[RISCV][FMV] Support target_clones (#85786)
[llvm-project.git] / clang / test / Driver / openmp-offload-gpu.c
blobef6cbdded6a6f21653433f4bb78150ffb9042d2d
1 ///
2 /// Perform several driver tests for OpenMP offloading
3 ///
5 /// ###########################################################################
7 /// Check -Xopenmp-target uses one of the archs provided when several archs are used.
8 // RUN: %clang -### -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda -nogpulib -nogpuinc \
9 // RUN: -Xopenmp-target -march=sm_52 -Xopenmp-target -march=sm_60 %s 2>&1 \
10 // RUN: | FileCheck -check-prefix=CHK-FOPENMP-TARGET-ARCHS %s
12 // CHK-FOPENMP-TARGET-ARCHS: ptxas{{.*}}" "--gpu-name" "sm_60"
14 /// ###########################################################################
16 /// Check -Xopenmp-target -march=sm_52 works as expected when two triples are present.
17 // RUN: %clang -### -fopenmp=libomp \
18 // RUN: -fopenmp-targets=powerpc64le-ibm-linux-gnu,nvptx64-nvidia-cuda \
19 // RUN: -nogpulib -nogpuinc -Xopenmp-target=nvptx64-nvidia-cuda -march=sm_52 %s 2>&1 \
20 // RUN: | FileCheck -check-prefix=CHK-FOPENMP-TARGET-COMPILATION %s
22 // CHK-FOPENMP-TARGET-COMPILATION: ptxas{{.*}}" "--gpu-name" "sm_52"
24 /// Check PTXAS is passed -c flag when offloading to an NVIDIA device using OpenMP.
25 // RUN: %clang -### -nogpulib -nogpuinc -fopenmp=libomp -Xopenmp-target=nvptx64-nvidia-cuda -march=sm_52 \
26 // RUN: -fopenmp-targets=nvptx64-nvidia-cuda %s 2>&1 | FileCheck -check-prefix=CHK-PTXAS-DEFAULT %s
28 // CHK-PTXAS-DEFAULT: ptxas{{.*}}" "-c"
30 /// ###########################################################################
32 /// PTXAS is passed -c flag by default when offloading to an NVIDIA device using OpenMP - disable it.
33 // RUN: %clang -### -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda -fnoopenmp-relocatable-target \
34 // RUN: -nogpulib -nogpuinc --offload-arch=sm_52 -save-temps %s 2>&1 \
35 // RUN: | FileCheck -check-prefix=CHK-PTXAS-NORELO %s
37 // CHK-PTXAS-NORELO-NOT: ptxas{{.*}}" "-c"
39 /// ###########################################################################
41 /// PTXAS is passed -c flag by default when offloading to an NVIDIA device using OpenMP
42 /// Check that the flag is passed when -fopenmp-relocatable-target is used.
43 // RUN: %clang -### -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda -fopenmp-relocatable-target \
44 // RUN: -nogpulib -nogpuinc --offload-arch=sm_52 -save-temps %s 2>&1 \
45 // RUN: | FileCheck -check-prefix=CHK-PTXAS-RELO %s
47 // CHK-PTXAS-RELO: ptxas{{.*}}" "-c"
49 /// ###########################################################################
51 /// Check that error is not thrown by toolchain when no cuda lib flag is used.
52 /// Check that the flag is passed when -fopenmp-relocatable-target is used.
53 // RUN: %clang -### -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda -Xopenmp-target -march=sm_60 \
54 // RUN: -nogpulib -nogpuinc -fopenmp-relocatable-target -save-temps %s 2>&1 \
55 // RUN: | FileCheck -check-prefix=CHK-FLAG-NOLIBDEVICE %s
57 // CHK-FLAG-NOLIBDEVICE-NOT: error:{{.*}}sm_60
59 /// ###########################################################################
61 /// Check that error is not thrown by toolchain when no cuda lib device is found when using -S.
62 /// Check that the flag is passed when -fopenmp-relocatable-target is used.
63 // RUN: %clang -### -S -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda -Xopenmp-target -march=sm_60 \
64 // RUN: -fopenmp-relocatable-target -save-temps %s 2>&1 \
65 // RUN: | FileCheck -check-prefix=CHK-NOLIBDEVICE %s
67 // CHK-NOLIBDEVICE-NOT: error:{{.*}}sm_60
69 /// ###########################################################################
71 /// Check that the runtime bitcode library is part of the compile line.
