[SYCL][E2E] Replace any-device-is- with target- features in %if markup (#16889)

ayylol · web-flow · commit 5b292cd43b0a · 2025-02-10T16:23:39.000Z
Using `target-` features is preferable for `%if` statements because this
feature will only evaluate to true if a device with the corresponding
build target is selected, or if we are in `build-only` and the
corresponding triple was selected to build for the test. `any-device-is`
on the other hand evaluates to true any time we have the device
available on the system, even if it is marked as XFAIL/UNSUPPORTED, and
on `build-only` it is never true because we have no devices in that
mode.
diff --git a/sycl/test-e2e/AtomicRef/atomic_memory_order_acq_rel.cpp b/sycl/test-e2e/AtomicRef/atomic_memory_order_acq_rel.cpp
@@ -1,4 +1,4 @@
-// RUN: %{build} -O3 -o %t.out %if any-device-is-cuda %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 %}
+// RUN: %{build} -O3 -o %t.out %if target-nvidia %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 %}
 // RUN: %{run} %t.out
 
 // NOTE: Tests fetch_add for acquire and release memory ordering.
diff --git a/sycl/test-e2e/AtomicRef/atomic_memory_order_seq_cst.cpp b/sycl/test-e2e/AtomicRef/atomic_memory_order_seq_cst.cpp
@@ -1,4 +1,4 @@
-// RUN: %{build} -O3 -o %t.out %if any-device-is-cuda %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 %}
+// RUN: %{build} -O3 -o %t.out %if target-nvidia %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 %}
 // RUN: %{run} %t.out
 // UNSUPPORTED: arch-intel_gpu_bmg_g21
 // UNSUPPORTED-TRACKER: https://github.com/intel/llvm/issues/16924
diff --git a/sycl/test-e2e/BFloat16/bfloat16_builtins.cpp b/sycl/test-e2e/BFloat16/bfloat16_builtins.cpp
@@ -5,11 +5,11 @@
 // + below sm_80 always uses generic impls
 
 // DEFINE: %{mathflags} = %if cl_options %{/clang:-fno-fast-math%} %else %{-fno-fast-math%}
-// RUN: %clangxx -fsycl %{sycl_target_opts} %if any-device-is-cuda %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_80 %} %s -o %t.out %{mathflags}
+// RUN: %clangxx -fsycl %{sycl_target_opts} %if target-nvidia %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_80 %} %s -o %t.out %{mathflags}
 // RUN: %{run} %t.out
 
 // Test "new" (ABI breaking) for all platforms ( sm_80/native if CUDA )
-// RUN:  %if preview-breaking-changes-supported %{  %clangxx -fsycl -fpreview-breaking-changes %{sycl_target_opts} %if any-device-is-cuda %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_80 %} %s -o %t2.out %{mathflags} %}
+// RUN:  %if preview-breaking-changes-supported %{  %clangxx -fsycl -fpreview-breaking-changes %{sycl_target_opts} %if target-nvidia %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_80 %} %s -o %t2.out %{mathflags} %}
 // RUN:  %if preview-breaking-changes-supported %{  %{run} %t2.out  %}
 
 // Flaky timeout on CPU. Enable when fixed.
diff --git a/sycl/test-e2e/BFloat16/bfloat16_builtins_cuda_generic.cpp b/sycl/test-e2e/BFloat16/bfloat16_builtins_cuda_generic.cpp
@@ -7,8 +7,8 @@
 // DEFINE: %{mathflags} = %if cl_options %{/clang:-fno-fast-math%} %else %{-fno-fast-math%}
 
 // If CUDA, test "new" again for sm_75/generic
-// RUN:  %if any-device-is-cuda %{ %if preview-breaking-changes-supported %{  %clangxx -fsycl -fpreview-breaking-changes %{sycl_target_opts}  -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_75  %s -o %t3.out %{mathflags} %} %}
-// RUN:  %if any-device-is-cuda %{ %if preview-breaking-changes-supported %{  %{run} %t3.out  %} %}
+// RUN:  %if target-nvidia %{ %if preview-breaking-changes-supported %{  %clangxx -fsycl -fpreview-breaking-changes %{sycl_target_opts}  -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_75  %s -o %t3.out %{mathflags} %} %}
+// RUN:  %if target-nvidia %{ %if preview-breaking-changes-supported %{  %{run} %t3.out  %} %}
 
