Catch2 Test Fixes (#2961)

* test fixes

* address PR comment

* PR comment fixing Nvidia pass

* add additional kernel launch checks

* pr comments

tests/catch/unit/multiThread/hipMultiThreadStreams1.cc

@@ -74,6 +74,7 @@ void simpleVectorAdd(size_t numElements, int iters, hipStream_t stream) {
 
         hipLaunchKernelGGL(HipTest::vectorADDReverse, dim3(blocks), dim3(threadsPerBlock), 0, 0,
                         static_cast<const T*>(A_d), static_cast<const T*>(B_d), C_d, numElements);
+        HIP_CHECK(hipGetLastError()); 
 
         MemTraits<C>::Copy(C_h, C_d, Nbytes, hipMemcpyDeviceToHost, stream);
 

tests/catch/unit/multiThread/hipMultiThreadStreams2.cc

@@ -56,6 +56,7 @@ void run1(size_t size, hipStream_t stream) {
     HIPCHECK(hipMemcpyAsync(Bh, Ah, size, hipMemcpyHostToHost, stream));
     HIPCHECK(hipMemcpyAsync(Cd, Bh, size, hipMemcpyHostToDevice, stream));
     hipLaunchKernelGGL(HIP_KERNEL_NAME(Inc), dim3(N / 500), dim3(500), 0, stream, Cd);
+    HIP_CHECK(hipGetLastError()); 
     HIPCHECK(hipMemcpyAsync(Dd, Cd, size, hipMemcpyDeviceToDevice, stream));
     HIPCHECK(hipMemcpyAsync(Eh, Dd, size, hipMemcpyDeviceToHost, stream));
     HIPCHECK(hipDeviceSynchronize());
@@ -97,7 +98,9 @@ void run(size_t size, hipStream_t stream1, hipStream_t stream2) {
     HIPCHECK(hipMemcpyAsync(Cd, Bh, size, hipMemcpyHostToDevice, stream1));
     HIPCHECK(hipMemcpyAsync(Cdd, Bhh, size, hipMemcpyHostToDevice, stream2));
     hipLaunchKernelGGL(HIP_KERNEL_NAME(Inc), dim3(N / 500), dim3(500), 0, stream1, Cd);
+    HIP_CHECK(hipGetLastError()); 
     hipLaunchKernelGGL(HIP_KERNEL_NAME(Inc), dim3(N / 500), dim3(500), 0, stream2, Cdd);
+    HIP_CHECK(hipGetLastError()); 
     HIPCHECK(hipMemcpyAsync(Dd, Cd, size, hipMemcpyDeviceToDevice, stream1));
     HIPCHECK(hipMemcpyAsync(Ddd, Cdd, size, hipMemcpyDeviceToDevice, stream2));
     HIPCHECK(hipMemcpyAsync(Eh, Dd, size, hipMemcpyDeviceToHost, stream1));