17ec3a51c3
This implementation was ported from the AMD builtin library
and has been tested with piglit, OpenCV, and the ocl conformance tests.
v2:
- Remove f suffix from constant in double implementations.
- Consolidate implementations using the .cl/.inc approach.
v3:
- Use __CLC_FPSIZE instead of __CLC_FP{32,64}
v4 (Jan Vesely):
- Limit to single precision.
llvm-svn: 236920
56 lines
893 B
C++
56 lines
893 B
C++
#define __CLC_FLOAT float
|
|
#define __CLC_FPSIZE 32
|
|
|
|
#define __CLC_FLOATN float
|
|
#define __CLC_SCALAR
|
|
#include __CLC_BODY
|
|
#undef __CLC_FLOATN
|
|
#undef __CLC_SCALAR
|
|
|
|
#define __CLC_FLOATN float2
|
|
#include __CLC_BODY
|
|
#undef __CLC_FLOATN
|
|
|
|
#define __CLC_FLOATN float3
|
|
#include __CLC_BODY
|
|
#undef __CLC_FLOATN
|
|
|
|
#define __CLC_FLOATN float4
|
|
#include __CLC_BODY
|
|
#undef __CLC_FLOATN
|
|
|
|
#undef __CLC_FLOAT
|
|
#undef __CLC_FPSIZE
|
|
|
|
#ifndef __FLOAT_ONLY
|
|
#ifdef cl_khr_fp64
|
|
|
|
#define __CLC_FLOAT double
|
|
#define __CLC_FPSIZE 64
|
|
|
|
#define __CLC_FLOATN double
|
|
#define __CLC_SCALAR
|
|
#include __CLC_BODY
|
|
#undef __CLC_FLOATN
|
|
#undef __CLC_SCALAR
|
|
|
|
#define __CLC_FLOATN double2
|
|
#include __CLC_BODY
|
|
#undef __CLC_FLOATN
|
|
|
|
#define __CLC_FLOATN double3
|
|
#include __CLC_BODY
|
|
#undef __CLC_FLOATN
|
|
|
|
#define __CLC_FLOATN double4
|
|
#include __CLC_BODY
|
|
#undef __CLC_FLOATN
|
|
|
|
#undef __CLC_FLOAT
|
|
#undef __CLC_FPSIZE
|
|
|
|
#endif
|
|
#endif
|
|
|
|
#undef __CLC_BODY
|