aboutsummaryrefslogtreecommitdiff
path: root/llvm/lib/Target/AMDGPU/SIInsertWaitcnts.cpp
AgeCommit message (Expand)AuthorFilesLines
2026-01-12[AMDGPU] Add DS loop waitcnt optimization for GFX12+ (#172728)hidekisaito1-61/+164
2026-01-12[AMDGPU] Fix crash in SIInsertWaitcnts debug output (#175518)Jay Foad1-32/+28
2026-01-12[AMDGPU] Inline two helpers in SIInsertWaitcnts. NFC. (#174557)Jay Foad1-22/+9
2026-01-12[AMDGPU][SIInsertWaitcnt] Implement Waitcnt Expansion for Profiling (#169345)Pankaj Dwivedi1-92/+203
2026-01-09[AMDGPU] Add support for GFX12 expert scheduling mode 2 (#170319)Jay Foad1-9/+287
2026-01-05[AMDGPU] Make WaitcntBrackets::simplifyWaitcnt const again (#173390)Jay Foad1-47/+61
2026-01-05[AMDGPU] Simplify and document waitcnt handling on call and return (#172453)Jay Foad1-25/+5
2026-01-05Revert "[AMDGPU] Fix excessive stack usage in SIInsertWaitcnts::run (#134835)...Jay Foad1-8/+3
2025-12-29AMDGPU: Avoid crashing on statepoint-like pseudoinstructions (#170657)Matt Arsenault1-1/+1
2025-12-17[AMDGPU][SIInsertWaitCnt] Optimize loadcnt insertion at function boundaries (...Pankaj Dwivedi1-3/+28
2025-12-16[AMDGPU] Simplify waitcnt insertion on function entry. NFC. (#172461)Jay Foad1-4/+3
2025-12-12[AMDGPU][InsertWaitCnts][NFC] Merge VMEM_ACCESS and VMEM_READ_ACCESS into a s...Pankaj Dwivedi1-8/+7
2025-12-12(reland) [AMDGPU][SIInsertWaitCnts] Use RegUnits-based tracking (#162077) (#1...Pierre van Houtryve1-282/+323
2025-12-10[AMDGPU][NFC] dump Waitcnt using an ostream operator (#171251)Sameer Sahasrabuddhe1-2/+2
2025-12-09Revert "[AMDGPU][SIInsertWaitCnts] Use RegUnits-based tracking (#162077)"pvanhout1-310/+281
2025-12-09[AMDGPU][SIInsertWaitCnts] Use RegUnits-based tracking (#162077)Pierre van Houtryve1-281/+310
2025-12-09[AMDGPU][NFC] cleanup whitespace in debug log of SIInsertWaitcntsSameer Sahasrabuddhe1-13/+13
2025-12-09[AMDGPU][NFC] fix function names in debug log for SIInsertWaitcntsSameer Sahasrabuddhe1-4/+4
2025-12-08[AMDGPU][NFC] Update a comment about FLAT v/s LDSDMASameer Sahasrabuddhe1-6/+5
2025-12-08[AMDGPU][SIInsertWaitcnts] Wait on all LDS DMA operations when no aliasing st...Pierre van Houtryve1-5/+17
2025-12-05[AMDGPU] Inherit constructors from WaitcntGenerator. NFC. (#170845)Jay Foad1-8/+3
2025-12-05[AMDGPU] Add an assertion. NFCI.Jay Foad1-0/+1
2025-12-04[AMDGPU][Waitcnts] Don't create a pending flat event for LDS DMA (#170263)Sameer Sahasrabuddhe1-4/+7
2025-12-03[AMDGPU] Take BUF instructions into account in mayAccessScratchThroughFlat (#...Pierre van Houtryve1-3/+1
2025-11-13[AMDGPU][SIInsertWaitCnts] Gfx12.5 - Refactor xcnt optimization (#164357)Ryan Mitchell1-40/+52
2025-11-12[AMDGPU] Fix missing S_WAIT_XCNT with multiple pending VMEMs (#166779)Jay Foad1-12/+6
2025-11-01[AMDGPU] Fix Xcnt handling between blocks (#165201)Aaditya1-4/+24
2025-10-13[AMDGPU][NFC] Use `getScoreUB` for XCNT insertion. (#162448)Aaditya1-1/+1
2025-10-03[AMDGPU] Account for implicit XCNT insertion (#160812)Aaditya1-0/+11
2025-10-01[AMDGPU][SIInsertWaitCnts] Remove redundant TII/TRI/MRI arguments (NFC) (#161...Pierre van Houtryve1-67/+54
2025-10-01[AMDGPU][SIInsertWaitCnts] De-duplicate code (NFC) (#161161)Pierre van Houtryve1-35/+29
2025-10-01[AMDGPU][InsertWaitCnts] Refactor some helper functions, NFC (#161160)Pierre van Houtryve1-102/+13
2025-09-26[AMDGPU] Skip debug uses in SIInsertWaitcnts::shouldFlushVmCnt (#160818)Jay Foad1-1/+1
2025-09-18[AMDGPU][SIInsertWaitcnts] Track SCC. Insert KM_CNT waits for SCC writes. (#1...Petar Avramovic1-6/+75
2025-09-17[AMDGPU][CodeGen][True16] Track waitcnt as vgpr32 instead of vgpr16 for D16 I...Brox Chen1-0/+9
2025-09-12[AMDGPU] Remove scope check in SIInsertWaitcnts::generateWaitcntInstBefore (#...choikwa1-7/+1
2025-09-03[AMDGPU] Define 1024 VGPRs on gfx1250 (#156765)Stanislav Mekhanoshin1-3/+2
2025-08-23AMDGPU/GFX12: Do not wait unnecessarily before barriers (#154970)Nicolai Hähnle1-5/+13
2025-08-22[AMDGPU][NFC] Only include CodeGenPassBuilder.h where needed. (#154769)Ivan Kosarev1-0/+4
2025-08-12[AMDGPU] Add s_barrier_init|join|leave instructions (#153296)Stanislav Mekhanoshin1-0/+1
2025-07-30[AMDGPU] introduce S_WAITCNT_LDS_DIRECT in the memory legalizer (#150887)Sameer Sahasrabuddhe1-0/+20
2025-07-28[AMDGPU][gfx1250] Use SCOPE_SE for stores that may hit scratch (#150586)Pierre van Houtryve1-28/+1
2025-07-24[AMDGPU] gfx1250 vmem prefetch target intrinsics and builtins (#150466)Stanislav Mekhanoshin1-5/+3
2025-07-21[AMDGPU] ISel & PEI for whole wave functions (#145858)Diana Picus1-0/+1
2025-07-17[AMDGPU] Move common fields out of WaitcntBrackets. NFC. (#148864)Jay Foad1-63/+59
2025-07-17[AMDGPU] Move class WaitcntBrackets after class SIInsertWaitcnts. NFC.Jay Foad1-234/+236
2025-07-02[AMDGPU] Remove HasSampler variable. NFC. (#146682)Jay Foad1-3/+8
2025-07-02[AMDGPU] Use isImage. NFC. (#146677)Jay Foad1-2/+1
2025-07-01[AMDGPU][SIInsertWaitCnts] skip meta instructions early (#145720)Sameer Sahasrabuddhe1-2/+5
2025-06-26[AMDGPU][NFC] rename some constants for readability (#145870)Sameer Sahasrabuddhe1-9/+9