[SYCL] Enable LIT testing with CUDA BE

vladimirlaz · sys_bbsycl · commit ea89d42d4af4 · 2020-04-03T21:26:26.000+03:00
Signed-off-by: Vladimir Lazarev &lt;vladimir.lazarev@intel.com&gt;
diff --git a/sycl/test/backend/cuda/primary_context.cpp b/sycl/test/backend/cuda/primary_context.cpp
diff --git a/sycl/test/basic_tests/boolean.cpp b/sycl/test/basic_tests/boolean.cpp
@@ -3,8 +3,6 @@
 // RUN: %CPU_RUN_PLACEHOLDER %t.out
 // RUN: %GPU_RUN_PLACEHOLDER %t.out
 // RUN: %ACC_RUN_PLACEHOLDER %t.out
-// XFAIL: cuda
-// TODO: investigate incorrect results on cuda backend
 #include <CL/sycl.hpp>
 
 #include <cassert>
diff --git a/sycl/test/basic_tests/buffer/buffer_full_copy.cpp b/sycl/test/basic_tests/buffer/buffer_full_copy.cpp
@@ -6,9 +6,6 @@
 // RUN: %GPU_RUN_PLACEHOLDER %t2.out
 // RUN: %ACC_RUN_PLACEHOLDER %t2.out
 
-// TODO: cuda_piEnqueueMemBufferCopy not implemented
-// XFAIL: cuda
-
 //==------------- buffer_full_copy.cpp - SYCL buffer basic test ------------==//
 //
 // Part of the LLVM Project, under the Apache License v2.0 with LLVM Exceptions.
diff --git a/sycl/test/group-algorithm/all_of.cpp b/sycl/test/group-algorithm/all_of.cpp
@@ -1,8 +1,9 @@
-// RUN: %clangxx -fsycl %s -o %t.out
+// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple %s -o %t.out
 // RUN: env SYCL_DEVICE_TYPE=HOST %t.out
 // RUN: %CPU_RUN_PLACEHOLDER %t.out
 // RUN: %GPU_RUN_PLACEHOLDER %t.out
 // RUN: %ACC_RUN_PLACEHOLDER %t.out
+// UNSUPPORTED: cuda
 
 #include <CL/sycl.hpp>
 #include <algorithm>
diff --git a/sycl/test/group-algorithm/any_of.cpp b/sycl/test/group-algorithm/any_of.cpp
@@ -1,8 +1,9 @@
-// RUN: %clangxx -fsycl %s -o %t.out
+// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple %s -o %t.out
 // RUN: env SYCL_DEVICE_TYPE=HOST %t.out
 // RUN: %CPU_RUN_PLACEHOLDER %t.out
 // RUN: %GPU_RUN_PLACEHOLDER %t.out
 // RUN: %ACC_RUN_PLACEHOLDER %t.out
+// UNSUPPORTED: cuda
 
 #include <CL/sycl.hpp>
 #include <algorithm>
diff --git a/sycl/test/group-algorithm/broadcast.cpp b/sycl/test/group-algorithm/broadcast.cpp
@@ -1,8 +1,9 @@
-// RUN: %clangxx -fsycl %s -o %t.out
+// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple %s -o %t.out
 // RUN: env SYCL_DEVICE_TYPE=HOST %t.out
 // RUN: %CPU_RUN_PLACEHOLDER %t.out
 // RUN: %GPU_RUN_PLACEHOLDER %t.out
 // RUN: %ACC_RUN_PLACEHOLDER %t.out
+// UNSUPPORTED: cuda
 
 #include <CL/sycl.hpp>
 #include <algorithm>
diff --git a/sycl/test/group-algorithm/exclusive_scan.cpp b/sycl/test/group-algorithm/exclusive_scan.cpp
@@ -1,8 +1,9 @@
-// RUN: %clangxx -fsycl %s -o %t.out
+// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple %s -o %t.out
 // RUN: env SYCL_DEVICE_TYPE=HOST %t.out
 // RUN: %CPU_RUN_PLACEHOLDER %t.out
 // RUN: %GPU_RUN_PLACEHOLDER %t.out
 // RUN: %ACC_RUN_PLACEHOLDER %t.out
+// UNSUPPORTED: cuda
 