72 /// Create a bogus bitcode library and specify it with libomptarget-nvptx-bc-path
73 // RUN: %clang -### -no-canonical-prefixes -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda \
74 // RUN: --libomptarget-nvptx-bc-path=%S/Inputs/libomptarget/libomptarget-nvptx-test.bc \
75 // RUN: -Xopenmp-target -march=sm_52 --cuda-path=%S/Inputs/CUDA_102/usr/local/cuda \
76 // RUN: -fopenmp-relocatable-target -save-temps %s 2>&1 \
77 // RUN: | FileCheck -check-prefix=CHK-BCLIB %s
79 /// Specify the directory containing the bitcode lib, check clang picks the right one
80 // RUN: %clang -### -no-canonical-prefixes -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda \
81 // RUN: --libomptarget-nvptx-bc-path=%S/Inputs/libomptarget \
82 // RUN: -Xopenmp-target -march=sm_52 --cuda-path=%S/Inputs/CUDA_102/usr/local/cuda \
83 // RUN: -fopenmp-relocatable-target -save-temps \
84 // RUN: %s 2>&1 | FileCheck -check-prefix=CHK-BCLIB-DIR %s
86 /// Create a bogus bitcode library and find it with LIBRARY_PATH
87 // RUN: env LIBRARY_PATH=%S/Inputs/libomptarget/subdir %clang -### -no-canonical-prefixes -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda \
88 // RUN: -Xopenmp-target -march=sm_52 --cuda-path=%S/Inputs/CUDA_102/usr/local/cuda \
89 // RUN: -fopenmp-relocatable-target -save-temps \
90 // RUN: %s 2>&1 | FileCheck -check-prefix=CHK-ENV-BCLIB %s
92 // CHK-BCLIB: clang{{.*}}-triple{{.*}}nvptx64-nvidia-cuda{{.*}}-mlink-builtin-bitcode{{.*}}libomptarget-nvptx-test.bc
93 // CHK-BCLIB-DIR: clang{{.*}}-triple{{.*}}nvptx64-nvidia-cuda{{.*}}-mlink-builtin-bitcode{{.*}}libomptarget{{/|\\\\}}libomptarget-nvptx-sm_52.bc
94 // CHK-ENV-BCLIB: clang{{.*}}-triple{{.*}}nvptx64-nvidia-cuda{{.*}}-mlink-builtin-bitcode{{.*}}subdir{{/|\\\\}}libomptarget-nvptx-sm_52.bc
95 // CHK-BCLIB-NOT: {{error:|warning:}}
97 /// ###########################################################################
99 /// Check that the error is thrown when the libomptarget bitcode library does not exist.
100 // RUN: not %clang -### -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda \
101 // RUN: -Xopenmp-target -march=sm_52 --cuda-path=%S/Inputs/CUDA_102/usr/local/cuda \
102 // RUN: --libomptarget-nvptx-bc-path=not-exist.bc \
103 // RUN: -fopenmp-relocatable-target -save-temps %s 2>&1 \
104 // RUN: | FileCheck -check-prefix=CHK-BCLIB-ERROR %s
106 // CHK-BCLIB-ERROR: bitcode library 'not-exist.bc' does not exist
108 /// ###########################################################################
110 /// Check that the error is thrown when CUDA 9.1 or lower version is used.
111 // RUN: not %clang -### -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda \
112 // RUN: -Xopenmp-target -march=sm_52 --cuda-path=%S/Inputs/CUDA_90/usr/local/cuda \
113 // RUN: -fopenmp-relocatable-target -save-temps %s 2>&1 \
114 // RUN: | FileCheck -check-prefix=CHK-CUDA-VERSION-ERROR %s
116 // CHK-CUDA-VERSION-ERROR: NVPTX target requires CUDA 9.2 or above; CUDA 9.0 detected
118 /// Check that debug info is emitted in dwarf-2
119 // RUN: %clang -### -nogpulib -nogpuinc -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda -Xopenmp-target -march=sm_60 %s -g -O1 --no-cuda-noopt-device-debug 2>&1 \
120 // RUN: | FileCheck -check-prefix=DEBUG_DIRECTIVES %s
121 // RUN: %clang -### -nogpulib -nogpuinc -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda -Xopenmp-target -march=sm_60 %s -g -O3 2>&1 \