 #include "bfloat16_builtins.hpp"
 
diff --git a/sycl/test-e2e/BFloat16/bfloat16_type.cpp b/sycl/test-e2e/BFloat16/bfloat16_type.cpp
@@ -1,4 +1,4 @@
-// RUN: %if any-device-is-cuda %{ %{build} -DUSE_CUDA_SM80=1 -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_80 -o %t.cuda.out %}
+// RUN: %if target-nvidia %{ %{build} -DUSE_CUDA_SM80=1 -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_80 -o %t.cuda.out %}
 // RUN: %if cuda %{ %{run} %t.cuda.out %}
 // RUN: %{build} -o %t.out
 // RUN: %{run} %t.out
diff --git a/sycl/test-e2e/Basic/interop/interop_all_backends.cpp b/sycl/test-e2e/Basic/interop/interop_all_backends.cpp
@@ -1,8 +1,8 @@
-// XFAIL: any-device-is-cuda
+// XFAIL: target-nvidia
 // XFAIL-TRACKER: https://github.com/intel/llvm/issues/16070
-// RUN: %if any-device-is-opencl %{ %{build} -o %t-opencl.out %}
-// RUN: %if any-device-is-cuda %{ %{build} -isystem %sycl_include -DBUILD_FOR_CUDA -o %t-cuda.out %}
-// RUN: %if any-device-is-hip %{ %{build} -DBUILD_FOR_HIP -o %t-hip.out %}
+// RUN: %if target-spir %{ %{build} -o %t-opencl.out %}
+// RUN: %if target-nvidia %{ %{build} -isystem %sycl_include -DBUILD_FOR_CUDA -o %t-cuda.out %}
+// RUN: %if target-amd %{ %{build} -DBUILD_FOR_HIP -o %t-hip.out %}
 
 #include <sycl/backend.hpp>
 #include <sycl/detail/core.hpp>
diff --git a/sycl/test-e2e/DeviceLib/cmath_test.cpp b/sycl/test-e2e/DeviceLib/cmath_test.cpp
@@ -7,7 +7,7 @@
 // RUN: %if !gpu %{ %{run} %t2.out %}
 //
 // // Check that --fast-math works with cmath funcs for CUDA
-// RUN: %if any-device-is-cuda %{ %{build} -Wno-nan-infinity-disabled -fno-builtin %{mathflags} -o %t3.out -ffast-math -DSYCL_E2E_FASTMATH %}
+// RUN: %if target-nvidia %{ %clangxx -fsycl -fsycl-targets=nvptx64-nvidia-cuda %s -Wno-nan-infinity-disabled -fno-builtin %{mathflags} -o %t3.out -ffast-math -DSYCL_E2E_FASTMATH %}
 // RUN: %if cuda %{ %{run} %t3.out %}
 
 #include "math_utils.hpp"
diff --git a/sycl/test-e2e/GroupAlgorithm/root_group.cpp b/sycl/test-e2e/GroupAlgorithm/root_group.cpp
@@ -5,7 +5,7 @@
 // TODO: Currently using the -Wno-deprecated-declarations flag due to issue
 // https://github.com/intel/llvm/issues/16451. Rewrite testRootGroup() amd
 // remove the flag once the issue is resolved.
-// RUN: %{build} -I . -o %t.out -Wno-deprecated-declarations %if any-device-is-cuda %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 %}
+// RUN: %{build} -I . -o %t.out -Wno-deprecated-declarations %if target-nvidia %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 %}
 // RUN: %{run} %t.out
 
 // Disabled temporarily while investigation into the failure is ongoing.
diff --git a/sycl/test-e2e/NewOffloadDriver/lit.local.cfg b/sycl/test-e2e/NewOffloadDriver/lit.local.cfg
@@ -3,5 +3,5 @@ import platform
 config.unsupported_features += ['accelerator']
 
 config.substitutions.append(
-    ("%{embed-ir}", "%if any-device-is-hip || any-device-is-cuda %{ -fsycl-embed-ir %}")
+    ("%{embed-ir}", "%if target-amd || target-nvidia %{ -fsycl-embed-ir %}")
 )
diff --git a/sycl/test-e2e/Reduction/reduction_range_1d_dw.cpp b/sycl/test-e2e/Reduction/reduction_range_1d_dw.cpp
@@ -1,4 +1,4 @@
-// RUN: %{build} -DENABLE_64_BIT=false -o %t.out %if any-device-is-cuda %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_60 %}
+// RUN: %{build} -DENABLE_64_BIT=false -o %t.out %if target-nvidia %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_60 %}
 // RUN: %{run} %t.out
 
