Merge branch 'intel' into cperkins-complex-group-algo-support

Author: cperkinsintel

.gitignore

@@ -2,3 +2,4 @@
 # External/*
 /test-suite-externals
 *.pyc
+/build*

SYCL/Assert/assert_in_kernels.cpp

@@ -1,7 +1,7 @@
 // REQUIRES: linux
 // FIXME unsupported on CUDA and HIP until fallback libdevice becomes available
 // UNSUPPORTED: cuda || hip
-// RUN: %clangxx -DSYCL_ENABLE_FALLBACK_ASSERT -fsycl -fsycl-targets=%sycl_triple %s -o %t.out
+// RUN: %clangxx -DSYCL_FALLBACK_ASSERT=1 -fsycl -fsycl-targets=%sycl_triple %s -o %t.out
 // RUN: %CPU_RUN_PLACEHOLDER %t.out &> %t.txt || true
 // RUN: %CPU_RUN_PLACEHOLDER FileCheck %s --input-file %t.txt
 // RUN: %GPU_RUN_PLACEHOLDER %t.out &> %t.txt || true

SYCL/Assert/assert_in_kernels_win.cpp

@@ -1,6 +1,6 @@
 // REQUIRES: windows
 // UNSUPPORTED: cuda || hip
-// RUN: %clangxx -DSYCL_ENABLE_FALLBACK_ASSERT -fsycl -fsycl-targets=%sycl_triple %s -o %t.out
+// RUN: %clangxx -DSYCL_FALLBACK_ASSERT=1 -fsycl -fsycl-targets=%sycl_triple %s -o %t.out
 // RUN: %CPU_RUN_PLACEHOLDER %t.out &> %t.txt || true
 // RUN: %CPU_RUN_PLACEHOLDER FileCheck %s --input-file %t.txt
 // RUN: %GPU_RUN_PLACEHOLDER %t.out &> %t.txt || true

SYCL/Assert/assert_in_multiple_tus.cpp

@@ -1,7 +1,7 @@
 // REQUIRES: linux
 // FIXME unsupported on CUDA and HIP until fallback libdevice becomes available
 // UNSUPPORTED: cuda || hip
-// RUN: %clangxx -DSYCL_ENABLE_FALLBACK_ASSERT -fsycl -fsycl-targets=%sycl_triple -I %S/Inputs %s %S/Inputs/kernels_in_file2.cpp -o %t.out
+// RUN: %clangxx -DSYCL_FALLBACK_ASSERT=1 -fsycl -fsycl-targets=%sycl_triple -I %S/Inputs %s %S/Inputs/kernels_in_file2.cpp -o %t.out
 // RUN: %CPU_RUN_PLACEHOLDER %t.out &> %t.txt || true
 // RUN: %CPU_RUN_PLACEHOLDER FileCheck %s --input-file %t.txt
 // RUN: %GPU_RUN_PLACEHOLDER %t.out &> %t.txt || true

SYCL/Assert/assert_in_multiple_tus_one_ndebug.cpp

@@ -1,7 +1,7 @@
 // REQUIRES: linux
 // FIXME unsupported on CUDA and HIP until fallback libdevice becomes available
 // UNSUPPORTED: cuda || hip
-// RUN: %clangxx -DSYCL_ENABLE_FALLBACK_ASSERT -fsycl -fsycl-targets=%sycl_triple -DDEFINE_NDEBUG_INFILE2 -I %S/Inputs %S/assert_in_multiple_tus.cpp %S/Inputs/kernels_in_file2.cpp -o %t.out
+// RUN: %clangxx -DSYCL_FALLBACK_ASSERT=1 -fsycl -fsycl-targets=%sycl_triple -DDEFINE_NDEBUG_INFILE2 -I %S/Inputs %S/assert_in_multiple_tus.cpp %S/Inputs/kernels_in_file2.cpp -o %t.out
 // RUN: %CPU_RUN_PLACEHOLDER %t.out &> %t.txt || true
 // RUN: %CPU_RUN_PLACEHOLDER FileCheck %s --input-file %t.txt
 // RUN: %GPU_RUN_PLACEHOLDER %t.out &> %t.txt || true