122 // RUN: | FileCheck -check-prefix=DEBUG_DIRECTIVES %s
123 // RUN: %clang -### -nogpulib -nogpuinc -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda -Xopenmp-target -march=sm_60 %s -g -O3 --no-cuda-noopt-device-debug 2>&1 \
124 // RUN: | FileCheck -check-prefix=DEBUG_DIRECTIVES %s
125 // RUN: %clang -### -nogpulib -nogpuinc -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda -Xopenmp-target -march=sm_60 %s -g0 2>&1 \
126 // RUN: | FileCheck -check-prefix=NO_DEBUG %s
127 // RUN: %clang -### -nogpulib -nogpuinc -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda -Xopenmp-target -march=sm_60 %s -ggdb0 -O3 --cuda-noopt-device-debug 2>&1 \
128 // RUN: | FileCheck -check-prefix=NO_DEBUG %s
129 // RUN: %clang -### -nogpulib -nogpuinc -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda -Xopenmp-target -march=sm_60 %s -gline-directives-only 2>&1 \
130 // RUN: | FileCheck -check-prefix=DEBUG_DIRECTIVES %s
132 // DEBUG_DIRECTIVES-NOT: warning: debug
133 // NO_DEBUG-NOT: warning: debug
134 // NO_DEBUG: "-fopenmp-is-target-device"
135 // NO_DEBUG-NOT: "-debug-info-kind=
136 // NO_DEBUG: ptxas
137 // DEBUG_DIRECTIVES: "-triple" "nvptx64-nvidia-cuda"
138 // DEBUG_DIRECTIVES-SAME: "-debug-info-kind=line-directives-only"
139 // DEBUG_DIRECTIVES-SAME: "-fopenmp-is-target-device"
140 // DEBUG_DIRECTIVES: ptxas
141 // DEBUG_DIRECTIVES: "-lineinfo"
143 // RUN: %clang -### -nogpulib -nogpuinc -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda -Xopenmp-target -march=sm_60 %s -g -O0 --no-cuda-noopt-device-debug 2>&1 \
144 // RUN: | FileCheck -check-prefix=HAS_DEBUG %s
145 // RUN: %clang -### -nogpulib -nogpuinc -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda -Xopenmp-target -march=sm_60 %s -g 2>&1 \
146 // RUN: | FileCheck -check-prefix=HAS_DEBUG %s
147 // RUN: %clang -### -nogpulib -nogpuinc -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda -Xopenmp-target -march=sm_60 %s -g -O0 --cuda-noopt-device-debug 2>&1 \
148 // RUN: | FileCheck -check-prefix=HAS_DEBUG %s
149 // RUN: %clang -### -nogpulib -nogpuinc -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda -Xopenmp-target -march=sm_60 %s -g -O3 --cuda-noopt-device-debug 2>&1 \
150 // RUN: | FileCheck -check-prefix=HAS_DEBUG %s
151 // RUN: %clang -### -nogpulib -nogpuinc -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda -Xopenmp-target -march=sm_60 %s -g2 2>&1 \
152 // RUN: | FileCheck -check-prefix=HAS_DEBUG %s
153 // RUN: %clang -### -nogpulib -nogpuinc -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda -Xopenmp-target -march=sm_60 %s -ggdb2 -O0 --cuda-noopt-device-debug 2>&1 \
154 // RUN: | FileCheck -check-prefix=HAS_DEBUG %s
155 // RUN: %clang -### -nogpulib -nogpuinc -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda -Xopenmp-target -march=sm_60 %s -g3 -O3 --cuda-noopt-device-debug 2>&1 \
156 // RUN: | FileCheck -check-prefix=HAS_DEBUG %s
157 // RUN: %clang -### -nogpulib -nogpuinc -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda -Xopenmp-target -march=sm_60 %s -ggdb3 -O2 --cuda-noopt-device-debug 2>&1 \
158 // RUN: | FileCheck -check-prefix=HAS_DEBUG %s
159 // RUN: %clang -### -nogpulib -nogpuinc -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda -Xopenmp-target -march=sm_60 %s -gline-tables-only 2>&1 \
160 // RUN: | FileCheck -check-prefix=HAS_DEBUG %s
161 // RUN: %clang -### -nogpulib -nogpuinc -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda -Xopenmp-target -march=sm_60 %s -ggdb1 -O2 --cuda-noopt-device-debug 2>&1 \
