Align with HC Next.

tests/src/deviceLib/hipDoublePrecisionIntrinsics.cpp

@@ -28,6 +28,7 @@ THE SOFTWARE.
 
 #include <hip/hip_runtime.h>
 #include <hip/device_functions.h>
+#include <hip/math_functions.h>
 #include "test_common.h"
 
 #pragma GCC diagnostic ignored "-Wall"

tests/src/deviceLib/hipFloatMath.cpp

@@ -27,6 +27,7 @@ THE SOFTWARE.
  */
 
 #include "test_common.h"
+#include <hip/math_functions.h>
 #include <hip/device_functions.h>
 
 #define LEN 512

tests/src/deviceLib/hipIntegerIntrinsics.cpp

@@ -31,6 +31,10 @@ THE SOFTWARE.
 #include <hip/device_functions.h>
 #include "test_common.h"
 
+#include <algorithm>
+
+using namespace std;
+
 #pragma GCC diagnostic ignored "-Wall"
 #pragma clang diagnostic ignored "-Wunused-variable"
 

tests/src/deviceLib/hipMathFunctions.cpp

@@ -27,6 +27,7 @@ THE SOFTWARE.
  */
 
 #include "hip/hip_runtime.h"
+#include "hip/math_functions.h"
 #include "test_common.h"
 
 #if __HIP_ARCH_GFX803__ || __HIP_ARCH_GFX900__ || __HIP_ARCH_GFX906__

tests/src/deviceLib/hipSimpleAtomicsTest.cpp

@@ -215,6 +215,7 @@ template<
     typename T, 
     typename enable_if<
         is_same<T, int>{} || is_same<T, unsigned int>{}>::type* = nullptr>
+__device__
 void testKernelSub(T* g_odata) {
     // Atomic subtraction (final should be 0)
     atomicSub(&g_odata[1], 10);

tests/src/deviceLib/hipSinglePrecisionIntrinsics.cpp

@@ -27,6 +27,8 @@ THE SOFTWARE.
 
 #include <hip/hip_runtime.h>
 #include <hip/device_functions.h>
+#include <hip/math_functions.h>
+
 #include "test_common.h"
 
 #pragma GCC diagnostic ignored "-Wall"

tests/src/deviceLib/hip_trig.cpp

@@ -28,9 +28,11 @@ THE SOFTWARE.
 
 #include <hip/hip_runtime_api.h>
 #include <hip/hip_runtime.h>
-#include <iostream>
 #include "test_common.h"
 #include <hip/device_functions.h>
+#include <hip/math_functions.h>
+
+#include <iostream>
 
 #define LEN 512
 #define SIZE LEN << 2

tests/src/kernel/hipLanguageExtensions.cpp

@@ -31,7 +31,7 @@ THE SOFTWARE.
 #include <test_common.h>
 
 #ifdef __HCC__
-#include <amp.h>
+    #include <hc.hpp>
 #endif
 
 // cudaA
@@ -93,11 +93,7 @@ __global__ void vectorADD(const hipLaunchParm lp, T __restrict__* A_d, T* B_d, T
     int b = threadIdx.x;
     int c;
 
-    // TODO - move to HIP atomics when ready.
-    concurrency ::atomic_fetch_add(&c, b);
-    // Concurrency::atomic_add_unsigned (&x, a);
-
-    // concurrency ::atomic_add_ (x, a);
+    atomicAdd(&c, b);
 #endif
 
     __syncthreads();

tests/src/runtimeApi/stream/hipNullStream.cpp

@@ -26,11 +26,15 @@ THE SOFTWARE.
 
 #include "hip/hip_runtime.h"
 #include "test_common.h"
+
+#include <algorithm>
 #include <vector>
+
 unsigned p_streams = 16;
 int p_repeat = 10;
 int p_db = 0;
 
+using namespace std;
 
 template <typename T>
 __global__ void vectorADDRepeat(hipLaunchParm lp, const T* A_d, const T* B_d, T* C_d, size_t NELEM,