SYCL/Assert/assert_in_multiple_tus_one_ndebug_win.cpp

@@ -1,6 +1,6 @@
 // REQUIRES: windows
 // UNSUPPORTED: cuda || hip
-// RUN: %clangxx -DSYCL_ENABLE_FALLBACK_ASSERT -fsycl -fsycl-targets=%sycl_triple -DDEFINE_NDEBUG_INFILE2 -I %S/Inputs %S/assert_in_multiple_tus.cpp %S/Inputs/kernels_in_file2.cpp -o %t.out
+// RUN: %clangxx -DSYCL_FALLBACK_ASSERT=1 -fsycl -fsycl-targets=%sycl_triple -DDEFINE_NDEBUG_INFILE2 -I %S/Inputs %S/assert_in_multiple_tus.cpp %S/Inputs/kernels_in_file2.cpp -o %t.out
 // RUN: %CPU_RUN_PLACEHOLDER %t.out &> %t.txt || true
 // RUN: %CPU_RUN_PLACEHOLDER FileCheck %s --input-file %t.txt
 // RUN: %GPU_RUN_PLACEHOLDER %t.out &> %t.txt || true

SYCL/Assert/assert_in_multiple_tus_win.cpp

@@ -1,6 +1,6 @@
 // REQUIRES: windows
 // UNSUPPORTED: cuda || hip
-// RUN: %clangxx -DSYCL_ENABLE_FALLBACK_ASSERT -fsycl -fsycl-targets=%sycl_triple -I %S/Inputs %s %S/Inputs/kernels_in_file2.cpp -o %t.out
+// RUN: %clangxx -DSYCL_FALLBACK_ASSERT=1 -fsycl -fsycl-targets=%sycl_triple -I %S/Inputs %s %S/Inputs/kernels_in_file2.cpp -o %t.out
 // RUN: %CPU_RUN_PLACEHOLDER %t.out &> %t.txt || true
 // RUN: %CPU_RUN_PLACEHOLDER FileCheck %s --input-file %t.txt
 // RUN: %GPU_RUN_PLACEHOLDER %t.out &> %t.txt || true

SYCL/Assert/assert_in_one_kernel.cpp

@@ -1,7 +1,7 @@
 // REQUIRES: linux
 // FIXME unsupported on CUDA and HIP until fallback libdevice becomes available
 // UNSUPPORTED: cuda || hip
-// RUN: %clangxx -DSYCL_ENABLE_FALLBACK_ASSERT -fsycl -fsycl-targets=%sycl_triple %s -o %t.out
+// RUN: %clangxx -DSYCL_FALLBACK_ASSERT=1 -fsycl -fsycl-targets=%sycl_triple %s -o %t.out
 // RUN: %CPU_RUN_PLACEHOLDER %t.out &> %t.txt || true
 // RUN: %CPU_RUN_PLACEHOLDER FileCheck %s --input-file %t.txt
 // RUN: %GPU_RUN_PLACEHOLDER %t.out &> %t.txt || true

SYCL/Assert/assert_in_one_kernel_win.cpp

@@ -1,6 +1,6 @@
 // REQUIRES: windows
 // UNSUPPORTED: cuda || hip
-// RUN: %clangxx -DSYCL_ENABLE_FALLBACK_ASSERT -fsycl -fsycl-targets=%sycl_triple %s -o %t.out
+// RUN: %clangxx -DSYCL_FALLBACK_ASSERT=1 -fsycl -fsycl-targets=%sycl_triple %s -o %t.out
 // RUN: %CPU_RUN_PLACEHOLDER %t.out &> %t.txt || true
 // RUN: %CPU_RUN_PLACEHOLDER FileCheck %s --input-file %t.txt
 // RUN: %GPU_RUN_PLACEHOLDER %t.out &> %t.txt || true

SYCL/Assert/assert_in_simultaneous_kernels.cpp

@@ -1,7 +1,7 @@
 // REQUIRES: linux
 // FIXME unsupported on CUDA and HIP until fallback libdevice becomes available
 // UNSUPPORTED: cuda || hip
-// RUN: %clangxx -DSYCL_ENABLE_FALLBACK_ASSERT -fsycl -fsycl-targets=%sycl_triple %s -o %t.out %threads_lib
+// RUN: %clangxx -DSYCL_FALLBACK_ASSERT=1 -fsycl -fsycl-targets=%sycl_triple %s -o %t.out %threads_lib
 // RUN: %CPU_RUN_PLACEHOLDER %t.out &> %t.txt || true
 // RUN: %CPU_RUN_PLACEHOLDER FileCheck %s --input-file %t.txt
 //