162 // RUN: | FileCheck -check-prefix=HAS_DEBUG %s
164 // HAS_DEBUG-NOT: warning: debug
165 // HAS_DEBUG: "-triple" "nvptx64-nvidia-cuda"
166 // HAS_DEBUG-SAME: "-debug-info-kind={{constructor|line-tables-only}}"
167 // HAS_DEBUG-SAME: "-dwarf-version=2"
168 // HAS_DEBUG-SAME: "-fopenmp-is-target-device"
169 // HAS_DEBUG: ptxas
170 // HAS_DEBUG-SAME: "-g"
171 // HAS_DEBUG-SAME: "--dont-merge-basicblocks"
172 // HAS_DEBUG-SAME: "--return-at-end"
174 // RUN: %clang -### -nogpulib -nogpuinc -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda -Xopenmp-target -march=sm_60 %s -fopenmp-cuda-mode 2>&1 \
175 // RUN: | FileCheck -check-prefix=CUDA_MODE %s
176 // RUN: %clang -### -nogpulib -nogpuinc -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda -Xopenmp-target -march=sm_60 %s -fno-openmp-cuda-mode -fopenmp-cuda-mode 2>&1 \
177 // RUN: | FileCheck -check-prefix=CUDA_MODE %s
178 // RUN: %clang -### -nogpulib -nogpuinc -fopenmp=libomp -fopenmp-targets=amdgcn-amd-amdhsa -Xopenmp-target -march=gfx906 %s -fopenmp-cuda-mode 2>&1 \
179 // RUN: | FileCheck -check-prefix=CUDA_MODE %s
180 // RUN: %clang -### -nogpulib -nogpuinc -fopenmp=libomp -fopenmp-targets=amdgcn-amd-amdhsa -Xopenmp-target -march=gfx906 %s -fno-openmp-cuda-mode -fopenmp-cuda-mode 2>&1 \
181 // RUN: | FileCheck -check-prefix=CUDA_MODE %s
182 // CUDA_MODE: "-cc1"{{.*}}"-triple" "{{nvptx64-nvidia-cuda|amdgcn-amd-amdhsa}}"
183 // CUDA_MODE-SAME: "-fopenmp-cuda-mode"
184 // RUN: %clang -### -nogpulib -nogpuinc -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda -Xopenmp-target -march=sm_60 %s -fno-openmp-cuda-mode 2>&1 \
185 // RUN: | FileCheck -check-prefix=NO_CUDA_MODE %s
186 // RUN: %clang -### -nogpulib -nogpuinc -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda -Xopenmp-target -march=sm_60 %s -fopenmp-cuda-mode -fno-openmp-cuda-mode 2>&1 \
187 // RUN: | FileCheck -check-prefix=NO_CUDA_MODE %s
188 // RUN: %clang -### -nogpulib -nogpuinc -fopenmp=libomp -fopenmp-targets=amdgcn-amd-amdhsa -Xopenmp-target -march=gfx906 %s -fno-openmp-cuda-mode 2>&1 \
189 // RUN: | FileCheck -check-prefix=NO_CUDA_MODE %s
190 // RUN: %clang -### -nogpulib -nogpuinc -fopenmp=libomp -fopenmp-targets=amdgcn-amd-amdhsa -Xopenmp-target -march=gfx906 %s -fopenmp-cuda-mode -fno-openmp-cuda-mode 2>&1 \
191 // RUN: | FileCheck -check-prefix=NO_CUDA_MODE %s
192 // NO_CUDA_MODE-NOT: "-{{fno-|f}}openmp-cuda-mode"
194 // RUN: %clang -### -nogpulib -nogpuinc -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda -Xopenmp-target -march=sm_60 %s -fopenmp-cuda-teams-reduction-recs-num=2048 2>&1 \
195 // RUN: | FileCheck -check-prefix=CUDA_RED_RECS %s
196 // CUDA_RED_RECS: "-cc1"{{.*}}"-triple" "nvptx64-nvidia-cuda"
197 // CUDA_RED_RECS-SAME: "-fopenmp-cuda-teams-reduction-recs-num=2048"
199 // RUN: %clang -### -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda --cuda-path=%S/Inputs/CUDA_102/usr/local/cuda \
200 // RUN: --offload-arch=sm_52 --libomptarget-nvptx-bc-path=%S/Inputs/libomptarget/libomptarget-nvptx-test.bc %s 2>&1 \
201 // RUN: | FileCheck -check-prefix=OPENMP_NVPTX_WRAPPERS %s
202 // OPENMP_NVPTX_WRAPPERS: "-cc1"{{.*}}"-triple" "nvptx64-nvidia-cuda"
203 // OPENMP_NVPTX_WRAPPERS-SAME: "-internal-isystem" "{{.*}}openmp_wrappers"
205 // RUN: %clang -### -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda -nogpulib -nogpuinc \
