Revert adoption of CUDA indexing in general - this can only work with later versions of the compiler, just like module based dispatch, and thus must be guarded against usage in earlier (e.g. 1.6) versions.

hipamd/tests/src/deviceLib/hipSimpleAtomicsTest.cpp

@@ -217,7 +217,7 @@ int computeGold(int *gpuData, const int len)
 __global__ void testKernel(hipLaunchParm lp,int *g_odata)
 {
     // access thread id
-    const unsigned int tid = blockDim.x * blockIdx.x + threadIdx.x;
+    const unsigned int tid = hipBlockDim_x * hipBlockIdx_x + hipThreadIdx_x;
 
     // Test various atomic instructions