SWDEV-286322 - clean up trailing space (#2361)

Change-Id: I03c07e67a8d1fa1a874718ffba43eb396c2aa05c

tests/src/deviceLib/hipBfloat16.cpp

@@ -55,17 +55,17 @@ __host__ __device__ void testOperations(float &fa, float &fb) {
   hip_bfloat16 bf_a(fa);
   hip_bfloat16 bf_b(fb);
   float fc = float(bf_a);
-  float fd = float(bf_b); 
+  float fd = float(bf_b);
 
   assert(testRelativeAccuracy(fa, bf_a));
   assert(testRelativeAccuracy(fb, bf_b));
 
   assert(testRelativeAccuracy(fc + fd, bf_a + bf_b));
-  //when checked as above for add, operation sub fails on GPU 
+  //when checked as above for add, operation sub fails on GPU
   assert(hip_bfloat16(fc - fd) == (bf_a - bf_b));
   assert(testRelativeAccuracy(fc * fd, bf_a * bf_b));
   assert(testRelativeAccuracy(fc / fd, bf_a / bf_b));
- 
+
   hip_bfloat16 bf_opNegate = -bf_a;
   assert(bf_opNegate == -bf_a);
 
@@ -75,7 +75,7 @@ __host__ __device__ void testOperations(float &fa, float &fb) {
   bf_x--;
   ++bf_x;
   --bf_x;
-  //hip_bfloat16 is converted to float and then inc/decremented, hence check with reduced precision 
+  //hip_bfloat16 is converted to float and then inc/decremented, hence check with reduced precision
   assert(testRelativeAccuracy(bf_x,bf_a));
 
   bf_x = bf_a;
@@ -95,7 +95,7 @@ __host__ __device__ void testOperations(float &fa, float &fb) {
   if (isnan(bf_rounded)) {
     assert(isnan(bf_rounded) || isinf(bf_rounded));
   }
-}  
+}
 
 __global__ void testOperationsGPU(float* d_a, float* d_b)
 {
@@ -126,7 +126,7 @@ int main(){
 
   hipLaunchKernelGGL(testOperationsGPU, 1, SIZE, 0, 0, d_fa, d_fb);
   hipDeviceSynchronize();
-  cout<<"Device bfloat16 Operations Successful!!"<<endl; 
+  cout<<"Device bfloat16 Operations Successful!!"<<endl;
 
   delete[] h_fa;
   delete[] h_fb;

tests/src/deviceLib/hipMathFunctions.cpp

@@ -56,21 +56,21 @@ __global__ void kernel_lgamma_double(double *input, double *output) {
 void check_lgamma_double() {
 
   using datatype_t = double;
-  
+
   const int NUM_INPUTS = 8;
   auto memsize = NUM_INPUTS * sizeof(datatype_t);
-  
+
   // allocate memories
   datatype_t *inputCPU = (datatype_t *) malloc(memsize);
   datatype_t *outputCPU = (datatype_t *) malloc(memsize);
   datatype_t *inputGPU = nullptr; hipMalloc((void**)&inputGPU, memsize);
   datatype_t *outputGPU = nullptr; hipMalloc((void**)&outputGPU, memsize);
-  
+
   // populate input
   for (int i=0; i<NUM_INPUTS; i++) {
     inputCPU[i] = -3.5 + i;
   }
-  
+
   // copy inputs to device
   hipMemcpy(inputGPU, inputCPU, memsize, hipMemcpyHostToDevice);
 
@@ -84,13 +84,13 @@ void check_lgamma_double() {
   for (int i=0; i<NUM_INPUTS; i++) {
     CHECK_LGAMMA_DOUBLE(inputCPU[i], outputCPU[i], lgamma(inputCPU[i]));
   }
-  
+
   // free memories
   hipFree(inputGPU);
   hipFree(outputGPU);
   free(inputCPU);
   free(outputCPU);
-  
+
   // done
   return;
 }
@@ -102,15 +102,15 @@ void check_abs_int64() {
 