 #include <CL/sycl.hpp>
 #include <algorithm>
diff --git a/sycl/test/group-algorithm/inclusive_scan.cpp b/sycl/test/group-algorithm/inclusive_scan.cpp
@@ -1,8 +1,9 @@
-// RUN: %clangxx -fsycl %s -o %t.out
+// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple %s -o %t.out
 // RUN: env SYCL_DEVICE_TYPE=HOST %t.out
 // RUN: %CPU_RUN_PLACEHOLDER %t.out
 // RUN: %GPU_RUN_PLACEHOLDER %t.out
 // RUN: %ACC_RUN_PLACEHOLDER %t.out
+// UNSUPPORTED: cuda
 
 #include <CL/sycl.hpp>
 #include <algorithm>
diff --git a/sycl/test/group-algorithm/leader.cpp b/sycl/test/group-algorithm/leader.cpp
@@ -1,4 +1,4 @@
-// RUN: %clangxx -fsycl %s -o %t.out
+// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple %s -o %t.out
 // RUN: env SYCL_DEVICE_TYPE=HOST %t.out
 // RUN: %CPU_RUN_PLACEHOLDER %t.out
 // RUN: %GPU_RUN_PLACEHOLDER %t.out
diff --git a/sycl/test/group-algorithm/none_of.cpp b/sycl/test/group-algorithm/none_of.cpp
@@ -1,8 +1,9 @@
-// RUN: %clangxx -fsycl %s -o %t.out
+// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple %s -o %t.out
 // RUN: env SYCL_DEVICE_TYPE=HOST %t.out
 // RUN: %CPU_RUN_PLACEHOLDER %t.out
 // RUN: %GPU_RUN_PLACEHOLDER %t.out
 // RUN: %ACC_RUN_PLACEHOLDER %t.out
+// UNSUPPORTED: cuda
 
 #include <CL/sycl.hpp>
 #include <algorithm>
diff --git a/sycl/test/group-algorithm/reduce.cpp b/sycl/test/group-algorithm/reduce.cpp
@@ -1,8 +1,9 @@
-// RUN: %clangxx -fsycl %s -o %t.out
+// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple %s -o %t.out
 // RUN: env SYCL_DEVICE_TYPE=HOST %t.out
 // RUN: %CPU_RUN_PLACEHOLDER %t.out
 // RUN: %GPU_RUN_PLACEHOLDER %t.out
 // RUN: %ACC_RUN_PLACEHOLDER %t.out
+// UNSUPPORTED: cuda
 
 #include <CL/sycl.hpp>
 #include <algorithm>
diff --git a/sycl/test/hier_par/hier_par_basic.cpp b/sycl/test/hier_par/hier_par_basic.cpp
@@ -12,9 +12,6 @@
 // RUN: %GPU_RUN_PLACEHOLDER %t.out
 // RUN: %ACC_RUN_PLACEHOLDER %t.out
 
-// TODO: ptxas fatal   : Unresolved extern function '__spirv_ControlBarrier'
-// XFAIL: cuda
-
 // This test checks hierarchical parallelism invocation APIs, but without any
 // data or code with side-effects between the work group and work item scopes.
 
diff --git a/sycl/test/kernel_from_file/hw.cpp b/sycl/test/kernel_from_file/hw.cpp
@@ -6,9 +6,6 @@
 
 // TODO: InvalidTargetTriple: Expects spir-unknown-unknown or spir64-unknown-unknown. Actual target triple is x86_64-unknown-linux-gnu
 
