SWDEV-277697 - [CatchTest] Fix Documentation, Add test to AMD specific, Add HIP Macros, New Binary for multiproc tests

Change-Id: I3783caf85c694b724ed55b778220b8ef9a39f84b

catch/unit/deviceLib/CMakeLists.txt

@@ -1,9 +1,18 @@
 # Common Tests - Test independent of all platforms
 set(TEST_SRC
     floatMath.cc
+)
+
+# AMD only tests
+set(AMD_TEST_SRC
     vectorTypesDevice.cc
 )
 
+if(HIP_PLATFORM MATCHES "amd")
+    set(TEST_SRC ${TEST_SRC} ${AMD_TEST_SRC})
+endif()
+
+
 # Create shared lib of all tests
 add_library(DeviceLibs SHARED EXCLUDE_FROM_ALL ${TEST_SRC})
 

catch/unit/kernels/add.cc

@@ -2,12 +2,11 @@
 #include <iostream>
 
 template <typename T> __global__ void add(T* a, T* b, T* c, size_t size) {
-  int i = threadIdx.x;
-  c[i] = a[i] + b[i];
+  size_t i = threadIdx.x;
+  if (i < size) c[i] = a[i] + b[i];
 }
 
 TEMPLATE_TEST_CASE("Add Kernel", "[kernel][add]", int, long, float, long long, double) {
-  auto addKernel = add<TestType>;
   auto size = GENERATE(as<size_t>{}, 100, 500, 1000);
   TestType *d_a, *d_b, *d_c;
   auto res = hipMalloc(&d_a, sizeof(TestType) * size);
@@ -18,7 +17,7 @@ TEMPLATE_TEST_CASE("Add Kernel", "[kernel][add]", int, long, float, long long, d
   REQUIRE(res == hipSuccess);
 
   std::vector<TestType> a, b, c;
-  for (int i = 0; i < size; i++) {
+  for (size_t i = 0; i < size; i++) {
     a.push_back(i + 1);
     b.push_back(i + 1);
     c.push_back(2 * (i + 1));
@@ -29,7 +28,7 @@ TEMPLATE_TEST_CASE("Add Kernel", "[kernel][add]", int, long, float, long long, d
   res = hipMemcpy(d_b, b.data(), sizeof(TestType) * size, hipMemcpyHostToDevice);
   REQUIRE(res == hipSuccess);
 
-  hipLaunchKernelGGL(addKernel, 1, size, 0, 0, d_a, d_b, d_c, size);
+  hipLaunchKernelGGL(add<TestType>, 1, size, 0, 0, d_a, d_b, d_c, size);
 
   res = hipMemcpy(a.data(), d_c, sizeof(TestType) * size, hipMemcpyDeviceToHost);
   REQUIRE(res == hipSuccess);

catch/unit/rtc/CMakeLists.txt

@@ -1,14 +1,12 @@
-# Common Tests - Test independent of all platforms
-set(TEST_SRC
+# AMD Tests
+set(AMD_TEST_SRC
     saxpy.cc
 )
 
-# Set source File properties
-set_source_files_properties(saxpy.cc PROPERTIES COMPILE_FLAGS " -std=c++14 ")
-set_source_files_properties(test.cc PROPERTIES COMPILE_FLAGS " -std=c++17 ")
+if(HIP_PLATFORM MATCHES "amd")
+    # Create shared lib of all tests
+    add_library(RTC SHARED EXCLUDE_FROM_ALL ${AMD_TEST_SRC})
 
-# Create shared lib of all tests
-add_library(RTC SHARED EXCLUDE_FROM_ALL ${TEST_SRC})
-
-# Add dependency on build_tests to build it on this custom target
-add_dependencies(build_tests RTC)
+    # Add dependency on build_tests to build it on this custom target
+    add_dependencies(build_tests RTC)
+endif()

catch/unit/rtc/saxpy.cc

@@ -15,7 +15,7 @@ static constexpr auto NUM_THREADS{128};
 static constexpr auto NUM_BLOCKS{32};
 
 static constexpr auto saxpy{
-    R"(
+R"(
 #include <hip/hip_runtime.h>
 extern "C"
 __global__
@@ -23,8 +23,7 @@ void saxpy(float a, float* x, float* y, float* out, size_t n)
 {
     size_t tid = blockIdx.x * blockDim.x + threadIdx.x;
         if (tid < n) {
-               out[tid] = a * x[tid] + y[tid] ;
-                   
+               out[tid] = a * x[tid] + y[tid];
         }
         