SYCL/Assert/assert_in_simultaneous_kernels_win.cpp

@@ -1,6 +1,6 @@
 // REQUIRES: windows
 // UNSUPPORTED: cuda || hip
-// RUN: %clangxx -DSYCL_ENABLE_FALLBACK_ASSERT -fsycl -fsycl-targets=%sycl_triple %s -o %t.out %threads_lib
+// RUN: %clangxx -DSYCL_FALLBACK_ASSERT=1 -fsycl -fsycl-targets=%sycl_triple %s -o %t.out %threads_lib
 // RUN: %CPU_RUN_PLACEHOLDER %t.out &> %t.txt || true
 // RUN: %CPU_RUN_PLACEHOLDER FileCheck %s --input-file %t.txt
 //

SYCL/Assert/assert_in_simultaneously_multiple_tus.cpp

@@ -1,6 +1,6 @@
 // FIXME unsupported on CUDA and HIP until fallback libdevice becomes available
 // UNSUPPORTED: cuda || hip
-// RUN: %clangxx -DSYCL_ENABLE_FALLBACK_ASSERT -fsycl -fsycl-targets=%sycl_triple -I %S/Inputs %s %S/Inputs/kernels_in_file2.cpp -o %t.out %threads_lib
+// RUN: %clangxx -DSYCL_FALLBACK_ASSERT=1 -fsycl -fsycl-targets=%sycl_triple -I %S/Inputs %s %S/Inputs/kernels_in_file2.cpp -o %t.out %threads_lib
 // RUN: %CPU_RUN_PLACEHOLDER %t.out &> %t.txt || true
 // RUN: %CPU_RUN_PLACEHOLDER FileCheck %s --input-file %t.txt
 //

SYCL/Assert/assert_in_simultaneously_multiple_tus_one_ndebug.cpp

@@ -1,6 +1,6 @@
 // FIXME unsupported on CUDA and HIP until fallback libdevice becomes available
 // UNSUPPORTED: cuda || hip
-// RUN: %clangxx -DSYCL_ENABLE_FALLBACK_ASSERT -fsycl -fsycl-targets=%sycl_triple -DDEFINE_NDEBUG_INFILE2 -I %S/Inputs %S/assert_in_simultaneously_multiple_tus.cpp %S/Inputs/kernels_in_file2.cpp -o %t.out %threads_lib
+// RUN: %clangxx -DSYCL_FALLBACK_ASSERT=1 -fsycl -fsycl-targets=%sycl_triple -DDEFINE_NDEBUG_INFILE2 -I %S/Inputs %S/assert_in_simultaneously_multiple_tus.cpp %S/Inputs/kernels_in_file2.cpp -o %t.out %threads_lib
 // RUN: %CPU_RUN_PLACEHOLDER %t.out &> %t.txt || true
 // RUN: %CPU_RUN_PLACEHOLDER FileCheck %s --input-file %t.txt
 //

SYCL/AtomicRef/atomic_memory_order_acq_rel.cpp

@@ -9,7 +9,7 @@
 
 // NOTE: Tests load and store for acquire-release memory ordering.
 
-#include "atomic_memory_order_acq_rel.h"
+#include "atomic_memory_order.h"
 #include <iostream>
 using namespace sycl;
 
@@ -24,21 +24,10 @@ int main() {
     return 0;
   }
 
-  constexpr int N = 32;
-
   // Acquire-release memory order must also support both acquire and release
   // orderings.
   assert(is_supported(supported_memory_orders, memory_order::acquire) &&
          is_supported(supported_memory_orders, memory_order::release));
-  acq_rel_test<int>(q, N);
-  acq_rel_test<unsigned int>(q, N);
-  acq_rel_test<float>(q, N);
-
-  // Include long tests if they are 32 bits wide
-  if constexpr (sizeof(long) == 4) {
-    acq_rel_test<long>(q, N);
-    acq_rel_test<unsigned long>(q, N);
-  }
 
   std::cout << "Test passed." << std::endl;
 }