Skip to content

[SYCL][CUDA] Add CoreOption flag to CUDA backend options #4930

New issue

Have a question about this project? Sign up for a free GitHub account to open an issue and contact its maintainers and the community.

Sign up for GitHub

By clicking “Sign up for GitHub”, you agree to our terms of service and privacy statement. We’ll occasionally send you account related emails.

Already on GitHub? Sign in to your account

Merged
merged 1 commit into from
Nov 12, 2021
Merged

[SYCL][CUDA] Add CoreOption flag to CUDA backend options #4930

Show file tree
Hide file tree
Changes from all commits
Commits
Show all changes
1 commit
Select commit
File filter

Filter by extension

Filter by extension
Conversations
Failed to load comments.
Loading
Jump to
Jump to file
Failed to load files.
Loading
Diff view
Diff view
8 changes: 4 additions & 4 deletions clang/include/clang/Driver/Options.td
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -937,7 +937,7 @@ def offload_arch_EQ : Joined<["--"], "offload-arch=">, Flags<[NoXarchOption]>,
"device architecture followed by target ID features delimited by a colon. Each target ID feature "
"is a pre-defined string followed by a plus or minus sign (e.g. gfx908:xnack+:sramecc-). May be "
"specified more than once.">;
def cuda_gpu_arch_EQ : Joined<["--"], "cuda-gpu-arch=">, Flags<[NoXarchOption]>,
def cuda_gpu_arch_EQ : Joined<["--"], "cuda-gpu-arch=">, Flags<[NoXarchOption, CoreOption]>,
Alias<offload_arch_EQ>;
def hip_link : Flag<["--"], "hip-link">,
HelpText<"Link clang-offload-bundler bundles for HIP">;
Expand All @@ -954,7 +954,7 @@ def no_cuda_version_check : Flag<["--"], "no-cuda-version-check">,
HelpText<"Don't error out if the detected version of the CUDA install is "
"too low for the requested CUDA gpu architecture.">;
def no_cuda_noopt_device_debug : Flag<["--"], "no-cuda-noopt-device-debug">;
def cuda_path_EQ : Joined<["--"], "cuda-path=">, Group<i_Group>,
def cuda_path_EQ : Joined<["--"], "cuda-path=">, Flags<[CoreOption]>, Group<i_Group>,
HelpText<"CUDA installation path">;
def cuda_path_ignore_env : Flag<["--"], "cuda-path-ignore-env">, Group<i_Group>,
HelpText<"Ignore environment variables to detect CUDA installation">;
Expand Down Expand Up @@ -2691,7 +2691,7 @@ def fsycl_help : Flag<["-"], "fsycl-help">, Alias<fsycl_help_EQ>,
"from all of the offline compilation tools">;
def fsycl_libspirv_path_EQ : Joined<["-"], "fsycl-libspirv-path=">,
Flags<[CC1Option, CoreOption]>, HelpText<"Path to libspirv library">;
def fno_sycl_libspirv : Flag<["-"], "fno-sycl-libspirv">, HelpText<"Disable check for libspirv">;
def fno_sycl_libspirv : Flag<["-"], "fno-sycl-libspirv">, Flags<[CoreOption]>, HelpText<"Disable check for libspirv">;
def fsycl_host_compiler_EQ : Joined<["-"], "fsycl-host-compiler=">,
Flags<[CoreOption]>, HelpText<"Specify C++ compiler binary to perform host "
"compilation with during SYCL offload compiles.">;
Expand Down Expand Up @@ -3843,7 +3843,7 @@ def nogpuinc : Flag<["-"], "nogpuinc">, HelpText<"Do not add include paths for C
def : Flag<["-"], "nocudainc">, Alias<nogpuinc>;
def nogpulib : Flag<["-"], "nogpulib">,
HelpText<"Do not link device library for CUDA/HIP device compilation">;
def : Flag<["-"], "nocudalib">, Alias<nogpulib>;
def : Flag<["-"], "nocudalib">, Flags<[CoreOption]>, Alias<nogpulib>;
def nodefaultlibs : Flag<["-"], "nodefaultlibs">;
def nofixprebinding : Flag<["-"], "nofixprebinding">;
def nolibc : Flag<["-"], "nolibc">;
Expand Down
61 changes: 56 additions & 5 deletions clang/test/Driver/sycl-offload-nvptx.cpp
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8,6 +8,11 @@
// RUN: -fsycl-targets=nvptx64-nvidia-cuda --cuda-path=%S/Inputs/CUDA/usr/local/cuda \
// RUN: -fsycl-libspirv-path=%S/Inputs/SYCL/libspirv.bc %s 2>&1 \
// RUN: | FileCheck -check-prefix=CHK-ACTIONS %s
//
// RUN: %clang_cl -### -fsycl \
// RUN: -fsycl-targets=nvptx64-nvidia-cuda --cuda-path=%S/Inputs/CUDA/usr/local/cuda \
// RUN: -fsycl-libspirv-path=%S/Inputs/SYCL/libspirv.bc %s 2>&1 \
// RUN: | FileCheck -check-prefix=CHK-ACTIONS-WIN %s

