Commits · f7f59b529293136d0f711b86dbfccb525d0677d4 · Roger Ferrer / llvm-epi · GitLab

Dec 18, 2017
- [X86] Use mattr instead of mcpu in some of the cost model tests. · 7034d401
  Craig Topper authored Dec 18, 2017
  7034d401
Nov 29, 2017
- [X86] Promote fp_to_sint v16f32->v16i16/v16i8 to avoid scalarization. · fbf7b3bf
  Craig Topper authored Nov 29, 2017
  fbf7b3bf
Nov 20, 2017
- [LV][X86] Support of AVX2 Gathers code generation and update the LV with this · 115f68ea
  Mohammed Agabaria authored Nov 20, 2017
  115f68ea
Nov 16, 2017
- [TTI][X86] update costs of interleaved load\store of i64\double · 6e6d5326
  Mohammed Agabaria authored Nov 16, 2017
  6e6d5326
Nov 06, 2017
- [LV][X86] update the cost of interleaving mem. access of floats · 66917583
  Mohammed Agabaria authored Nov 06, 2017
  66917583
Nov 05, 2017
- [REVERT][LV][X86] update the cost of interleaving mem. access of floats · acd69dbc
  Mohammed Agabaria authored Nov 05, 2017
  acd69dbc
- [LV][X86] update the cost of interleaving mem. access of floats · f74c767d
  Mohammed Agabaria authored Nov 05, 2017
  f74c767d
Oct 18, 2017
- [AVX512][AVX2]Cost calculation for interleave load/store patterns {v8i8,v16i8,v32i8,v64i8} · 49293264
  Michael Zuckerman authored Oct 18, 2017
  49293264
Oct 13, 2017
- Revert r314923: "Recommit : Use the basic cost if a GEP is not used as addressing mode" · 3344a212
  Daniel Jasper authored Oct 13, 2017
  3344a212
Oct 04, 2017
- Convert an APInt to int64_t properly in TTI::getGEPCost(). · f84f7c74
  Justin Lebar authored Oct 04, 2017
  f84f7c74
- [TargetTransformInfo] Check if function pointer is valid before calling isLoweredToCall · eb301875
  Guozhi Wei authored Oct 04, 2017
  eb301875
- Recommit : Use the basic cost if a GEP is not used as addressing mode · d40e03c2
  Jun Bum Lim authored Oct 04, 2017
  d40e03c2
Oct 03, 2017
- [X86] Add AVX512 check lines to the cost model truncate test. · 360c816c
  Craig Topper authored Oct 03, 2017
  360c816c
Sep 30, 2017
- Revert "Use the basic cost if a GEP is not used as addressing mode" · e76aa3b0
  Alex Shlyapnikov authored Sep 29, 2017
  e76aa3b0
Sep 29, 2017
- Use the basic cost if a GEP is not used as addressing mode · 0e16a59e
  Jun Bum Lim authored Sep 29, 2017
  0e16a59e
Sep 28, 2017
- Check for overflows when calculating the offset in GetGEPCost. · 8ea84426
  Justin Lebar authored Sep 27, 2017
  8ea84426
Sep 22, 2017
- [TargetTransformInfo] Handle intrinsic call in getInstructionLatency() · bce228ca
  Guozhi Wei authored Sep 22, 2017
  bce228ca
Sep 16, 2017
- [TargetTransformInfo] Static alloca has 0 cost · 3d1305f6
  Guozhi Wei authored Sep 15, 2017
  3d1305f6
Sep 14, 2017
- [TargetTransformInfo] Detect 0 latency instructions · 21f8fad9
  Guozhi Wei authored Sep 14, 2017
  21f8fad9
Sep 09, 2017
- [TargetTransformInfo] Add a new public interface getInstructionCost · 62d64144
  Guozhi Wei authored Sep 08, 2017
  62d64144
Sep 07, 2017
- X86: Improve AVX512 fptoui lowering · 25799d93
  Zvi Rackover authored Sep 07, 2017
  25799d93
Aug 31, 2017
- AMDGPU: Don't assert in TTI with fp32 denorms enabled · 376f1bd7
  Matt Arsenault authored Aug 31, 2017
  376f1bd7
Aug 16, 2017
- [CostModel][X86][XOP] Improve costs for XOP shuffles · c63f93a1
  Simon Pilgrim authored Aug 16, 2017
  c63f93a1
Aug 10, 2017
- [CostModel][X86] Add SSE2 two-src shuffle costs · b59c2d9d
  Simon Pilgrim authored Aug 10, 2017
  b59c2d9d
- [CostModel][X86] Add avx1 two-src shuffle costs · 7354531b
  Simon Pilgrim authored Aug 10, 2017
  7354531b
- [CostModel][X86] Add avx2 two-src shuffle costs · ac2e50a4
  Simon Pilgrim authored Aug 10, 2017
  ac2e50a4
- [CostModel][X86] Extend two src shuffle cost tests · 2f529412
  Simon Pilgrim authored Aug 10, 2017
  2f529412
- [CostModel][X86] Add avx512vbmi broadcast/reverse/single-src shuffle cost tests · fe67612e
  Simon Pilgrim authored Aug 10, 2017
  fe67612e
- [CostModel][X86] Improve single src shuffle costs · 702e5fa3
  Simon Pilgrim authored Aug 10, 2017
  702e5fa3
- [CostModel][X86] Added v2f64/v2i64 single src shuffle model tests · 419215ab
  Simon Pilgrim authored Aug 10, 2017
  419215ab
Jul 17, 2017
- [SystemZ] Add support for IBM z14 processor (2/3) · 33435c4c
  Ulrich Weigand authored Jul 17, 2017
  33435c4c
Jul 02, 2017
- [X86][CM] update add\sub costs of vectors of 64 in X86\SLM arch · eb09a810
  Mohammed Agabaria authored Jul 02, 2017
  eb09a810
Jun 25, 2017
- [AVX2] [TTI CostModel] Add cost of interleaved loads/stores for AVX2 · e0e0f1dd
  Dorit Nuzman authored Jun 25, 2017
  e0e0f1dd
Jun 20, 2017
- [CostModel][X86] Add scalar arithmetic cost tests · 68204b83
  Simon Pilgrim authored Jun 20, 2017
  68204b83
- [CostModel][X86] Declare costs variables based on type · 36c17935
  Simon Pilgrim authored Jun 20, 2017
  36c17935
May 24, 2017
- Revert r291254: [AArch64] Reduce vector insert/extract cost for Falkor · 6349380f
  Matthew Simpson authored May 24, 2017
  6349380f
May 19, 2017
- Fix line-endings. · c74e7f0a
  Simon Pilgrim authored May 19, 2017
  c74e7f0a
May 18, 2017
- [X86][AVX512] Add 512-bit vector ctpop costs + tests · 6bba6068
  Simon Pilgrim authored May 18, 2017
  6bba6068
May 17, 2017
- [X86][AVX512] Add 512-bit vector ctlz costs + tests · 23ef2672
  Simon Pilgrim authored May 17, 2017
  23ef2672
- [X86][AVX512] Add 512-bit vector cttz costs + tests · d0365967
  Simon Pilgrim authored May 17, 2017
  d0365967