   const int NUM_INPUTS = 8;
   auto memsize = NUM_INPUTS * sizeof(datatype_t);
-  
+
   // allocate memories
   datatype_t *inputCPU = (datatype_t *) malloc(memsize);
   datatype_t *outputCPU = (datatype_t *) malloc(memsize);
   datatype_t *inputGPU = nullptr; hipMalloc((void**)&inputGPU, memsize);
   datatype_t *outputGPU = nullptr; hipMalloc((void**)&outputGPU, memsize);
-  
+
   // populate input
-  inputCPU[0] = -81985529216486895ll; 
+  inputCPU[0] = -81985529216486895ll;
   inputCPU[1] =  81985529216486895ll;
   inputCPU[2] = -1250999896491ll;
   inputCPU[3] =  1250999896491ll;
@@ -118,7 +118,7 @@ void check_abs_int64() {
   inputCPU[5] =  19088743ll;
   inputCPU[6] = -291ll;
   inputCPU[7] =  291ll;
-  
+
   // copy inputs to device
   hipMemcpy(inputGPU, inputCPU, memsize, hipMemcpyHostToDevice);
 
@@ -137,17 +137,17 @@ void check_abs_int64() {
   CHECK_ABS_INT64(inputCPU[5], outputCPU[5], outputCPU[5]);
   CHECK_ABS_INT64(inputCPU[6], outputCPU[6], outputCPU[7]);
   CHECK_ABS_INT64(inputCPU[7], outputCPU[7], outputCPU[7]);
-  
+
   // free memories
   hipFree(inputGPU);
   hipFree(outputGPU);
   free(inputCPU);
   free(outputCPU);
-  
+
   // done
   return;
 }
-  
+
 
 template<class T, class F>
 __global__ void kernel_simple(F f, T *out) {
@@ -191,7 +191,7 @@ int main(int argc, char* argv[]) {
     check_abs_int64();
 
     // check_lgamma_double();
-    
+
     test_fp16();
 
     test_pown();

tests/src/deviceLib/hipStdComplex.cpp

@@ -82,7 +82,7 @@ __device__ __host__ complex<FloatT> calc(complex<FloatT> A,
     return A * B;
   case CK_div:
     return A / B;
-    
+
     ONE_ARG(abs)
     ONE_ARG(arg)
     ONE_ARG(sin)
@@ -111,7 +111,7 @@ void test() {
     hipMalloc((void**)&Ad, sizeof(ComplexT)*LEN);
     hipMalloc((void**)&Bd, sizeof(ComplexT)*LEN);
     hipMalloc((void**)&Cd, sizeof(ComplexT)*LEN);
-    
+
     for (uint32_t i = 0; i < LEN; i++) {
         A[i] = ComplexT((i + 1) * 1.0f, (i + 2) * 1.0f);
         B[i] = A[i];
@@ -119,7 +119,7 @@ void test() {
     }
     hipMemcpy(Ad, A, sizeof(ComplexT)*LEN, hipMemcpyHostToDevice);
     hipMemcpy(Bd, B, sizeof(ComplexT)*LEN, hipMemcpyHostToDevice);
-    
+
     // Run kernel for a calculation kind and verify by comparing with host
     // calculation result. Returns false if fails.
     auto test_fun = [&](enum CalcKind CK) {
@@ -145,7 +145,7 @@ void test() {
       }
       return true;
     };
-    
+
 #define OP(x) assert(test_fun(CK_##x));
     ALL_FUN
 #undef OP

tests/src/deviceLib/hipTestHalf.cpp

@@ -84,7 +84,7 @@ void kernel_hisnan(__half* input, int* output) {
 }
 
 __global__
-void kernel_hisinf(__half* input, int* output) { 
+void kernel_hisinf(__half* input, int* output) {
   int tx = threadIdx.x;
   output[tx] = __hisinf(input[tx]);
 }

tests/src/deviceLib/hipTestNew.cpp

@@ -41,7 +41,7 @@ THE SOFTWARE.
   private:
       int a;
   };
-     
+
   static __global__ void kernel(int* Ad) {
       int tid = threadIdx.x + blockIdx.x * blockDim.x;
       new(Ad+tid) A();