[PATCH] D85236: [CUDA] Work around a bug in rint() caused by a broken implementation provided by CUDA.

[Artem Belevich via Phabricator via cfe-commits]

tra updated this revision to Diff 283359.
tra edited the summary of this revision.
tra added a comment.

Also fixed the same bug in nearbyint().


Repository:
  rG LLVM Github Monorepo

CHANGES SINCE LAST ACTION
  https://reviews.llvm.org/D85236/new/

https://reviews.llvm.org/D85236

Files:
  clang/lib/Headers/__clang_cuda_math.h


Index: clang/lib/Headers/__clang_cuda_math.h
===================================================================
--- clang/lib/Headers/__clang_cuda_math.h
+++ clang/lib/Headers/__clang_cuda_math.h
@@ -195,8 +195,8 @@
 __DEVICE__ int min(int __a, int __b) { return __nv_min(__a, __b); }
 __DEVICE__ double modf(double __a, double *__b) { return __nv_modf(__a, __b); }
 __DEVICE__ float modff(float __a, float *__b) { return __nv_modff(__a, __b); }
-__DEVICE__ double nearbyint(double __a) { return __nv_nearbyint(__a); }
-__DEVICE__ float nearbyintf(float __a) { return __nv_nearbyintf(__a); }
+__DEVICE__ double nearbyint(double __a) { return __builtin_nearbyint(__a); }
+__DEVICE__ float nearbyintf(float __a) { return __builtin_nearbyintf(__a); }
 __DEVICE__ double nextafter(double __a, double __b) {
   return __nv_nextafter(__a, __b);
 }
@@ -249,8 +249,9 @@
 __DEVICE__ float rhypotf(float __a, float __b) {
   return __nv_rhypotf(__a, __b);
 }
-__DEVICE__ double rint(double __a) { return __nv_rint(__a); }
-__DEVICE__ float rintf(float __a) { return __nv_rintf(__a); }
+// __nv_rint* in libdevice is buggy and produces incorrect results.
+__DEVICE__ double rint(double __a) { return __builtin_rint(__a); }
+__DEVICE__ float rintf(float __a) { return __builtin_rintf(__a); }
 __DEVICE__ double rnorm(int __a, const double *__b) {
   return __nv_rnorm(__a, __b);
 }


-------------- next part --------------
A non-text attachment was scrubbed...
Name: D85236.283359.patch
Type: text/x-patch
Size: 1371 bytes
Desc: not available
URL: <http://lists.llvm.org/pipermail/cfe-commits/attachments/20200805/a085d497/attachment.bin>