// CHK-ACTIONS: "-cc1" "-triple" "nvptx64-nvidia-cuda" "-aux-triple" "x86_64-unknown-linux-gnu"{{.*}} "-fsycl-is-device"{{.*}} "-Wno-sycl-strict"{{.*}} "-sycl-std=2020" {{.*}} "-emit-llvm-bc" {{.*}} "-internal-isystem" "{{.*}}bin{{[/\\]+}}..{{[/\\]+}}include{{[/\\]+}}sycl"{{.*}} "-mlink-builtin-bitcode" "{{.*}}libspirv.bc"{{.*}} "-mlink-builtin-bitcode" "{{.*}}libdevice{{.*}}.10.bc"{{.*}} "-target-feature" "+ptx42"{{.*}} "-target-sdk-version=[[CUDA_VERSION:[0-9.]+]]"{{.*}} "-target-cpu" "sm_50"{{.*}} "-std=c++11"{{.*}}
// CHK-ACTIONS: sycl-post-link{{.*}} "-split=auto"
Expand All @@ -19,17 +24,32 @@
// CHK-ACTIONS-NOT: "-mllvm -sycl-opt"
// CHK-ACTIONS: clang-offload-wrapper"{{.*}} "-host=x86_64-unknown-linux-gnu" "-target=nvptx64" "-kind=sycl"{{.*}}

// CHK-ACTIONS-WIN: "-cc1" "-triple" "nvptx64-nvidia-cuda" "-aux-triple" "x86_64-pc-windows-msvc"{{.*}} "-fsycl-is-device"{{.*}} "-Wno-sycl-strict"{{.*}} "-sycl-std=2020" {{.*}} "-emit-llvm-bc" {{.*}} "-internal-isystem" "{{.*}}bin{{[/\\]+}}..{{[/\\]+}}include{{[/\\]+}}sycl"{{.*}} "-mlink-builtin-bitcode" "{{.*}}libspirv.bc"{{.*}} "-mlink-builtin-bitcode" "{{.*}}libdevice{{.*}}.10.bc"{{.*}} "-target-feature" "+ptx42"{{.*}} "-target-sdk-version=[[CUDA_VERSION:[0-9.]+]]"{{.*}} "-target-cpu" "sm_50"{{.*}}
// CHK-ACTIONS-WIN: sycl-post-link{{.*}} "-split=auto"
// CHK-ACTIONS-WIN: file-table-tform" "-extract=Code" "-drop_titles"
// CHK-ACTIONS-WIN: llvm-foreach" {{.*}} "--" "{{.*}}clang-{{[0-9]+}}"
// CHK-ACTIONS-WIN: llvm-foreach" {{.*}} "--" "{{.*}}ptxas"
// CHK-ACTIONS-WIN: llvm-foreach" {{.*}} "--" "{{.*}}fatbinary"
// CHK-ACTIONS-WIN: file-table-tform" "-replace=Code,Code"
// CHK-ACTIONS-WIN-NOT: "-mllvm -sycl-opt"
// CHK-ACTIONS-WIN: clang-offload-wrapper"{{.*}} "-host=x86_64-pc-windows-msvc" "-target=nvptx64" "-kind=sycl"{{.*}}