 // Windows doesn't yet have full shutdown().
diff --git a/sycl/test-e2e/Reduction/reduction_range_1d_dw_64bit.cpp b/sycl/test-e2e/Reduction/reduction_range_1d_dw_64bit.cpp
@@ -1,4 +1,4 @@
-// RUN: %{build} -DENABLE_64_BIT=true -o %t.out %if any-device-is-cuda %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_60 %}
+// RUN: %{build} -DENABLE_64_BIT=true -o %t.out %if target-nvidia %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_60 %}
 // RUN: %{run} %t.out
 
 // Windows doesn't yet have full shutdown().
diff --git a/sycl/test-e2e/Reduction/reduction_range_1d_reducer_skip.cpp b/sycl/test-e2e/Reduction/reduction_range_1d_reducer_skip.cpp
@@ -1,4 +1,4 @@
-// RUN: %{build} -o %t.out %if any-device-is-cuda %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_60 %}
+// RUN: %{build} -o %t.out %if target-nvidia %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_60 %}
 // RUN: %{run} %t.out
 
 // Windows doesn't yet have full shutdown().
diff --git a/sycl/test-e2e/Reduction/reduction_range_1d_rw.cpp b/sycl/test-e2e/Reduction/reduction_range_1d_rw.cpp
@@ -1,4 +1,4 @@
-// RUN: %{build} -o %t.out %if any-device-is-cuda %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_60 %}
+// RUN: %{build} -o %t.out %if target-nvidia %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_60 %}
 // RUN: %{run} %t.out
 
 // Windows doesn't yet have full shutdown().
diff --git a/sycl/test-e2e/Reduction/reduction_range_2d_dw.cpp b/sycl/test-e2e/Reduction/reduction_range_2d_dw.cpp
@@ -1,4 +1,4 @@
-// RUN: %{build} -o %t.out %if any-device-is-cuda %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_60 %}
+// RUN: %{build} -o %t.out %if target-nvidia %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_60 %}
 // RUN: %{run} %t.out
 
 // Windows doesn't yet have full shutdown().
diff --git a/sycl/test-e2e/Reduction/reduction_range_2d_dw_reducer_skip.cpp b/sycl/test-e2e/Reduction/reduction_range_2d_dw_reducer_skip.cpp
@@ -1,4 +1,4 @@
-// RUN: %{build} -o %t.out %if any-device-is-cuda %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_60 %}
+// RUN: %{build} -o %t.out %if target-nvidia %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_60 %}
 // RUN: %{run} %t.out
 
 // Windows doesn't yet have full shutdown().
diff --git a/sycl/test-e2e/Reduction/reduction_range_2d_rw.cpp b/sycl/test-e2e/Reduction/reduction_range_2d_rw.cpp
@@ -1,4 +1,4 @@
-// RUN: %{build} -o %t.out %if any-device-is-cuda %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_60 %}
+// RUN: %{build} -o %t.out %if target-nvidia %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_60 %}
 // RUN: %{run} %t.out
 
 // Windows doesn't yet have full shutdown().
diff --git a/sycl/test-e2e/Reduction/reduction_range_3d_dw.cpp b/sycl/test-e2e/Reduction/reduction_range_3d_dw.cpp
@@ -1,4 +1,4 @@
-// RUN: %{build} -o %t.out %if any-device-is-cuda %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_60 %}
+// RUN: %{build} -o %t.out %if target-nvidia %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_60 %}
 // RUN: %{run} %t.out
 
 // Windows doesn't yet have full shutdown().
diff --git a/sycl/test-e2e/Reduction/reduction_range_3d_rw.cpp b/sycl/test-e2e/Reduction/reduction_range_3d_rw.cpp
@@ -1,4 +1,4 @@
-// RUN: %{build} -o %t.out %if any-device-is-cuda %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_60 %}
+// RUN: %{build} -o %t.out %if target-nvidia %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_60 %}
 // RUN: %{run} %t.out
 