-// XFAIL: cuda
-// Currently unsupported on cuda as this test specifically tests a SPV path.
-
 #include <CL/sycl.hpp>
 #include <iostream>
 
diff --git a/sycl/test/lit.cfg.py b/sycl/test/lit.cfg.py
@@ -155,6 +155,7 @@ def getDeviceCount(device_type):
     config.available_features.add('gpu')
     if cuda:
        config.available_features.add('cuda')
+       gpu_run_substitute += " SYCL_BE=PI_CUDA "
 
     if platform.system() == "Linux":
         gpu_run_on_linux_substitute = "env SYCL_DEVICE_TYPE=GPU "
diff --git a/sycl/test/multi_ptr/multi_ptr.cpp b/sycl/test/multi_ptr/multi_ptr.cpp
@@ -3,7 +3,7 @@
 // RUN: %CPU_RUN_PLACEHOLDER %t.out
 // RUN: %GPU_RUN_PLACEHOLDER %t.out
 // RUN: %ACC_RUN_PLACEHOLDER %t.out
-// RUN: %clangxx -DRESTRICT_WRITE_ACCESS_TO_CONSTANT_PTR -fsycl %s -o %t1.out
+// RUN: %clangxx -DRESTRICT_WRITE_ACCESS_TO_CONSTANT_PTR -fsycl -fsycl-targets=%sycl_triple %s -o %t1.out
 // RUN: env SYCL_DEVICE_TYPE=HOST %t1.out
 // RUN: %CPU_RUN_PLACEHOLDER %t1.out
 // RUN: %GPU_RUN_PLACEHOLDER %t1.out
diff --git a/sycl/test/program_manager/env_vars.cpp b/sycl/test/program_manager/env_vars.cpp
@@ -1,4 +1,4 @@
-// RUN: %clangxx -O0 -fsycl %s -o %t.out -lsycl
+// RUN: %clangxx -O0 -fsycl -fsycl-targets=%sycl_triple %s -o %t.out -lsycl
 //
 // Deprecated SYCL_PROGRAM_BUILD_OPTIONS should work as an alias to
 // SYCL_PROGRAM_COMPILE_OPTIONS:
@@ -16,7 +16,6 @@
 // RUN: %CPU_RUN_PLACEHOLDER SYCL_PROGRAM_COMPILE_OPTIONS="-enable-link-options -cl-denorms-are-zero" SHOULD_CRASH=1 %t.out
 // RUN: %CPU_RUN_PLACEHOLDER SYCL_PROGRAM_LINK_OPTIONS="-g" SHOULD_CRASH=1 %t.out
 
-
 #include <CL/sycl.hpp>
 #include <cassert>
 #include <memory>
diff --git a/sycl/test/regression/private_array_init_test.cpp b/sycl/test/regression/private_array_init_test.cpp
@@ -1,4 +1,4 @@
-// RUN: %clangxx -fsycl %s -o %t.out -lOpenCL
+// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple %s -o %t.out -lOpenCL
 
 // RUN: %CPU_RUN_PLACEHOLDER %t.out
 // RUN: %GPU_RUN_PLACEHOLDER %t.out
diff --git a/sycl/test/regression/static-buffer-dtor.cpp b/sycl/test/regression/static-buffer-dtor.cpp
@@ -16,7 +16,6 @@
 
 // TODO: terminate called after throwing an instance of 'cl::sycl::runtime_error'
 // TODO: what():  OpenCL API failed. OpenCL API returns: -999 (Unknown OpenCL error code) -999 (Unknown OpenCL error code)
-// XFAIL: cuda
 
 #include <CL/sycl.hpp>
 
diff --git a/sycl/test/scheduler/DataMovement.cpp b/sycl/test/scheduler/DataMovement.cpp
@@ -1,5 +1,9 @@
 // RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple -I %sycl_source_dir %s -o %t.out
-// RUN: %t.out
+// RUN: %CPU_RUN_PLACEHOLDER %t.out
+// RUN: %GPU_RUN_PLACEHOLDER %t.out
+// RUN: %ACC_RUN_PLACEHOLDER %t.out
+// Incorrect event callback processing for host device.
+// XFAIL: cuda
 //
 //==-------------------------- DataMovement.cpp ----------------------------==//
 //
diff --git a/sycl/test/usm/allocator_vector.cpp b/sycl/test/usm/allocator_vector.cpp
@@ -2,6 +2,7 @@
 // RUN: env SYCL_DEVICE_TYPE=HOST %t1.out
 // RUN: %CPU_RUN_PLACEHOLDER %t1.out
 // RUN: %GPU_RUN_PLACEHOLDER %t1.out
+// UNSUPPORTED: cuda
 