206 // RUN: --offload-arch=sm_52 -save-temps -ccc-print-bindings %s -o openmp-offload-gpu 2>&1 \
207 // RUN: | FileCheck -check-prefix=SAVE_TEMPS_NAMES %s
209 // SAVE_TEMPS_NAMES-NOT: "GNU::Linker"{{.*}}["[[SAVE_TEMPS_INPUT1:.*\.o]]", "[[SAVE_TEMPS_INPUT1]]"]
211 // RUN: %clang -### -fopenmp=libomp -fopenmp-targets=nvptx64 -Xopenmp-target=nvptx64 -march=sm_52 \
212 // RUN: -nogpulib -nogpuinc -save-temps %s -o openmp-offload-gpu 2>&1 \
213 // RUN: | FileCheck -check-prefix=TRIPLE %s
215 // TRIPLE: "-triple" "nvptx64-nvidia-cuda"
216 // TRIPLE: "-target-cpu" "sm_52"
218 // RUN: %clang -### --target=x86_64-unknown-linux-gnu -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda \
219 // RUN: -Xopenmp-target=nvptx64-nvidia-cuda -march=sm_52 --cuda-path=%S/Inputs/CUDA_102/usr/local/cuda \
220 // RUN: --libomptarget-nvptx-bc-path=%S/Inputs/libomptarget/libomptarget-nvptx-test.bc %s 2>&1 \
221 // RUN: | FileCheck %s
222 // RUN: %clang -### --target=x86_64-unknown-linux-gnu -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda \
223 // RUN: --offload-arch=sm_52 --cuda-path=%S/Inputs/CUDA_102/usr/local/cuda \
224 // RUN: --libomptarget-nvptx-bc-path=%S/Inputs/libomptarget/libomptarget-nvptx-test.bc %s 2>&1 \
225 // RUN: | FileCheck %s
227 // verify the tools invocations
228 // CHECK: "-cc1" "-triple" "x86_64-unknown-linux-gnu"{{.*}}"-emit-llvm-bc"{{.*}}"-x" "c"
229 // CHECK: "-cc1" "-triple" "nvptx64-nvidia-cuda" "-aux-triple" "x86_64-unknown-linux-gnu"{{.*}}"-target-cpu" "sm_52"
230 // CHECK: "-cc1" "-triple" "x86_64-unknown-linux-gnu"{{.*}}"-emit-obj"
231 // CHECK: clang-linker-wrapper{{.*}} "-o" "a.out"
233 // RUN: %clang -ccc-print-phases --target=x86_64-unknown-linux-gnu -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda -Xopenmp-target=nvptx64-nvidia-cuda -march=sm_52 %s 2>&1 \
234 // RUN: | FileCheck --check-prefix=CHECK-PHASES %s
235 // CHECK-PHASES: 0: input, "[[INPUT:.+]]", c, (host-openmp)
236 // CHECK-PHASES: 1: preprocessor, {0}, cpp-output, (host-openmp)
237 // CHECK-PHASES: 2: compiler, {1}, ir, (host-openmp)
238 // CHECK-PHASES: 3: input, "[[INPUT]]", c, (device-openmp)
239 // CHECK-PHASES: 4: preprocessor, {3}, cpp-output, (device-openmp)
240 // CHECK-PHASES: 5: compiler, {4}, ir, (device-openmp)
241 // CHECK-PHASES: 6: offload, "host-openmp (x86_64-unknown-linux-gnu)" {2}, "device-openmp (nvptx64-nvidia-cuda)" {5}, ir
242 // CHECK-PHASES: 7: backend, {6}, assembler, (device-openmp)
243 // CHECK-PHASES: 8: assembler, {7}, object, (device-openmp)
244 // CHECK-PHASES: 9: offload, "device-openmp (nvptx64-nvidia-cuda)" {8}, object
245 // CHECK-PHASES: 10: clang-offload-packager, {9}, image
246 // CHECK-PHASES: 11: offload, "host-openmp (x86_64-unknown-linux-gnu)" {2}, "device-openmp (x86_64-unknown-linux-gnu)" {10}, ir
247 // CHECK-PHASES: 12: backend, {11}, assembler, (host-openmp)
248 // CHECK-PHASES: 13: assembler, {12}, object, (host-openmp)
249 // CHECK-PHASES: 14: clang-linker-wrapper, {13}, image, (host-openmp)
251 // RUN: %clang -### --target=x86_64-unknown-linux-gnu -ccc-print-bindings -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda -Xopenmp-target=nvptx64-nvidia-cuda -march=sm_52 -nogpulib %s 2>&1 | FileCheck %s --check-prefix=CHECK-BINDINGS
252 // CHECK-BINDINGS: "x86_64-unknown-linux-gnu" - "clang", inputs: ["[[INPUT:.+]]"], output: "[[HOST_BC:.+]]"
253 // CHECK-BINDINGS: "nvptx64-nvidia-cuda" - "clang", inputs: ["[[INPUT]]", "[[HOST_BC]]"], output: "[[DEVICE_BC:.+]]"
