Commits · ff630c2cdc7f805091e8d8cce40692bb9441fd3a · Roger Ferrer / llvm-epi · GitLab

Jun 21, 2018
- atom: Use volatile pointers for cl_khr_{global,local}_int32_{base,extended}_atomics · 8382e5bc
  Jan Vesely authored Jun 21, 2018
  8382e5bc
- atom: Consolidate cl_khr_{local,global}_int32_{base,extended}_atomics implementation · 65e3541b
  Jan Vesely authored Jun 21, 2018
  65e3541b
- atomic: Provide function implementation of atomic_{dec,inc} · f965b46c
  Jan Vesely authored Jun 21, 2018
  f965b46c
- atom: Consolidate cl_khr_int64_{base,extended}_atomics declarations · b9cbe0bf
  Jan Vesely authored Jun 21, 2018
  b9cbe0bf
- atom: Consolidate cl_khr_{local,global}_int32_{base,extended}_atomics declarations · d1c3811f
  Jan Vesely authored Jun 21, 2018
  d1c3811f
- atomic: Cleanup atomic_cmpxchg header · fe08de0c
  Jan Vesely authored Jun 21, 2018
  fe08de0c
- atomic: Move define cleanup to shared include · eabc1103
  Jan Vesely authored Jun 21, 2018
  eabc1103
Jun 18, 2018
- Update copyright year to 2018. · 7555c589
  Paul Robinson authored Jun 18, 2018
  7555c589
Jun 07, 2018
- r600/fmin: Flush denormals before calling builtin. · e0edcaa4
  Jan Vesely authored Jun 07, 2018
  e0edcaa4
- r600/fmax: Flush denormals before calling builtin. · e23c0ec0
  Jan Vesely authored Jun 07, 2018
  e23c0ec0
- math/fma: Add fp32 software implementation · 6e85e630
  Jan Vesely authored Jun 07, 2018
  6e85e630
May 18, 2018
- Add initial support for half precision builtins · 70a270da
  Jan Vesely authored May 17, 2018
  70a270da
May 15, 2018
- rootn: Use denormal path only · 58fdb3b0
  Jan Vesely authored May 15, 2018
  58fdb3b0
May 03, 2018
- remquo: Flush denormals if not supported · 21e77037
  Jan Vesely authored May 03, 2018
  21e77037
- remquo: Port from amd builtins · 8db45e4c
  Jan Vesely authored May 03, 2018
  8db45e4c
- math: Add helper function to flush denormals if not supported. · 6146eda7
  Jan Vesely authored May 03, 2018
  6146eda7
May 02, 2018
- clc_sqrt: Reuse unary_decl.inc · 616a38a6
  Jan Vesely authored May 02, 2018
  616a38a6
Apr 25, 2018
- relational/select: Condition types for half are short/ushort, not char/uchar · 1647e503
  Jan Vesely authored Apr 25, 2018
  1647e503
Apr 23, 2018
- log10: Use sw implementation from amd builtins · e7d567ee
  Jan Vesely authored Apr 23, 2018
  e7d567ee
Apr 17, 2018
- powr: Use denormal path only · 96591b62
  Jan Vesely authored Apr 17, 2018
  96591b62
- pown: Use denormal path only · 4388d288
  Jan Vesely authored Apr 17, 2018
  4388d288
- pow: Use denormal path only · 0d92f304
  Jan Vesely authored Apr 17, 2018
  0d92f304
- amdgcn/fmin: Fix typos that reduced precision · 8fa100df
  Jan Vesely authored Apr 17, 2018
  8fa100df
- exp10: Port from amd builtins · 15c388cd
  Jan Vesely authored Apr 17, 2018
  15c388cd
Apr 10, 2018
- hypot: Port from amd builtins · 4be03390
  Jan Vesely authored Apr 10, 2018
  4be03390
Apr 07, 2018
- select: simplify implementation and fix fp16 · 4c111261
  Jan Vesely authored Apr 06, 2018
  4c111261
Apr 06, 2018
- fmod: Port from amd_builtins · 93af9667
  Jan Vesely authored Apr 06, 2018
  93af9667
Apr 05, 2018
- r600: Update datalayout after LLVM r328656 · 92357a23
  Jan Vesely authored Apr 05, 2018
  92357a23
- amdgcn: Update datalayout after LLVM r328656 · fd11db19
  Jan Vesely authored Apr 05, 2018
  fd11db19
Mar 19, 2018
- remainder: Port from amd builtins · 5b10494f
  Jan Vesely authored Mar 19, 2018
  5b10494f
Mar 12, 2018
- nan: Implement · b672f7a2
  Jan Vesely authored Mar 12, 2018
  b672f7a2
- travis: Add build using llvm-6 · 08c96acb
  Jan Vesely authored Mar 12, 2018
  08c96acb
Mar 09, 2018
- amdgcn/fmax: fcanonicalize operands · f96b1b88
  Jan Vesely authored Mar 08, 2018
  f96b1b88
- amdgcn/fmin: fcanonicalize operands · e724e346
  Jan Vesely authored Mar 08, 2018
  e724e346
Mar 08, 2018
- amdgcn,popcount: Workaround broken llvm.ctpop intrinsic on some GCN ASICs · 04a46bf0
  Jan Vesely authored Mar 08, 2018
  04a46bf0
- integer/gentype: Add __CLC_VECSIZE macro · 0883c4d3
  Jan Vesely authored Mar 08, 2018
  0883c4d3
- popcount: Provide function implementation rather than intrinsic redirect · 17e86794
  Jan Vesely authored Mar 08, 2018
  17e86794
Mar 06, 2018
- lgamma_r: Move code from .inc to .cl file · c15a48dd
  Jan Vesely authored Mar 06, 2018
  c15a48dd
- frexp: Reuse types provided by gentype.inc · 2dcb382e
  Jan Vesely authored Mar 06, 2018
  2dcb382e
- select: Add vector implementation · ae156b66
  Jan Vesely authored Mar 06, 2018
  ae156b66