/// Check phases w/out specifying a compute capability.
// RUN: %clangxx -ccc-print-phases -std=c++11 -target x86_64-unknown-linux-gnu -fsycl \
// RUN: -fsycl-targets=nvptx64-nvidia-cuda %s 2>&1 \
// RUN: | FileCheck -check-prefix=CHK-PHASES-NO-CC %s
//
// RUN: %clang_cl -ccc-print-phases -fsycl \
// RUN: -fsycl-targets=nvptx64-nvidia-cuda %s 2>&1 \
// RUN: | FileCheck -check-prefix=CHK-PHASES-NO-CC %s
//
// CHK-PHASES-NO-CC: 0: input, "{{.*}}", c++, (host-sycl)
// CHK-PHASES-NO-CC: 1: append-footer, {0}, c++, (host-sycl)
// CHK-PHASES-NO-CC: 2: preprocessor, {1}, c++-cpp-output, (host-sycl)
// CHK-PHASES-NO-CC: 3: input, "{{.*}}", c++, (device-sycl, sm_50)
// CHK-PHASES-NO-CC: 4: preprocessor, {3}, c++-cpp-output, (device-sycl, sm_50)
// CHK-PHASES-NO-CC: 5: compiler, {4}, ir, (device-sycl, sm_50)
// CHK-PHASES-NO-CC: 6: offload, "host-sycl (x86_64-unknown-linux-gnu)" {2}, "device-sycl (nvptx64-nvidia-cuda:sm_50)" {5}, c++-cpp-output
// CHK-PHASES-NO-CC: 6: offload, "host-sycl (x86_64-{{.*}})" {2}, "device-sycl (nvptx64-nvidia-cuda:sm_50)" {5}, c++-cpp-output
// CHK-PHASES-NO-CC: 7: compiler, {6}, ir, (host-sycl)
// CHK-PHASES-NO-CC: 8: backend, {7}, assembler, (host-sycl)
// CHK-PHASES-NO-CC: 9: assembler, {8}, object, (host-sycl)
Expand All @@ -43,21 +63,26 @@
// CHK-PHASES-NO-CC: 17: foreach, {13, 16}, cuda-fatbin, (device-sycl, sm_50)
// CHK-PHASES-NO-CC: 18: file-table-tform, {12, 17}, tempfiletable, (device-sycl, sm_50)
// CHK-PHASES-NO-CC: 19: clang-offload-wrapper, {18}, object, (device-sycl, sm_50)
// CHK-PHASES-NO-CC: 20: offload, "host-sycl (x86_64-unknown-linux-gnu)" {10}, "device-sycl (nvptx64-nvidia-cuda:sm_50)" {19}, image
// CHK-PHASES-NO-CC: 20: offload, "host-sycl (x86_64-{{.*}})" {10}, "device-sycl (nvptx64-nvidia-cuda:sm_50)" {19}, image

/// Check phases specifying a compute capability.
// RUN: %clangxx -ccc-print-phases -std=c++11 -target x86_64-unknown-linux-gnu -fsycl \
// RUN: -fsycl-targets=nvptx64-nvidia-cuda \
N.)
// RUN: -Xsycl-target-backend "--cuda-gpu-arch=sm_35" %s 2>&1 \
// RUN: | FileCheck -check-prefix=CHK-PHASES %s
//
// RUN: %clang_cl -ccc-print-phases -fsycl \
// RUN: -fsycl-targets=nvptx64-nvidia-cuda \
// RUN: -Xsycl-target-backend "--cuda-gpu-arch=sm_35" %s 2>&1 \
// RUN: | FileCheck -check-prefix=CHK-PHASES %s
//
// CHK-PHASES: 0: input, "{{.*}}", c++, (host-sycl)
// CHK-PHASES: 1: append-footer, {0}, c++, (host-sycl)
// CHK-PHASES: 2: preprocessor, {1}, c++-cpp-output, (host-sycl)
// CHK-PHASES: 3: input, "{{.*}}", c++, (device-sycl, sm_35)
// CHK-PHASES: 4: preprocessor, {3}, c++-cpp-output, (device-sycl, sm_35)
// CHK-PHASES: 5: compiler, {4}, ir, (device-sycl, sm_35)
// CHK-PHASES: 6: offload, "host-sycl (x86_64-unknown-linux-gnu)" {2}, "device-sycl (nvptx64-nvidia-cuda:sm_35)" {5}, c++-cpp-output
// CHK-PHASES: 6: offload, "host-sycl (x86_64-{{.*}})" {2}, "device-sycl (nvptx64-nvidia-cuda:sm_35)" {5}, c++-cpp-output
// CHK-PHASES: 7: compiler, {6}, ir, (host-sycl)
// CHK-PHASES: 8: backend, {7}, assembler, (host-sycl)
// CHK-PHASES: 9: assembler, {8}, object, (host-sycl)
Expand All @@ -71,11 +96,37 @@ N.)
// CHK-PHASES: 17: foreach, {13, 16}, cuda-fatbin, (device-sycl, sm_35)
// CHK-PHASES: 18: file-table-tform, {12, 17}, tempfiletable, (device-sycl, sm_35)
// CHK-PHASES: 19: clang-offload-wrapper, {18}, object, (device-sycl, sm_35)
// CHK-PHASES: 20: offload, "host-sycl (x86_64-unknown-linux-gnu)" {10}, "device-sycl (nvptx64-nvidia-cuda:sm_35)" {19}, image
// CHK-PHASES: 20: offload, "host-sycl (x86_64-{{.*}})" {10}, "device-sycl (nvptx64-nvidia-cuda:sm_35)" {19}, image

