Commits · d37ce53ad30f3d5e7fb37b5bb9c49793ca33d2df · Roger Ferrer / llvm-epi · GitLab

May 27, 2020
- AMDGPU: Set StackPointerRegisterToSaveRestore · d37ce53a
  Matt Arsenault authored May 27, 2020
  d37ce53a
- AMDGPU: Fix dropping MI flags when rewriting instructions · 07cd19ef
  Matt Arsenault authored May 26, 2020
  07cd19ef
- AMDGPU: Fix backwards s_cselect_* operands · 833996ce
  Matt Arsenault authored May 19, 2020
  833996ce
- GlobalISel: Basic legalization for G_PTRMASK · ef3e8312
  Matt Arsenault authored May 23, 2020
  
  ef3e8312
May 26, 2020
- [AMDGPU] Bail alloca vectorization if GEP not found · 512e806a
  Stanislav Mekhanoshin authored May 26, 2020
  512e806a
- AMDGPU: Fix wrong null value for private address space · bb10fa3a
  Matt Arsenault authored May 24, 2020
  bb10fa3a
- AMDGPU: Update store node checks for atomics · e09064e9
  Matt Arsenault authored May 24, 2020
  e09064e9
- Revert "[AMDGPU] NFC target dependent requiresUniformRegister refactored out" · 9786e755
  Matt Arsenault authored May 26, 2020
  9786e755
- [AMDGPU] NFC target dependent requiresUniformRegister refactored out · fb38b983
  alex-t authored May 26, 2020
  fb38b983
- AMDGPU/GlobalISel: Fix assert on 16-bit G_EXTRACT results · 50d4b22c
  Matt Arsenault authored May 20, 2020
  50d4b22c
- GlobalISel: Merge G_PTR_MASK with llvm.ptrmask intrinsic · 8bc03d21
  Matt Arsenault authored May 15, 2020
  8bc03d21
- AMDGPU/GlobalISel: Don't select boolean phi by default · 2dd7714b
  Matt Arsenault authored May 19, 2020
  2dd7714b
- [AMDGPU] Fix wait counts in the presence of 16bit subregisters · 92f3828d
  vpykhtin authored May 26, 2020
  92f3828d
- [CostModel] Unify Intrinsic Costs. · 871556a4
  Sam Parker authored May 26, 2020
  871556a4
May 25, 2020
- [AMDGPU][MC][GFX8+] Enabled clamp for v_add_u16, v_sub_u16 and v_subrev_u16 · 77aec3b4
  Dmitry Preobrazhensky authored May 25, 2020
  77aec3b4
- [AMDGPU][CODEGEN] Added 'A' constraint for inline assembler · b087b91c
  Dmitry Preobrazhensky authored May 25, 2020
  b087b91c
May 24, 2020
- AMDGPU.h - reduce TargetMachine.h include. NFC. · 71bed820
  Simon Pilgrim authored May 24, 2020
  71bed820
- AMDGPUInstPrinter.cpp - add CommandLine.h include. NFC. · b05b69e0
  Simon Pilgrim authored May 24, 2020
  b05b69e0
- AMDGPUTargetObjectFile.h - remove unnecessary includes. NFC. · 725b3463
  Simon Pilgrim authored May 24, 2020
  725b3463
- AMDGPULibFunc - fix include order. NFC. · a6502560
  Simon Pilgrim authored May 24, 2020
  a6502560
May 23, 2020
- AMDGPU: Implement isConstantPhysReg · 76e3dd0a
  Matt Arsenault authored May 20, 2020
  76e3dd0a
- AMDGPU: Define mode register · 2e82667f
  Matt Arsenault authored Jan 03, 2020
  2e82667f
May 22, 2020
- [AMDGPU] Define 6 dword subregs · 62fb3fa6
  Stanislav Mekhanoshin authored May 22, 2020
  62fb3fa6
- AMDGPU/GlobalISel: Fix masked control flow with fallthrough blocks · 66fe6022
  Matt Arsenault authored May 17, 2020
  66fe6022
- [AMDGPU][MC][GFX8+] Enabled clamp for v_mul_i32_i24_e64 and v_mul_u32_u24_e64 · 933ebc40
  Dmitry Preobrazhensky authored May 22, 2020
  933ebc40
May 21, 2020
- [AMDGPU] Fixed incorrect PAL metadata register naming · d13a5088
  Tim Renouf authored Apr 25, 2020
  d13a5088
- [AMDGPU] Promote alloca to vector in opt · 689e616e
  Stanislav Mekhanoshin authored May 20, 2020
  689e616e
- [AMDGPU] Tune threshold for cmp/select vector lowering · 1dfd1b3e
  Stanislav Mekhanoshin authored May 20, 2020
  1dfd1b3e
- Revert "[CostModel] Unify Intrinsic Costs." · 259eb619
  Sam Parker authored May 21, 2020
  259eb619
- [CostModel] Unify Intrinsic Costs. · de71def3
  Sam Parker authored May 18, 2020
  de71def3
- [AMDGPU] Always expand ext/insertelement with divergent idx · 4eecf171
  Stanislav Mekhanoshin authored May 15, 2020
  4eecf171
May 20, 2020
- AMDGPU/GlobalISel: Fix splitting 64-bit extensions · e8f6b0e5
  Matt Arsenault authored May 18, 2020
  e8f6b0e5
- [NFCI][CostModel] Refactor getIntrinsicInstrCost · 8cc911fa
  Sam Parker authored May 20, 2020
  8cc911fa
- [AMDGPU] Process V_MOV_B32_indirect in SET_GPR_IDX optimization · 677929e3
  Stanislav Mekhanoshin authored May 19, 2020
  677929e3
- [DAGCombine] Remove the getNegatibleCost to avoid the out of sync with getNegatedExpression · 2b59e9f1
  QingShan Zhang authored May 20, 2020
  2b59e9f1
- AMDGPU: Annotate functions that have stack objects · 21d2884a
  Matt Arsenault authored May 19, 2020
  21d2884a
- AMDGPU: Fix DAG divergence for implicit function arguments · 074b8026
  Matt Arsenault authored May 13, 2020
  074b8026
- AMDGPU: Use member initializers in MFI · 61813b80
  Matt Arsenault authored May 19, 2020
  
  61813b80
May 19, 2020
- CodeGen: Use Register · 4dad4914
  Matt Arsenault authored May 19, 2020
  
  4dad4914
May 18, 2020
- [AMDGPU] Fixed selection error for 64 bit extract_subvector · 50f3bb13
  Stanislav Mekhanoshin authored May 18, 2020
  50f3bb13