 // Windows doesn't yet have full shutdown().
diff --git a/sycl/test-e2e/Reduction/reduction_range_3d_rw_reducer_skip.cpp b/sycl/test-e2e/Reduction/reduction_range_3d_rw_reducer_skip.cpp
@@ -1,4 +1,4 @@
-// RUN: %{build} -o %t.out %if any-device-is-cuda %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_60 %}
+// RUN: %{build} -o %t.out %if target-nvidia %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_60 %}
 // RUN: %{run} %t.out
 
 // Windows doesn't yet have full shutdown().
diff --git a/sycl/test-e2e/Reduction/reduction_range_usm_dw.cpp b/sycl/test-e2e/Reduction/reduction_range_usm_dw.cpp
@@ -1,4 +1,4 @@
-// RUN: %{build} -o %t.out %if any-device-is-cuda %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_60 %}
+// RUN: %{build} -o %t.out %if target-nvidia %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_60 %}
 // RUN: %{run} %t.out
 
 // Windows doesn't yet have full shutdown().
diff --git a/sycl/test-e2e/USM/P2P/p2p_atomics.cpp b/sycl/test-e2e/USM/P2P/p2p_atomics.cpp
@@ -1,5 +1,5 @@
 // REQUIRES: cuda || hip || level_zero
-// RUN:  %{build} %if any-device-is-cuda %{ -Xsycl-target-backend --cuda-gpu-arch=sm_61 %} -o %t.out
+// RUN:  %{build} %if target-nvidia %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_61 %} -o %t.out
 // RUN:  %{run} %t.out
 
 #include <cassert>
diff --git a/sycl/test-e2e/bindless_images/dx12_interop/read_write_unsampled.cpp b/sycl/test-e2e/bindless_images/dx12_interop/read_write_unsampled.cpp
@@ -2,7 +2,7 @@
 // REQUIRES: windows
 
 // DEFINE: %{link-flags}=%if cl_options %{ /clang:-ld3d12 /clang:-ldxgi /clang:-ldxguid %} %else %{ -ld3d12 -ldxgi -ldxguid %}
-// RUN: %{build} %{link-flags} -o %t.out %if any-device-is-level_zero %{ -DDISABLE_UNORM_TESTS %}
+// RUN: %{build} %{link-flags} -o %t.out %if target-spir %{ -DDISABLE_UNORM_TESTS %}
 // RUN: %{run-unfiltered-devices} env NEOReadDebugKeys=1 UseBindlessMode=1 UseExternalAllocatorForSshAndDsh=1 %t.out
 
 #pragma clang diagnostic ignored "-Waddress-of-temporary"
diff --git a/sycl/test-e2e/bindless_images/vulkan_interop/sampled_images.cpp b/sycl/test-e2e/bindless_images/vulkan_interop/sampled_images.cpp
@@ -1,7 +1,7 @@
 // REQUIRES: aspect-ext_oneapi_external_memory_import || (windows && level_zero && aspect-ext_oneapi_bindless_images)
 // REQUIRES: vulkan
 
-// RUN: %{build} %link-vulkan -o %t.out %if any-device-is-level_zero %{ -Wno-ignored-attributes -DENABLE_LINEAR_TILING -DTEST_L0_SUPPORTED_VK_FORMAT %}
+// RUN: %{build} %link-vulkan -o %t.out %if target-spir %{ -Wno-ignored-attributes -DENABLE_LINEAR_TILING -DTEST_L0_SUPPORTED_VK_FORMAT %}
 // RUN: %{run} env NEOReadDebugKeys=1 UseBindlessMode=1 UseExternalAllocatorForSshAndDsh=1 %t.out
 
 // Uncomment to print additional test information
diff --git a/sycl/test-e2e/bindless_images/vulkan_interop/unsampled_images.cpp b/sycl/test-e2e/bindless_images/vulkan_interop/unsampled_images.cpp
@@ -1,7 +1,7 @@
 // REQUIRES: aspect-ext_oneapi_external_memory_import || (windows && level_zero && aspect-ext_oneapi_bindless_images)
 // REQUIRES: vulkan
 