254 // CHECK-BINDINGS: "nvptx64-nvidia-cuda" - "NVPTX::Assembler", inputs: ["[[DEVICE_BC]]"], output: "[[DEVICE_OBJ:.+]]"
255 // CHECK-BINDINGS: "x86_64-unknown-linux-gnu" - "Offload::Packager", inputs: ["[[DEVICE_OBJ]]"], output: "[[BINARY:.+.out]]"
256 // CHECK-BINDINGS: "x86_64-unknown-linux-gnu" - "clang", inputs: ["[[HOST_BC]]", "[[BINARY]]"], output: "[[HOST_OBJ:.+]]"
257 // CHECK-BINDINGS: "x86_64-unknown-linux-gnu" - "Offload::Linker", inputs: ["[[HOST_OBJ]]"], output: "a.out"
259 // RUN: %clang -### --target=x86_64-unknown-linux-gnu -ccc-print-bindings -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda -Xopenmp-target=nvptx64-nvidia-cuda -march=sm_52 -nogpulib -save-temps %s 2>&1 | FileCheck %s --check-prefix=CHECK-TEMP-BINDINGS
260 // CHECK-TEMP-BINDINGS: "x86_64-unknown-linux-gnu" - "Offload::Packager", inputs: ["[[DEVICE_OBJ:.+]]"], output: "[[BINARY:.+.out]]"
262 // RUN: %clang -### --target=x86_64-unknown-linux-gnu -ccc-print-bindings -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda --offload-arch=sm_52 --offload-arch=sm_70 -nogpuinc -nogpulib %s 2>&1 | FileCheck %s --check-prefix=CHECK-ARCH-BINDINGS
263 // RUN: %clang -### --target=x86_64-unknown-linux-gnu -ccc-print-bindings -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda --offload-arch=sm_52,sm_70 -nogpuinc -nogpulib %s 2>&1 | FileCheck %s --check-prefix=CHECK-ARCH-BINDINGS
264 // RUN: %clang -### --target=x86_64-unknown-linux-gnu -ccc-print-bindings -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda --offload-arch=sm_52,sm_60,sm_70,sm_80 --no-offload-arch=sm_60,sm_80 -nogpuinc -nogpulib %s 2>&1 | FileCheck %s --check-prefix=CHECK-ARCH-BINDINGS
265 // CHECK-ARCH-BINDINGS: "x86_64-unknown-linux-gnu" - "clang", inputs: ["[[INPUT:.*]]"], output: "[[HOST_BC:.*]]"
266 // CHECK-ARCH-BINDINGS: "nvptx64-nvidia-cuda" - "clang", inputs: ["[[INPUT]]", "[[HOST_BC]]"], output: "[[DEVICE_BC_SM_52:.*]]"
267 // CHECK-ARCH-BINDINGS: "nvptx64-nvidia-cuda" - "NVPTX::Assembler", inputs: ["[[DEVICE_BC_SM_52]]"], output: "[[DEVICE_OBJ_SM_52:.*]]"
268 // CHECK-ARCH-BINDINGS: "nvptx64-nvidia-cuda" - "clang", inputs: ["[[INPUT]]", "[[HOST_BC]]"], output: "[[DEVICE_BC_SM_70:.*]]"
269 // CHECK-ARCH-BINDINGS: "nvptx64-nvidia-cuda" - "NVPTX::Assembler", inputs: ["[[DEVICE_BC_SM_70]]"], output: "[[DEVICE_OBJ_SM_70:.*]]"
270 // CHECK-ARCH-BINDINGS: "x86_64-unknown-linux-gnu" - "Offload::Packager", inputs: ["[[DEVICE_OBJ_SM_52]]", "[[DEVICE_OBJ_SM_70]]"], output: "[[BINARY:.*]]"
271 // CHECK-ARCH-BINDINGS: "x86_64-unknown-linux-gnu" - "clang", inputs: ["[[HOST_BC]]", "[[BINARY]]"], output: "[[HOST_OBJ:.*]]"
272 // CHECK-ARCH-BINDINGS: "x86_64-unknown-linux-gnu" - "Offload::Linker", inputs: ["[[HOST_OBJ]]"], output: "a.out"
274 // RUN: %clang -### --target=x86_64-unknown-linux-gnu -ccc-print-bindings -fopenmp=libomp \
275 // RUN: -fopenmp-targets=nvptx64-nvidia-cuda,amdgcn-amd-amdhsa -Xopenmp-target=nvptx64-nvidia-cuda --offload-arch=sm_70 \
276 // RUN: -fopenmp-targets=nvptx64-nvidia-cuda,amdgcn-amd-amdhsa -Xopenmp-target=amdgcn-amd-amdhsa --offload-arch=gfx908 \
277 // RUN: -nogpulib %s 2>&1 | FileCheck %s --check-prefix=CHECK-NVIDIA-AMDGPU
279 // CHECK-NVIDIA-AMDGPU: "x86_64-unknown-linux-gnu" - "clang", inputs: ["[[INPUT:.+]]"], output: "[[HOST_BC:.+]]"
280 // CHECK-NVIDIA-AMDGPU: "amdgcn-amd-amdhsa" - "clang", inputs: ["[[INPUT]]", "[[HOST_BC]]"], output: "[[AMD_BC:.+]]"
