[SYCL] Add test for multiple compiler targets (#593)

Author: AidanBeltonS

SYCL/Regression/multiple-targets.cpp

@@ -0,0 +1,72 @@
+// This test checks if CUDA and HIP can be compiled and run with spirv.
+// It tests if the target triples can be specified with any order.
+// The test is repeated for per_kernel device code splitting.
+//
+// REQUIRES: CUDA || HIP
+// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple,spirv64 -o %t.out %s \
+// RUN: %GPU_RUN_PLACEHOLDER %t.out
+//
+// RUN: %clangxx -fsycl -fsycl-targets=spirv64,%sycl_triple -o %t.out %s \
+// RUN: %GPU_RUN_PLACEHOLDER %t.out
+//
+// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple,spirv64 -fsycl-device-code-split=per_kernel -o %t.out %s \
+// RUN: %GPU_RUN_PLACEHOLDER %t.out
+//
+// RUN: %clangxx -fsycl -fsycl-targets=spirv64,%sycl_triple -fsycl-device-code-split=per_kernel -o %t.out %s \
+// RUN: %GPU_RUN_PLACEHOLDER %t.out
+//
+// XFAIL: hip_nvidia
+//
+
+#include <CL/sycl.hpp>
+
+int main() {
+  cl::sycl::queue Q;
+  int A_Data[10] = {0};
+  int B_Data[10] = {4};
+  int C_Data[10] = {-1};
+
+  {
+    cl::sycl::buffer<int, 1> A_Buf(A_Data, cl::sycl::range<1>(10));
+
+    Q.submit([&](cl::sycl::handler &Cgh) {
+      auto A_Acc = A_Buf.get_access<cl::sycl::access::mode::write>(Cgh);
+      Cgh.parallel_for(cl::sycl::range<1>{5},
+                       [=](cl::sycl::id<1> index) { A_Acc[index] = 5; });
+    });
+  }
+
+  assert(A_Data[0] == 5);
+
+  {
+    cl::sycl::buffer<int, 1> B_Buf(B_Data, cl::sycl::range<1>(10));
+    cl::sycl::buffer<int, 1> C_Buf(C_Data, cl::sycl::range<1>(10));
+
+    Q.submit([&](cl::sycl::handler &Cgh) {
+      auto B_Acc = B_Buf.get_access<cl::sycl::access::mode::read_write>(Cgh);
+      auto C_Acc = C_Buf.get_access<cl::sycl::access::mode::read>(Cgh);
+      Cgh.parallel_for(cl::sycl::range<1>{5}, [=](cl::sycl::id<1> index) {
+        B_Acc[index] += C_Acc[index];
+      });
+    });
+  }
+
+  assert(B_Data[0] == 3);
+
+  {
+    cl::sycl::buffer<int, 1> B_Buf(B_Data, cl::sycl::range<1>(10));
+    cl::sycl::buffer<int, 1> C_Buf(C_Data, cl::sycl::range<1>(10));
+
+    Q.submit([&](cl::sycl::handler &Cgh) {
+      auto B_Acc = B_Buf.get_access<cl::sycl::access::mode::read>(Cgh);
+      auto C_Acc = C_Buf.get_access<cl::sycl::access::mode::write>(Cgh);
+      Cgh.parallel_for(cl::sycl::range<1>{5}, [=](cl::sycl::id<1> index) {
+        C_Acc[index] = B_Acc[index];
+      });
+    });
+  }
+
+  assert(C_Data[0] == 3);
+
+  return 0;
+}