Commits · 8b175672cbbb19d328038be74fd917e405b46168 · Roger Ferrer / llvm-epi-0.8 · GitLab

This project is archived. Its data is read-only.

Feb 02, 2016
- AMDGPU: Skip promote alloca with no optimizations · 8b175672
  Matt Arsenault authored Feb 02, 2016
  8b175672
- AMDGPU: Whitelist handled intrinsics · ad134845
  Matt Arsenault authored Feb 02, 2016
  ad134845
- AMDGPU: Use inbounds when calculating workitem offset · 853a1fc6
  Matt Arsenault authored Feb 02, 2016
  853a1fc6
- Refactor backend diagnostics for unsupported features · 7e7d983a
  Oliver Stannard authored Feb 02, 2016
  7e7d983a
Jan 30, 2016
- AMDGPU: Fix emitting invalid workitem intrinsics for HSA · e0132464
  Matt Arsenault authored Jan 30, 2016
  e0132464
- AMDGPU: Stop checking intrinsics not used by HSA for dispatch-ptr · d0799df7
  Matt Arsenault authored Jan 30, 2016
  d0799df7
- AMDGPU: Add new amdgcn workitem intrinsics · 43976df0
  Matt Arsenault authored Jan 30, 2016
  43976df0
Jan 29, 2016
- AMDGPU: Remove 24-bit intrinsics · 295875ef
  Matt Arsenault authored Jan 29, 2016
  295875ef
Jan 28, 2016
- AMDGPU: Match fmed3 patterns with legacy fmin/fmax · 5b39b34c
  Matt Arsenault authored Jan 28, 2016
  5b39b34c
- AMDGPU: Match some med3 patterns · f639c327
  Matt Arsenault authored Jan 28, 2016
  f639c327
- AMDGPU: Set DX10Clamp bit · 7293f989
  Matt Arsenault authored Jan 28, 2016
  7293f989
- Revert r259035, it introduces a cyclic library dependency · 02fa1c80
  Oliver Stannard authored Jan 28, 2016
  02fa1c80
- Add backend dignostic printer for unsupported features · b4b092ea
  Oliver Stannard authored Jan 28, 2016
  b4b092ea
- Revert r258951 (and r258950), "Refactor backend diagnostics for unsupported features" · 628a7a0a
  NAKAMURA Takumi authored Jan 28, 2016
  628a7a0a
Jan 27, 2016
- Refactor backend diagnostics for unsupported features · 1e67a9f1
  Oliver Stannard authored Jan 27, 2016
  1e67a9f1
- AMDGPU/SI: Stoney has only 16 LDS banks · e86f2522
  Marek Olsak authored Jan 27, 2016
  e86f2522
- AMDGPU: Fix default device handling · b22828f2
  Matt Arsenault authored Jan 27, 2016
  b22828f2
Jan 26, 2016
- AMDGPU: Make v32i8/v64i8 illegal types · c5f61529
  Matt Arsenault authored Jan 26, 2016
  c5f61529
- AMDGPU: Remove old sample intrinsics · 018179fc
  Matt Arsenault authored Jan 26, 2016
  018179fc
- AMDGPU: Add new amdgcn intrinsics for cube instructions · 051d6f9f
  Matt Arsenault authored Jan 26, 2016
  051d6f9f
- AMDGPU: Implement read_register and write_register intrinsics · 9a10cea7
  Matt Arsenault authored Jan 26, 2016
  9a10cea7
- AMDGPU: Restore AMDGPU prefixed rsq intrinsic for now · 0c3e2338
  Matt Arsenault authored Jan 26, 2016
  0c3e2338
- [MC] Use .p2align instead of .align · 61d15ae4
  Dan Gohman authored Jan 26, 2016
  61d15ae4
Jan 23, 2016
- AMDGPU: Replace some deprecated intrinsic uses in tests · 8aa56781
  Matt Arsenault authored Jan 23, 2016
  8aa56781
- AMDGPU: Run instnamer on a few tests · 325cca33
  Matt Arsenault authored Jan 23, 2016
  325cca33
- AMDGPU: Remove more unused intrinsics · 7713162c
  Matt Arsenault authored Jan 23, 2016
  7713162c
Jan 22, 2016
- AMDGPU: Add new name for barrier intrinsic · 10ca39ca
  Matt Arsenault authored Jan 22, 2016
  10ca39ca
- AMDGPU: Rename intrinsics to use amdgcn prefix · bef34e21
  Matt Arsenault authored Jan 22, 2016
  bef34e21
- AMDGPU: Fix crash with invariant markers · 0b783ef0
  Matt Arsenault authored Jan 22, 2016
  0b783ef0
- AMDGPU: Rename some r600 intrinsics to use correct TargetPrefix · 59bd3014
  Matt Arsenault authored Jan 22, 2016
  59bd3014
- AMDGPU: Remove AMDGPU.fract intrinsic · 0cbaa176
  Matt Arsenault authored Jan 22, 2016
  0cbaa176
Jan 20, 2016
- AMDGPU/SI: Promote i1 SETCC operations · d1efda8e
  Tom Stellard authored Jan 20, 2016
  d1efda8e
- AMDGPU: Remove AMDGPU.trunc intrinsic · 7ba334a7
  Matt Arsenault authored Jan 20, 2016
  7ba334a7
- AMDGPU: Remove AMDIL.fraction intrinsic · 15fbe49d
  Matt Arsenault authored Jan 20, 2016
  15fbe49d
- AMDGPU: Remove AMDIL.round.nearest intrinsic · 7cccd267
  Matt Arsenault authored Jan 20, 2016
  7cccd267
- AMDGPU: Remove abs intrinsic · 1c9e4ef0
  Matt Arsenault authored Jan 20, 2016
  1c9e4ef0
- AMDGPU: Remove min/max intrinsics · f7e6e897
  Matt Arsenault authored Jan 20, 2016
  f7e6e897
- AMDGPU/SI: Prevent the DAGCombiner from creating setcc with i1 inputs · 2e045bbc
  Tom Stellard authored Jan 20, 2016
  2e045bbc
Jan 18, 2016
- AMDGPU: Reduce 64-bit SRAs · 33e3ecee
  Matt Arsenault authored Jan 18, 2016
  33e3ecee
- AMDGPU: Split 64-bit and of constant up · 6e3a4519
  Matt Arsenault authored Jan 18, 2016
  6e3a4519