281 // CHECK-NVIDIA-AMDGPU: "nvptx64-nvidia-cuda" - "clang", inputs: ["[[INPUT]]", "[[HOST_BC]]"], output: "[[NVIDIA_PTX:.+]]"
282 // CHECK-NVIDIA-AMDGPU: "nvptx64-nvidia-cuda" - "NVPTX::Assembler", inputs: ["[[NVIDIA_PTX]]"], output: "[[NVIDIA_CUBIN:.+]]"
283 // CHECK-NVIDIA-AMDGPU: "x86_64-unknown-linux-gnu" - "Offload::Packager", inputs: ["[[AMD_BC]]", "[[NVIDIA_CUBIN]]"], output: "[[BINARY:.*]]"
284 // CHECK-NVIDIA-AMDGPU: "x86_64-unknown-linux-gnu" - "clang", inputs: ["[[HOST_BC]]", "[[BINARY]]"], output: "[[HOST_OBJ:.+]]"
285 // CHECK-NVIDIA-AMDGPU: "x86_64-unknown-linux-gnu" - "Offload::Linker", inputs: ["[[HOST_OBJ]]"], output: "a.out"
287 // RUN: %clang -x ir -### --target=x86_64-unknown-linux-gnu -ccc-print-bindings -fopenmp=libomp --offload-arch=sm_52 -nogpulib %s 2>&1 | FileCheck %s --check-prefix=CHECK-IR
289 // CHECK-IR: "x86_64-unknown-linux-gnu" - "clang", inputs: ["[[INPUT_IR:.+]]"], output: "[[OBJECT:.+]]"
290 // CHECK-IR: "x86_64-unknown-linux-gnu" - "Offload::Linker", inputs: ["[[OBJECT]]"], output: "a.out"
292 // RUN: %clang -### --target=x86_64-unknown-linux-gnu -emit-llvm -S -fopenmp=libomp --offload-device-only \
293 // RUN: -fopenmp-targets=nvptx64-nvidia-cuda -Xopenmp-target=nvptx64-nvidia-cuda -march=sm_52 -nogpulib %s 2>&1 \
294 // RUN: | FileCheck %s --check-prefix=CHECK-EMIT-LLVM-IR
295 // CHECK-EMIT-LLVM-IR: "-cc1"{{.*}}"-triple" "nvptx64-nvidia-cuda"{{.*}}"-emit-llvm"
297 // RUN: %clang -### --target=x86_64-unknown-linux-gnu -emit-llvm -S -fopenmp=libomp \
298 // RUN: -fopenmp-targets=nvptx64-nvidia-cuda -Xopenmp-target=nvptx64-nvidia-cuda -march=sm_52 -nogpulib %s 2>&1 \
299 // RUN: | FileCheck %s --check-prefix=CHECK-EMIT-LLVM-IR-BC
300 // CHECK-EMIT-LLVM-IR-BC: "-cc1"{{.*}}"-triple" "nvptx64-nvidia-cuda"{{.*}}"-emit-llvm-bc"
302 // RUN: %clang -### -fopenmp=libomp --offload-arch=sm_89 \
303 // RUN: --no-cuda-version-check \
304 // RUN: -nogpulib %s -o openmp-offload-gpu 2>&1 \
305 // RUN: | FileCheck -check-prefix=DRIVER_EMBEDDING %s
307 // DRIVER_EMBEDDING: -fembed-offload-object={{.*}}.out
309 // RUN: %clang -### --target=x86_64-unknown-linux-gnu -ccc-print-bindings -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda \
310 // RUN: --offload-host-only -nogpulib %s 2>&1 | FileCheck %s --check-prefix=CHECK-HOST-ONLY
311 // CHECK-HOST-ONLY: "x86_64-unknown-linux-gnu" - "clang", inputs: ["[[INPUT:.*]]"], output: "[[OUTPUT:.*]]"
312 // CHECK-HOST-ONLY: "x86_64-unknown-linux-gnu" - "Offload::Linker", inputs: ["[[OUTPUT]]"], output: "a.out"
314 // RUN: %clang -### --target=x86_64-unknown-linux-gnu -ccc-print-bindings -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda \
315 // RUN: -Xopenmp-target=nvptx64-nvidia-cuda -march=sm_52 --offload-device-only -nogpulib %s 2>&1 | FileCheck %s --check-prefix=CHECK-DEVICE-ONLY
316 // CHECK-DEVICE-ONLY: "x86_64-unknown-linux-gnu" - "clang", inputs: ["[[INPUT:.*]]"], output: "[[HOST_BC:.*]]"
317 // CHECK-DEVICE-ONLY: "nvptx64-nvidia-cuda" - "clang", inputs: ["[[INPUT]]", "[[HOST_BC]]"], output: "[[DEVICE_ASM:.*]]"
318 // CHECK-DEVICE-ONLY: "nvptx64-nvidia-cuda" - "NVPTX::Assembler", inputs: ["[[DEVICE_ASM]]"], output: "{{.*}}-openmp-nvptx64-nvidia-cuda.o"
320 // RUN: %clang -### --target=x86_64-unknown-linux-gnu -ccc-print-bindings -fopenmp=libomp -fopenmp-targets=nvptx64-nvidia-cuda \