/// Check calling preprocessor only
// RUN: %clangxx -E -fsycl -fsycl-targets=nvptx64-nvidia-cuda -ccc-print-phases %s 2>&1 \
// RUN: | FileCheck -check-prefix=CHK-PREPROC %s
// CHK-PREPROC: 1: preprocessor, {0}, c++-cpp-output, (device-sycl, sm_[[CUDA_VERSION:[0-9.]+]])
// CHK-PREPROC: 2: offload, "device-sycl (nvptx64-nvidia-cuda:sm_[[CUDA_VERSION]])" {1}, c++-cpp-output
// CHK-PREPROC: 4: compiler, {1}, none, (device-sycl, sm_[[CUDA_VERSION]])
//
// RUN: %clangxx -### -std=c++11 -target x86_64-unknown-linux-gnu -fsycl \
// RUN: -fsycl-targets=nvptx64-nvidia-cuda --cuda-path=%S/Inputs/no/CUDA/path/here \
// RUN: -fsycl-libspirv-path=%S/Inputs/SYCL/libspirv.bc %s 2>&1 \
// RUN: | FileCheck -check-prefix=CHK-CUDA-PATH-ERROR %s
//
// RUN: %clang_cl -### -fsycl \
// RUN: -fsycl-targets=nvptx64-nvidia-cuda --cuda-path=%S/Inputs/no/CUDA/path/here \
// RUN: -fsycl-libspirv-path=%S/Inputs/SYCL/libspirv.bc %s 2>&1 \
// RUN: | FileCheck -check-prefix=CHK-CUDA-PATH-ERROR %s
//
// CHK-CUDA-PATH-ERROR: provide path to different CUDA installation via '--cuda-path', or pass '-nocudalib' to build without linking with libdevice
//
//
// RUN: %clangxx -### -std=c++11 -target x86_64-unknown-linux-gnu -fsycl \
// RUN: -fsycl-targets=nvptx64-nvidia-cuda --cuda-path=%S/Inputs/no/CUDA/path/here \
// RUN: -fsycl-libspirv-path=%S/Inputs/SYCL/libspirv.bc -nocudalib %s 2>&1 \
// RUN: | FileCheck -check-prefix=CHK-CUDA-NO-LIB %s
//
// RUN: %clang_cl -### -fsycl \
// RUN: -fsycl-targets=nvptx64-nvidia-cuda --cuda-path=%S/Inputs/no/CUDA/path/here \
// RUN: -fsycl-libspirv-path=%S/Inputs/SYCL/libspirv.bc -nocudalib %s 2>&1 \
// RUN: | FileCheck -check-prefix=CHK-CUDA-NO-LIB %s
//
// CHK-CUDA-NO-LIB-NOT: provide path to different CUDA installation via '--cuda-path', or pass '-nocudalib' to build without linking with libdevice
//