AI/rocm-systems
2
0
Fork 0
Code Issues Pull Requests Actions Packages Projects Releases Wiki Activity

Guard rcp rounded implementation as well

Browse Source 
Since rcp implementations of non-default rounded versions are not correct or supported in OCML, guard them using the same macro OCML_BASIC_ROUNDED_OPERATIONS. Also update the docs and tests.
This commit is contained in:
Aaron Enye Shi
2018-11-06 19:53:28 +00:00
parent e59c33250a
commit 0cfaa52d15
5 changed files with 15 additions and 21 deletions
Download Patch File Download Diff File Expand all files Collapse all files
include/hip/hcc_detail/math_functions.h
-4
View File
@@ -580,7 +580,6 @@ float __fmul_ru(float x, float y) { return __ocml_mul_rtp_f32(x, y); }
__DEVICE__
inline
float __fmul_rz(float x, float y) { return __ocml_mul_rtz_f32(x, y); }
#endif
__DEVICE__
inline
float __frcp_rd(float x) { return __llvm_amdgcn_rcp_f32(x); }
@@ -596,7 +595,6 @@ float __frcp_rz(float x) { return __llvm_amdgcn_rcp_f32(x); }
__DEVICE__
inline
float __frsqrt_rn(float x) { return __llvm_amdgcn_rsq_f32(x); }
#if defined OCML_BASIC_ROUNDED_OPERATIONS
__DEVICE__
inline
float __fsqrt_rd(float x) { return __ocml_sqrt_rtn_f32(x); }
@@ -1077,7 +1075,6 @@ double __dmul_ru(double x, double y) { return __ocml_mul_rtp_f64(x, y); }
__DEVICE__
inline
double __dmul_rz(double x, double y) { return __ocml_mul_rtz_f64(x, y); }
#endif
__DEVICE__
inline
double __drcp_rd(double x) { return __llvm_amdgcn_rcp_f64(x); }
@@ -1090,7 +1087,6 @@ double __drcp_ru(double x) { return __llvm_amdgcn_rcp_f64(x); }
__DEVICE__
inline
double __drcp_rz(double x) { return __llvm_amdgcn_rcp_f64(x); }
#if defined OCML_BASIC_ROUNDED_OPERATIONS
__DEVICE__
inline
double __dsqrt_rd(double x) { return __ocml_sqrt_rtn_f64(x); }