[SYCL] Remove host run and dependencies from SYCL/Reduction tests (intel#1216)

This commit removes the host run and any assumptions and operations
related to the host device from the tests in SYCL/Reduction.

Signed-off-by: Larsen, Steffen <steffen.larsen@intel.com>
Co-authored-by: Sachkov, Alexey <alexey.sachkov@intel.com>

Author: 2 people

SYCL/Reduction/reduction_big_data.cpp

@@ -3,14 +3,10 @@
 // RUN: %ACC_RUN_PLACEHOLDER %t.out
 // RUN: %CPU_RUN_PLACEHOLDER %t.out
 //
-// `Group algorithms are not supported on host device` on Nvidia.
+// Group algorithms are not supported on Nvidia.
 // XFAIL: hip_nvidia
 //
 
-// RUNx: %HOST_RUN_PLACEHOLDER %t.out
-// TODO: Enable the test for HOST when it supports ext::oneapi::reduce() and
-// barrier()
-
 // This test performs basic checks of parallel_for(nd_range, reduction, func)
 // where the bigger data size and/or non-uniform work-group sizes may cause
 // errors.

SYCL/Reduction/reduction_nd_N_queue_shortcut.cpp

@@ -3,13 +3,9 @@
 // RUN: %ACC_RUN_PLACEHOLDER %t.out
 // RUN: %CPU_RUN_PLACEHOLDER %t.out
 
-// `Group algorithms are not supported on host device.` on NVidia.
+// Group algorithms are not supported on NVidia.
 // XFAIL: hip_nvidia
 
-// RUNx: %HOST_RUN_PLACEHOLDER %t.out
-// TODO: Enable the test for HOST when it supports ext::oneapi::reduce() and
-// barrier()
-
 // This test only checks that the method queue::parallel_for() accepting
 // reduction, can be properly translated into queue::submit + parallel_for().
 

SYCL/Reduction/reduction_nd_conditional.cpp

@@ -1,5 +1,4 @@
 // RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple %s -o %t.out
-// RUNx: %HOST_RUN_PLACEHOLDER %t.out
 // RUN: %CPU_RUN_PLACEHOLDER %t.out
 // RUN: %GPU_RUN_PLACEHOLDER %t.out
 // RUN: %ACC_RUN_PLACEHOLDER %t.out

SYCL/Reduction/reduction_nd_dw.cpp

@@ -1,10 +1,9 @@
 // RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple %s -o %t.out
-// RUNx: %HOST_RUN_PLACEHOLDER %t.out
 // RUN: %CPU_RUN_PLACEHOLDER %t.out
 // RUN: %GPU_RUN_PLACEHOLDER %t.out
 // RUN: %ACC_RUN_PLACEHOLDER %t.out
 //
-// `Group algorithms are not supported on host device.` on Nvidia.
+// Group algorithms are not supported on Nvidia.
 // XFAIL: hip_nvidia
 
 // This test performs basic checks of parallel_for(nd_range, reduction, func)

SYCL/Reduction/reduction_nd_ext_double.cpp

@@ -9,9 +9,6 @@
 
 // XFAIL: hip_nvidia
 
-// TODO: Enable the test for HOST when it supports intel::reduce() and barrier()
-// RUNx: %HOST_RUN_PLACEHOLDER %t.out
-
 // This test performs basic checks of parallel_for(nd_range, reduction, func)
 // used with 'double' type.
 

SYCL/Reduction/reduction_nd_ext_half.cpp

@@ -8,9 +8,6 @@
 // work group size not bigger than 1` on Nvidia.
 // XFAIL: hip_amd || hip_nvidia
 
-// TODO: Enable the test for HOST when it supports intel::reduce() and barrier()
-// RUNx: %HOST_RUN_PLACEHOLDER %t.out
-
 // This test performs basic checks of parallel_for(nd_range, reduction, func)
 // used with 'half' type.
 

SYCL/Reduction/reduction_nd_ext_type.hpp

@@ -20,7 +20,7 @@ template <typename T> int runTests(sycl::aspect ExtAspect) {
   queue Q;
   printDeviceInfo(Q);
   device D = Q.get_device();
-  if (!D.is_host() && !D.has(ExtAspect)) {
+  if (!D.has(ExtAspect)) {
     std::cout << "Test skipped\n";
     return 0;
   }

SYCL/Reduction/reduction_nd_lambda.cpp

@@ -1,11 +1,9 @@
 // RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple %s -o %t.out
-// RUNx: %HOST_RUN_PLACEHOLDER %t.out
 // RUN: %CPU_RUN_PLACEHOLDER %t.out
 // RUN: %GPU_RUN_PLACEHOLDER %t.out
 // RUN: %ACC_RUN_PLACEHOLDER %t.out
 //
-// Inconsistently fails on HIP AMD, error message `Barrier is not supported on
-// the host device yet.` on HIP Nvidia.
+// Inconsistently fails on HIP AMD, HIP Nvidia.
 // UNSUPPORTED: hip_amd || hip_nvidia
 
 // This test performs basic checks of parallel_for(nd_range, reduction, lambda)

SYCL/Reduction/reduction_nd_queue_shortcut.cpp

@@ -3,13 +3,9 @@
 // RUN: %ACC_RUN_PLACEHOLDER %t.out
 // RUN: %CPU_RUN_PLACEHOLDER %t.out
 
-// `Group algorithms are not supported on host device.` on NVidia.
+// Group algorithms are not supported on NVidia.
 // XFAIL: hip_nvidia
 
-// RUNx: %HOST_RUN_PLACEHOLDER %t.out
-// TODO: Enable the test for HOST when it supports ext::oneapi::reduce() and
-// barrier()
-
 // This test only checks that the method queue::parallel_for() accepting
 // reduction, can be properly translated into queue::submit + parallel_for().
 

SYCL/Reduction/reduction_range_queue_shortcut.cpp

@@ -3,13 +3,9 @@
 // RUN: %ACC_RUN_PLACEHOLDER %t.out
 // RUN: %CPU_RUN_PLACEHOLDER %t.out
 
-// `Group algorithms are not supported on host device.` on NVidia.
+// Group algorithms are not supported on NVidia.
 // XFAIL: hip_nvidia
 
-// RUNx: %HOST_RUN_PLACEHOLDER %t.out
-// TODO: Enable the test for HOST when it supports ext::oneapi::reduce() and
-// barrier()
-
 // This test only checks that the shortcut method queue::parallel_for()
 // can accept 2 or more reduction variables.