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

Fix inline compilation issue with LL (#1806)

Browse Source
This commit is contained in:
Wenkai Du authored and GitHub committed 2025-07-15 08:39:18 -07:00
1 parent 8d0f1a1cef
commit 106024b0db
1 file changed
+1 -1
Download Patch File Download Diff File Expand all files Collapse all files
src/device/prims_ll.h
+1 -1
View File
@@ -431,7 +431,7 @@ private:
}
template <int RECV, int SEND, int SrcBuf, int DstBuf>
__device__ void LLGenericOp(intptr_t srcIx, intptr_t dstIx, int nelem, bool postOp) {
__device__ __attribute__((noinline)) void LLGenericOp(intptr_t srcIx, intptr_t dstIx, int nelem, bool postOp) {
constexpr int SRC = SrcBuf != -1 ? 1 : 0;
constexpr int DST = DstBuf != -1 ? 1 : 0;
T *srcElts = SrcBuf == -1 ? nullptr : userBufs[SrcBuf] + srcIx;