8: Fix OpenCL instructions and code examples

ENCCS · Nov 12, 2024 · 49e33d1 · 49e33d1
1 parent 9e85eb7
commit 49e33d1
Show file tree

Hide file tree

Showing 2 changed files with 3 additions and 3 deletions.
diff --git a/content/8-portable-kernel-models.rst b/content/8-portable-kernel-models.rst
@@ -197,8 +197,8 @@ To compile a program, you can use the AMD compilers on a GPU partition:
 
 .. code-block:: console
 
-    $ module load LUMI/23.03 partition/G
-    $ module load rocm/5.2.3
+    $ module load LUMI/24.03 partition/G
+    $ module load rocm/6.0.3
     $ module load PrgEnv-cray-amd
     $ CC program.cpp -lOpenCL -o program # C++ program
     $ cc program.c -lOpenCL -o program # C program

diff --git a/content/examples/portable-kernel-models/opencl-reduction.cpp b/content/examples/portable-kernel-models/opencl-reduction.cpp
@@ -19,7 +19,7 @@ static const std::string kernel_source = R"(
              
              // Perform reduction across the local work group
              for (int s = 1; s < lsize; s *= 2) { // loop over local memory with stride doubling each iteration
-               if (lid % (2 * s) == 0) {
+               if (lid % (2 * s) == 0 && (lid + s) < lsize) {
                  local_mem[lid] += local_mem[lid + s];
                }
                barrier(CLK_LOCAL_MEM_FENCE); // synchronize local memory