 }
@@ -72,42 +71,42 @@ TEST_CASE("saxpy", "[hiprtc][saxpy]") {
   unique_ptr<float[]> hX{new float[n]};
   unique_ptr<float[]> hY{new float[n]};
   unique_ptr<float[]> hOut{new float[n]};
-for (size_t i = 0; i < n; ++i) {
-        hX[i] = static_cast<float>(i);
-        hY[i] = static_cast<float>(i * 2);
-    }
+  for (size_t i = 0; i < n; ++i) {
+    hX[i] = static_cast<float>(i);
+    hY[i] = static_cast<float>(i * 2);
+  }
 
-    hipDeviceptr_t dX, dY, dOut;
-    hipMalloc(&dX, bufferSize);
-    hipMalloc(&dY, bufferSize);
-    hipMalloc(&dOut, bufferSize);
-    hipMemcpyHtoD(dX, hX.get(), bufferSize);
-    hipMemcpyHtoD(dY, hY.get(), bufferSize);
+  hipDeviceptr_t dX, dY, dOut;
+  hipMalloc(&dX, bufferSize);
+  hipMalloc(&dY, bufferSize);
+  hipMalloc(&dOut, bufferSize);
+  hipMemcpyHtoD(dX, hX.get(), bufferSize);
+  hipMemcpyHtoD(dY, hY.get(), bufferSize);
 
-    struct {
-        float a_;
-        hipDeviceptr_t b_;
-        hipDeviceptr_t c_;
-        hipDeviceptr_t d_;
-        size_t e_;
-    } args{a, dX, dY, dOut, n};
+  struct {
+    float a_;
+    hipDeviceptr_t b_;
+    hipDeviceptr_t c_;
+    hipDeviceptr_t d_;
+    size_t e_;
+  } args{a, dX, dY, dOut, n};
 
-    auto size = sizeof(args);
-    void* config[] = {HIP_LAUNCH_PARAM_BUFFER_POINTER, &args,
-                      HIP_LAUNCH_PARAM_BUFFER_SIZE, &size,
-                      HIP_LAUNCH_PARAM_END};
+  auto size = sizeof(args);
+  void* config[] = {HIP_LAUNCH_PARAM_BUFFER_POINTER, &args, HIP_LAUNCH_PARAM_BUFFER_SIZE, &size,
+                    HIP_LAUNCH_PARAM_END};
 
-    hipModuleLaunchKernel(kernel, NUM_BLOCKS, 1, 1, NUM_THREADS, 1, 1,
-                          0, nullptr, nullptr, config);
-    hipMemcpyDtoH(hOut.get(), dOut, bufferSize);
+  hipModuleLaunchKernel(kernel, NUM_BLOCKS, 1, 1, NUM_THREADS, 1, 1, 0, nullptr, nullptr, config);
+  hipMemcpyDtoH(hOut.get(), dOut, bufferSize);
 
-    for (size_t i = 0; i < n; ++i) {
-      REQUIRE(fabs(a * hX[i] + hY[i] - hOut[i]) > fabs(hOut[i]) * 1e-6);
-    }
+  for (size_t i = 0; i < n; ++i) {
+    INFO("For " << i << " Value: " << fabs(a * hX[i] + hY[i] - hOut[i])
+                << " with: " << (fabs(hOut[i] * 1.0f) * 1e-6));
+    REQUIRE(fabs(a * hX[i] + hY[i] - hOut[i]) <= fabs(hOut[i]) * 1e-6);
+  }
 
-    hipFree(dX);
-    hipFree(dY);
-    hipFree(dOut);
+  hipFree(dX);
+  hipFree(dY);
+  hipFree(dOut);
 
-    hipModuleUnload(module);
+  hipModuleUnload(module);
 }

catch/unit/rtc/test.cc

@@ -1,6 +0,0 @@
-#include <hip_test_common.hh>
-
-TEST_CASE("cpp17 test") {
-  constexpr auto l = []() { return 2 * 10 * 30; };
-  REQUIRE(l() == 600);
-}