-// RUN: %{build} %link-vulkan -o %t.out %if any-device-is-level_zero %{ -Wno-ignored-attributes -DTEST_L0_SUPPORTED_VK_FORMAT %}
+// RUN: %{build} %link-vulkan -o %t.out %if target-spir %{ -Wno-ignored-attributes -DTEST_L0_SUPPORTED_VK_FORMAT %}
 // RUN: %{run} env NEOReadDebugKeys=1 UseBindlessMode=1 UseExternalAllocatorForSshAndDsh=1 %t.out
 
 // Uncomment to print additional test information
diff --git a/sycl/test-e2e/syclcompat/atomic/atomic_class.cpp b/sycl/test-e2e/syclcompat/atomic/atomic_class.cpp
@@ -32,7 +32,7 @@
 
 // UNSUPPORTED: hip || (windows && level_zero)
 
-// RUN: %{build} %if any-device-is-cuda %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 %} -o %t.out
+// RUN: %{build} %if target-nvidia %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 %} -o %t.out
 // RUN: %{run} %t.out
 
 #include <sycl/detail/core.hpp>
diff --git a/sycl/test-e2e/syclcompat/atomic/atomic_memory_acq_rel.cpp b/sycl/test-e2e/syclcompat/atomic/atomic_memory_acq_rel.cpp
@@ -32,7 +32,7 @@
 
 // UNSUPPORTED: hip
 
-// RUN: %{build} %if any-device-is-cuda %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 %} -o %t.out
+// RUN: %{build} %if target-nvidia %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 %} -o %t.out
 // RUN: %{run} %t.out
 
 #include <iostream>
diff --git a/sycl/test-e2e/syclcompat/math/math_byte_dot_product.cpp b/sycl/test-e2e/syclcompat/math/math_byte_dot_product.cpp
@@ -29,7 +29,7 @@
 //
 // ===---------------------------------------------------------------------===//
 
-// RUN: %{build} %if any-device-is-cuda %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_61 %} -o %t.out
+// RUN: %{build} %if target-nvidia %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_61 %} -o %t.out
 // RUN: %{run} %t.out
 
 #include <sycl/detail/core.hpp>

Original file line number	Diff line number	Diff line change
`@@ -1,4 +1,4 @@`
`1`		`-// RUN: %{build} -O3 -o %t.out %if any-device-is-cuda %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 %}`
	`1`	`+// RUN: %{build} -O3 -o %t.out %if target-nvidia %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_70 %}`
`2`	`2`	`// RUN: %{run} %t.out`
`3`	`3`
`4`	`4`	`// NOTE: Tests fetch_add for acquire and release memory ordering.`
Original file line number	Diff line number	Diff line change
`@@ -1,4 +1,4 @@`
`1`		`-// RUN: %if any-device-is-cuda %{ %{build} -DUSE_CUDA_SM80=1 -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_80 -o %t.cuda.out %}`
	`1`	`+// RUN: %if target-nvidia %{ %{build} -DUSE_CUDA_SM80=1 -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_80 -o %t.cuda.out %}`
`2`	`2`	`// RUN: %if cuda %{ %{run} %t.cuda.out %}`
`3`	`3`	`// RUN: %{build} -o %t.out`
`4`	`4`	`// RUN: %{run} %t.out`
Original file line number	Diff line number	Diff line change
`@@ -3,5 +3,5 @@ import platform`
`3`	`3`	`config.unsupported_features += ['accelerator']`
`4`	`4`
`5`	`5`	`config.substitutions.append(`
`6`		`- ("%{embed-ir}", "%if any-device-is-hip \|\| any-device-is-cuda %{ -fsycl-embed-ir %}")`
	`6`	`+ ("%{embed-ir}", "%if target-amd \|\| target-nvidia %{ -fsycl-embed-ir %}")`
`7`	`7`	`)`
Original file line number	Diff line number	Diff line change
`@@ -1,4 +1,4 @@`
`1`		`-// RUN: %{build} -DENABLE_64_BIT=false -o %t.out %if any-device-is-cuda %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_60 %}`
	`1`	`+// RUN: %{build} -DENABLE_64_BIT=false -o %t.out %if target-nvidia %{ -Xsycl-target-backend=nvptx64-nvidia-cuda --cuda-gpu-arch=sm_60 %}`
`2`	`2`	`// RUN: %{run} %t.out`
`3`	`3`
`4`	`4`	`// Windows doesn't yet have full shutdown().`