aboutsummaryrefslogtreecommitdiff
path: root/llvm/test/Transforms/AtomicExpand/AMDGPU
AgeCommit message (Expand)AuthorFilesLines
14 days[AMDGPU] Fix duplicate RUN lines in tests (#182286)Jay Foad1-1/+0
2026-02-05IR: Promote "denormal-fp-math" to a first class attribute (#174293)Matt Arsenault2-6/+6
2026-01-18AtomicExpand: Use LibcallLoweringInfo analysis (#176384)Matt Arsenault32-144/+144
2025-12-05AMDGPU: Add codegen for atomicrmw operations usub_cond and usub_sat (#141068)anjenner8-379/+4056
2025-09-01[AMDGPU] Expand scratch atomics to flat atomics if GAS is enabled (#154710)Pierre van Houtryve1-0/+235
2025-08-22AMDGPU: Start considering new atomicrmw metadata on integer operations (#122138)Matt Arsenault8-951/+2956
2025-08-22AMDGPU: Expand remaining system atomic operations (#122137)Matt Arsenault5-62/+614
2025-08-19[AMDGPU] Check noalias.addrspace in mayAccessScratchThroughFlat (#151319)Pierre van Houtryve1-20/+4
2025-07-09AtomicExpand: Stop using report_fatal_error (#147300)Matt Arsenault5-1115/+7
2025-07-08AMDGPU: Fix atomic expand tests accidentally underaligning (#147299)Matt Arsenault2-837/+2900
2025-06-05[AMDGPU][AtomicExpand] Use full flat emulation if a target supports f64 globa...Shilei Tian2-103/+628
2025-05-09[AtomicExpand] Drop explicit datalayout from testAlexander Richardson1-8/+1
2025-04-30[AMDGPU] Remove explicit datalayout from tests where not neededAlexander Richardson1-2/+0
2025-02-13[AMDGPU][NFC] Replace gfx940 and gfx941 with gfx942 in llvm/test (#125711)Fabian Ritter18-1236/+1236
2024-11-04AMDGPU: Custom expand flat cmpxchg which may access private (#109410)Matt Arsenault3-17/+170
2024-10-31AtomicExpand: Copy metadata from atomicrmw to cmpxchg (#109409)Matt Arsenault18-935/+979
2024-10-31AMDGPU: Add baseline tests for cmpxchg custom expansion (#109408)Matt Arsenault4-8/+365
2024-10-31AMDGPU: Expand flat atomics that may access private memory (#109407)Matt Arsenault6-221/+949
2024-10-15AMDGPU: Add baseline tests for flat-may-alias private atomic expansions (#109...Matt Arsenault1-0/+1523
2024-08-14AMDGPU: Preserve alignment when custom expanding atomicrmw (#103768)Matt Arsenault1-0/+55
2024-08-13AMDGPU: Stop handling legacy amdgpu-unsafe-fp-atomics attribute (#101699)Matt Arsenault2-93/+115
2024-08-13AMDGPU: Do not create phi user for atomicrmw with no uses (#103061)Matt Arsenault2-3/+0
2024-08-09AMDGPU: Preserve atomicrmw name when specializing address space (#102470)Matt Arsenault3-8/+18
2024-08-09AMDGPU: Avoid creating unnecessary block split in atomic expansion (#102440)Matt Arsenault3-22/+10
2024-08-02AMDGPU: Handle new atomicrmw metadata for fadd case (#96760)Matt Arsenault12-4739/+964
2024-08-01AMDGPU: Handle remote/fine-grained memory in atomicrmw fmin/fmax lowering (#9...Matt Arsenault4-672/+4590
2024-06-27AMDGPU: Fix no return atomicrmw fadd v2f16 selection for gfx908 (#96948)Matt Arsenault1-13/+1
2024-06-23AMDGPU: Start selecting flat/global atomicrmw fmin/fmax. (#95592)Matt Arsenault6-100/+106
2024-06-18AMDGPU: Support local atomicrmw fmin/fmax for float/double (#95590)Matt Arsenault2-56/+8
2024-06-18AMDGPU: Add some tests for i128 and fp128 atomic expansionMatt Arsenault2-0/+323
2024-06-17Reapply "AMDGPU: Handle legal v2f16/v2bf16 atomicrmw fadd for global/flat (#9...Matt Arsenault6-1083/+6415
2024-06-15Revert "AMDGPU: Handle legal v2f16/v2bf16 atomicrmw fadd for global/flat (#95...Nico Weber6-6415/+1083
2024-06-15AMDGPU: Handle legal v2f16/v2bf16 atomicrmw fadd for global/flat (#95394)Matt Arsenault6-1083/+6415
2024-06-15AMDGPU: Legalize atomicrmw fadd for v2f16/v2bf16 for local memory (#95393)Matt Arsenault1-60/+332
2024-06-12AtomicExpand: Fix creating invalid ptrmask for fat pointers (#94955)Matt Arsenault2-0/+208
2024-06-11AMDGPU: Add more tests for vector typed atomicrmw faddMatt Arsenault1-9/+178
2024-05-23AtomicExpand: Preserve metadata when expanding partword RMW (#89769)Matt Arsenault1-10/+102
2024-05-07AMDGPU: Don't bitcast float typed atomic store in IR (#90116)Matt Arsenault1-32/+19
2024-05-07AMDGPU: Do not bitcast atomicrmw in IR (#90045)Matt Arsenault6-88/+112
2024-05-03AMDGPU: Add more tests for atomicrmw handlingMatt Arsenault8-21/+8779
2024-05-03AMDGPU: Update name for amdgpu.no.remote.memory metadataMatt Arsenault7-634/+634
2024-04-26AMDGPU: Do not bitcast atomic load in IR (#90060)Matt Arsenault1-34/+20
2024-04-25AMDGPU: Add baseline tests for bad bitcasting of atomic load/storeMatt Arsenault2-0/+374
2024-04-24AtomicExpand: Fix dropping a syncscope when bitcasting atomicrmwMatt Arsenault1-6/+6
2024-04-24[IR] Memory Model Relaxation Annotations (#78569)Pierre van Houtryve1-0/+204
2024-04-23AtomicExpand: Emit or with constant on RHSMatt Arsenault4-5/+5
2024-04-22AMDGPU: Simplify DS atomicrmw fadd handling (#89468)Matt Arsenault1-60/+284
2024-04-20AMDGPU: Add tests for atomicrmw handling of new metadata (#89248)Matt Arsenault8-9/+10986
2024-04-19AMDGPU: Use common check prefix in atomic expand testMatt Arsenault1-2403/+393
2024-04-17AMDGPU: Fix not handling atomicrmw fadd in exotic address spaces correctlyMatt Arsenault1-0/+414
generated by cgit v1.1 at 2026-03-05 05:34:01 +0000