo
committed