Merge pull request #135 from LatticeQCD/hip_hotfix

Unrolling dslash loop only with CUDA backand.
LatticeQCD · Jun 1, 2023 · 207164b · 207164b
2 parents 795abfb + f01e2bd
commit 207164b
Show file tree

Hide file tree

Showing 2 changed files with 3 additions and 1 deletion.
diff --git a/README.md b/README.md
@@ -25,7 +25,7 @@ please have a look at the [Getting Started](https://latticeqcd.github.io/SIMULAT
 You will need to install [`git-lfs`](https://git-lfs.github.com/) before continuing or you will need to use a git client which natively supports it.
 This is needed for downloading configurations used in the unit tests. Then run 
 ```shell
-git clone https://github.com/LatticeQCD/SIMULATeQCD.git -b v1.0.0
+git clone https://github.com/LatticeQCD/SIMULATeQCD.git -b v1.0.1
 ```
 
 ### Compile Using Container

diff --git a/src/modules/dslash/dslash.cpp b/src/modules/dslash/dslash.cpp
@@ -13,7 +13,9 @@ __host__ __device__ auto HisqDslashFunctor<floatT, LatLayoutRHS, HaloDepthGauge,
 
     gVect3<floatT> Stmp(0.0);
 
+#ifdef USE_CUDA
 #pragma unroll
+#endif
     for (int mu = 0; mu < 4; mu++) {
 
         Stmp += static_cast<floatT>(C_1000) * _gAcc_smeared.getLink(GInd::template convertSite<All, HaloDepthGauge>(GInd::getSiteMu(site, mu)))