 //==---- allocator_vector.cpp - Allocator Container test -------------------==//
 //
diff --git a/sycl/test/usm/allocator_vector_fail.cpp b/sycl/test/usm/allocator_vector_fail.cpp
@@ -2,6 +2,7 @@
 // RUN: env SYCL_DEVICE_TYPE=HOST %t1.out
 // RUN: %CPU_RUN_PLACEHOLDER %t1.out
 // RUN: %GPU_RUN_PLACEHOLDER %t1.out
+// UNSUPPORTED: cuda
 
 //==-- allocator_vector_fail.cpp - Device Memory Allocator fail test -------==//
 //
diff --git a/sycl/test/usm/allocatorll.cpp b/sycl/test/usm/allocatorll.cpp
@@ -2,6 +2,7 @@
 // RUN: env SYCL_DEVICE_TYPE=HOST %t1.out
 // RUN: %CPU_RUN_PLACEHOLDER %t1.out
 // RUN: %GPU_RUN_PLACEHOLDER %t1.out
+// UNSUPPORTED: cuda
 
 //==---- allocatorll.cpp - Device Memory Linked List Allocator test --------==//
 //
diff --git a/sycl/test/usm/badmalloc.cpp b/sycl/test/usm/badmalloc.cpp
@@ -1,9 +1,9 @@
-// RUN: %clangxx -fsycl %s -o %t1.out
+// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple %s -o %t1.out
 // RUN: env SYCL_DEVICE_TYPE=HOST %t1.out
 // RUN: %CPU_RUN_PLACEHOLDER %t1.out
 // RUN: %GPU_RUN_PLACEHOLDER %t1.out
 
-// UNSUPPORTED: windows
+// UNSUPPORTED: windows,cuda
 
 //==----------------- badmalloc.cpp - Bad Mallocs test ---------------------==//
 //
diff --git a/sycl/test/usm/depends_on.cpp b/sycl/test/usm/depends_on.cpp
@@ -2,6 +2,7 @@
 // RUN: env SYCL_DEVICE_TYPE=HOST %t1.out
 // RUN: %CPU_RUN_PLACEHOLDER %t1.out
 // RUN: %GPU_RUN_PLACEHOLDER %t1.out
+// UNSUPPORTED: cuda
 
 //==----------------- depends_on.cpp - depends_on test ---------------------==//
 //
diff --git a/sycl/test/usm/dmemll.cpp b/sycl/test/usm/dmemll.cpp
@@ -2,6 +2,7 @@
 // RUN: env SYCL_DEVICE_TYPE=HOST %t1.out
 // RUN: %CPU_RUN_PLACEHOLDER %t1.out
 // RUN: %GPU_RUN_PLACEHOLDER %t1.out
+// UNSUPPORTED: cuda
 
 //==------------------- dmemll.cpp - Device Memory Linked List test --------==//
 //
diff --git a/sycl/test/usm/hmemll.cpp b/sycl/test/usm/hmemll.cpp
@@ -2,6 +2,7 @@
 // RUN: env SYCL_DEVICE_TYPE=HOST %t1.out
 // RUN: %CPU_RUN_PLACEHOLDER %t1.out
 // RUN: %GPU_RUN_PLACEHOLDER %t1.out
+// UNSUPPORTED: cuda
 
 //==------------------- hmemll.cpp - Host Memory Linked List test ----------==//
 //
diff --git a/sycl/test/usm/math.cpp b/sycl/test/usm/math.cpp
@@ -3,8 +3,6 @@
 // RUN: %CPU_RUN_PLACEHOLDER %t.out
 
 // REQUIRES: cpu
-// TODO: ptxas fatal   : Unresolved extern function '_Z20__spirv_ocl_lgamma_rfPi'
-// XFAIL: cuda
 
