Commits · d3ffd47df94b362ebf52c0681bc2cdcfef00b98f · Roger Ferrer / llvm-epi · GitLab

Mar 29, 2019
- AMDGPU: Remove dx10-clamp from subtarget features · 055e4dce
  Matt Arsenault authored Mar 29, 2019
  055e4dce
- [AMDGPU][MC] Corrected conversion rules for inlinable constants to match rules for literals · d6827ce3
  Dmitry Preobrazhensky authored Mar 29, 2019
  d6827ce3
- [AMDGPU][MC] Corrected handling of tied src for atomic return MUBUF opcodes · 7f33574b
  Dmitry Preobrazhensky authored Mar 29, 2019
  7f33574b
- AMDGPU: Make sram-ecc off by default for Vega20 · 2b766ed7
  Konstantin Zhuravlyov authored Mar 29, 2019
  2b766ed7
- [ScheduleDAG] Move `Topo` and `addEdge` to base class. · b70355f0
  Clement Courbet authored Mar 29, 2019
  b70355f0
- AMDGPU/GlobalISel: Insert waterfall loop for vector indexing · 5fddf091
  Matt Arsenault authored Mar 29, 2019
  5fddf091
Mar 28, 2019
- Delay initialization of three static global maps, NFC · 85e2cdac
  Reid Kleckner authored Mar 28, 2019
  85e2cdac
- AMDGPU: Make exec mask optimzations more resistant to block splits · a353fd57
  Matt Arsenault authored Mar 28, 2019
  a353fd57
Mar 27, 2019
- [LegalizeVectorTypes] Allow single loads and stores for more short vectors · b1650f0d
  Justin Bogner authored Mar 27, 2019
  b1650f0d
- Reapply "AMDGPU: Scavenge register instead of findUnusedReg" · 7b14b242
  Matt Arsenault authored Mar 27, 2019
  7b14b242
- AMDGPU: Enable the scavenger for large frames · 17e39100
  Matt Arsenault authored Mar 27, 2019
  17e39100
- AMDGPU: Add additional MIR tests for exec mask optimizations · 4d47ac3b
  Matt Arsenault authored Mar 27, 2019
  4d47ac3b
- AMDGPU: Skip debug_instr when collapsing end_cf · 4ab28b64
  Matt Arsenault authored Mar 27, 2019
  4ab28b64
- AMDGPU: Fix missing scc implicit def on s_andn2_b64_term · a42b7247
  Matt Arsenault authored Mar 27, 2019
  a42b7247
- AMDGPU: Don't hardcode num defs for MUBUF instructions · 28f97f1d
  Matt Arsenault authored Mar 27, 2019
  28f97f1d
- AMDGPU: wave_barrier is not isBarrier · e9ad7e9a
  Matt Arsenault authored Mar 27, 2019
  e9ad7e9a
- AMDGPU: Fix areLoadsFromSameBasePtr for DS atomics · bbc59d8d
  Matt Arsenault authored Mar 27, 2019
  bbc59d8d
- Revert of 357063 [AMDGPU][MC] Corrected handling of tied src for atomic return MUBUF opcodes · 40f0162a
  Dmitry Preobrazhensky authored Mar 27, 2019
  40f0162a
- [AMDGPU][MC] Corrected handling of tied src for atomic return MUBUF opcodes · bcc4d538
  Dmitry Preobrazhensky authored Mar 27, 2019
  bcc4d538
Mar 25, 2019
- Revert "AMDGPU: Scavenge register instead of findUnusedReg" · 8bbc1597
  Matt Arsenault authored Mar 25, 2019
  8bbc1597
- AMDGPU: Remove unnecessary check for isFullCopy · 77bf2e37
  Matt Arsenault authored Mar 25, 2019
  77bf2e37
- AMDGPU: Set hasSideEffects 0 on _term instructions · bc978872
  Matt Arsenault authored Mar 25, 2019
  bc978872
- AMDGPU: Add support for cross address space synchronization scopes · 51809cbc
  Konstantin Zhuravlyov authored Mar 25, 2019
  51809cbc
- AMDGPU: Preserve LiveIntervals in WQM · fa284551
  Matt Arsenault authored Mar 25, 2019
  fa284551
Mar 22, 2019
- [AliasAnalysis] Second prototype to cache BasicAA / anyAA state. · bfc779e4
  Alina Sbirlea authored Mar 22, 2019
  bfc779e4
- [AMDGPU] Use three- and five-dword result type in image ops · 6f0191a5
  Tim Renouf authored Mar 22, 2019
  6f0191a5
- [AMDGPU] Implemented dwordx3 variants of buffer/tbuffer load/store intrinsics · 677387d8
  Tim Renouf authored Mar 22, 2019
  677387d8
- [AMDGPU] Added v5i32 and v5f32 register classes · 033f99a2
  Tim Renouf authored Mar 22, 2019
  033f99a2
Mar 21, 2019
- [AMDGPU] Support for v3i32/v3f32 · 361b5b21
  Tim Renouf authored Mar 21, 2019
  361b5b21
Mar 20, 2019
- [AMDGPU] Do not generate spurious PAL metadata · 2327c231
  Tim Renouf authored Mar 20, 2019
  2327c231
- [AMDGPU] Fix dependency on `BinaryFormat` · bbcb95a6
  Michael Liao authored Mar 20, 2019
  bbcb95a6
- AMDGPU: Don't look for constant in insert/extract_vector_elt regbankselect · 2065206a
  Matt Arsenault authored Mar 20, 2019
  2065206a
- [AMDGPU] Fix clamp bit DAG operand · eea5177d
  Michael Liao authored Mar 20, 2019
  eea5177d
- AMDHSA: Fix COMPUTE_PGM_RSRC2.USER_SGPR calculation when parsing ISA assembly · 88268e3e
  Konstantin Zhuravlyov authored Mar 20, 2019
  88268e3e
- [AMDGPU] Added MsgPack format PAL metadata · e7bd52f8
  Tim Renouf authored Mar 20, 2019
  e7bd52f8
- [AMDGPU] Factored PAL metadata handling out into its own class · d737b551
  Tim Renouf authored Mar 20, 2019
  d737b551
- [AMDGPU][MC] Corrected checks for DS offset0 range · 04bd1185
  Dmitry Preobrazhensky authored Mar 20, 2019
  04bd1185
- [AMDGPU][MC][GFX9] Added support of operands shared_base, shared_limit,... · 137976fa
  Dmitry Preobrazhensky authored Mar 20, 2019
  137976fa
- [AMDGPU] Allow MIMG with no uses in adjustWritemask in isel · fc2a7473
  David Stuttard authored Mar 20, 2019
  fc2a7473
Mar 19, 2019
- CodeGen: Refactor regallocator command line and target selection · cf55a657
  Matt Arsenault authored Mar 19, 2019
  cf55a657