321 // RUN: -Xopenmp-target=nvptx64-nvidia-cuda -march=sm_52 --offload-device-only -E -nogpulib %s 2>&1 | FileCheck %s --check-prefix=CHECK-DEVICE-ONLY-PP
322 // CHECK-DEVICE-ONLY-PP: "nvptx64-nvidia-cuda" - "clang", inputs: ["[[INPUT:.*]]"], output: "-"
324 // RUN: %clang -### --target=x86_64-unknown-linux-gnu -fopenmp=libomp --offload-arch=sm_52 --cuda-path=%S/Inputs/CUDA_102/usr/local/cuda \
325 // RUN: -foffload-lto %s 2>&1 | FileCheck --check-prefix=CHECK-LTO-LIBRARY %s
327 // CHECK-LTO-LIBRARY: {{.*}}-lomptarget{{.*}}-lomptarget.devicertl
329 // RUN: %clang -### --target=x86_64-unknown-linux-gnu -fopenmp=libomp --offload-arch=sm_52 --cuda-path=%S/Inputs/CUDA_102/usr/local/cuda \
330 // RUN: --libomptarget-nvptx-bc-path=%S/Inputs/libomptarget/libomptarget-nvptx-test.bc %s 2>&1 \
331 // RUN: | FileCheck --check-prefix=CHECK-NO-LTO-LIBRARY %s
333 // CHECK-NO-LTO-LIBRARY: {{.*}}-lomptarget{{.*}}-lomptarget.devicertl
335 // RUN: %clang -### --target=x86_64-unknown-linux-gnu -fopenmp=libomp --offload-arch=sm_52 -nogpulib \
336 // RUN: -foffload-lto %s 2>&1 | FileCheck --check-prefix=CHECK-NO-LIBRARY %s
338 // CHECK-NO-LIBRARY-NOT: {{.*}}-lomptarget{{.*}}-lomptarget.devicertl
340 // RUN: %clang -### --target=x86_64-unknown-linux-gnu -fopenmp=libomp --offload-arch=sm_52 -nogpulib \
341 // RUN: -Xoffload-linker a -Xoffload-linker-nvptx64-nvidia-cuda b -Xoffload-linker-nvptx64 c \
342 // RUN: %s 2>&1 | FileCheck --check-prefix=CHECK-XLINKER %s
344 // CHECK-XLINKER: -device-linker=a{{.*}}-device-linker=nvptx64-nvidia-cuda=b{{.*}}-device-linker=nvptx64-nvidia-cuda=c{{.*}}--
346 // RUN: %clang -### --target=x86_64-unknown-linux-gnu -fopenmp=libomp --offload-arch=sm_52 -nogpulib \
347 // RUN: -foffload-lto %s 2>&1 | FileCheck --check-prefix=CHECK-LTO-FEATURES %s
349 // CHECK-LTO-FEATURES: clang-offload-packager{{.*}}--image={{.*}}feature=+ptx{{[0-9]+}}
351 // RUN: %clang -### --target=x86_64-unknown-linux-gnu -fopenmp=libomp --offload-arch=sm_52 -nogpulib \
352 // RUN: -Xopenmp-target=nvptx64-nvidia-cuda --cuda-feature=+ptx64 -foffload-lto %s 2>&1 \
353 // RUN: | FileCheck --check-prefix=CHECK-SET-FEATURES %s
355 // CHECK-SET-FEATURES: clang-offload-packager{{.*}}--image={{.*}}feature=+ptx64
358 // Check that `-Xarch_host` works for OpenMP offloading.
360 // RUN: %clang -### --target=x86_64-unknown-linux-gnu -fopenmp=libomp \
361 // RUN: --offload-arch=sm_52 -nogpulib -nogpuinc -Xarch_host -O3 %s 2>&1 \
362 // RUN: | FileCheck --check-prefix=XARCH-HOST %s
363 // XARCH-HOST: "-cc1" "-triple" "x86_64-unknown-linux-gnu"{{.*}}"-O3"
364 // XARCH-HOST-NOT: "-cc1" "-triple" "nvptx64-nvidia-cuda"{{.*}}"-O3"
367 // Check that `-Xarch_device` works for OpenMP offloading.
369 // RUN: %clang -### --target=x86_64-unknown-linux-gnu -fopenmp=libomp \
370 // RUN: --offload-arch=sm_52 -nogpulib -nogpuinc -Xarch_device -O3 %s 2>&1 \
371 // RUN: | FileCheck --check-prefix=XARCH-DEVICE %s
372 // XARCH-DEVICE: "-cc1" "-triple" "nvptx64-nvidia-cuda"{{.*}}"-O3"
373 // XARCH-DEVICE-NOT: "-cc1" "-triple" "x86_64-unknown-linux-gnu"{{.*}}"-O3"
375 // RUN: %clang -### --target=x86_64-unknown-linux-gnu -fopenmp=libomp \
376 // RUN: --libomptarget-nvptx-bc-path=%S/Inputs/libomptarget/libomptarget-nvptx-test.bc \
377 // RUN: --cuda-path=%S/Inputs/CUDA_102/usr/local/cuda \
378 // RUN: --offload-arch=sm_52 -nogpulibc -nogpuinc %s 2>&1 \
379 // RUN: | FileCheck --check-prefix=LIBC-GPU %s
380 // LIBC-GPU: clang-linker-wrapper{{.*}}"--device-compiler=-nolibc"