 #include <CL/sycl.hpp>
 
diff --git a/sycl/test/usm/memadvise.cpp b/sycl/test/usm/memadvise.cpp
@@ -2,6 +2,7 @@
 // RUN: env SYCL_DEVICE_TYPE=HOST %t1.out
 // RUN: %CPU_RUN_PLACEHOLDER %t1.out
 // RUN: %GPU_RUN_PLACEHOLDER %t1.out
+// UNSUPPORTED: cuda
 
 //==---------------- memadvise.cpp - Shared Memory Linked List test --------==//
 //
diff --git a/sycl/test/usm/memcpy.cpp b/sycl/test/usm/memcpy.cpp
@@ -5,9 +5,10 @@
 // SPDX-License-Identifier: Apache-2.0 WITH LLVM-exception
 //
 //===----------------------------------------------------------------------===//
-// RUN: %clangxx -fsycl %s -o %t1.out
+// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple  %s -o %t1.out
 // RUN: %CPU_RUN_PLACEHOLDER %t1.out
 // RUN: %GPU_RUN_PLACEHOLDER %t1.out
+// UNSUPPORTED: cuda
 
 #include <CL/sycl.hpp>
 
diff --git a/sycl/test/usm/memset.cpp b/sycl/test/usm/memset.cpp
@@ -1,6 +1,7 @@
-// RUN: %clangxx -fsycl %s -o %t1.out
+// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple  %s -o %t1.out
 // RUN: %CPU_RUN_PLACEHOLDER %t1.out
 // RUN: %GPU_RUN_PLACEHOLDER %t1.out
+// UNSUPPORTED: cuda
 
 //==---- memset.cpp - USM memset test --------------------------------------==//
 //
diff --git a/sycl/test/usm/mixed.cpp b/sycl/test/usm/mixed.cpp
diff --git a/sycl/test/usm/mixed2.cpp b/sycl/test/usm/mixed2.cpp
diff --git a/sycl/test/usm/mixed2template.cpp b/sycl/test/usm/mixed2template.cpp
diff --git a/sycl/test/usm/mixed_queue.cpp b/sycl/test/usm/mixed_queue.cpp
diff --git a/sycl/test/usm/pfor_flatten.cpp b/sycl/test/usm/pfor_flatten.cpp
diff --git a/sycl/test/usm/queue_wait.cpp b/sycl/test/usm/queue_wait.cpp
diff --git a/sycl/test/usm/smemll.cpp b/sycl/test/usm/smemll.cpp
diff --git a/sycl/tools/CMakeLists.txt b/sycl/tools/CMakeLists.txt
diff --git a/sycl/tools/get_device_count_by_type.cpp b/sycl/tools/get_device_count_by_type.cpp
diff --git a/sycl/unittests/pi/cuda/CMakeLists.txt b/sycl/unittests/pi/cuda/CMakeLists.txt
diff --git a/sycl/unittests/pi/cuda/test_primary_context.cpp b/sycl/unittests/pi/cuda/test_primary_context.cpp

Original file line number	Diff line number	Diff line change
`@@ -1,4 +1,4 @@`
`1`		`-// RUN: %clangxx -fsycl %s -o %t.out`
	`1`	`+// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple %s -o %t.out`
`2`	`2`	`// RUN: env SYCL_DEVICE_TYPE=HOST %t.out`
`3`	`3`	`// RUN: %CPU_RUN_PLACEHOLDER %t.out`
`4`	`4`	`// RUN: %GPU_RUN_PLACEHOLDER %t.out`
Original file line number	Diff line number	Diff line change
`@@ -1,4 +1,4 @@`
`1`		`-// RUN: %clangxx -fsycl %s -o %t.out -lOpenCL`
	`1`	`+// RUN: %clangxx -fsycl -fsycl-targets=%sycl_triple %s -o %t.out -lOpenCL`
`2`	`2`
`3`	`3`	`// RUN: %CPU_RUN_PLACEHOLDER %t.out`
`4`	`4`	`// RUN: %GPU_RUN_PLACEHOLDER %t.out`