Home
last modified time | relevance | path

Searched refs:CUDA (Results 1 – 25 of 150) sorted by relevance

123456

/llvm-project-15.0.7/openmp/libomptarget/plugins/cuda/
H A DCMakeLists.txt9 # Build a plugin for a CUDA machine if available.
13 "Whether to build CUDA plugin")
20 …libomptarget_say("Not building CUDA offloading plugin: only support CUDA in Linux x86_64, ppc64le,…
23 libomptarget_say("Not building CUDA offloading plugin: libelf dependency not found.")
27 libomptarget_say("Building CUDA offloading plugin.")
30 add_definitions(-DTARGET_NAME=CUDA)
41 libomptarget_say("Building CUDA plugin linked against libcuda")
64 libomptarget_say("Building CUDA plugin for dlopened libcuda")
101 # Report to the parent scope that we are building a plugin for CUDA.
106 libomptarget_say("Enable tests using CUDA plugin")
[all …]
/llvm-project-15.0.7/llvm/docs/
H A DCompileCudaWithLLVM.rst2 Compiling CUDA with clang
14 This document assumes a basic familiarity with CUDA. Information about CUDA
16 `CUDA programming guide
19 Compiling CUDA Code
25 CUDA is supported since llvm 3.9. Clang currently supports CUDA 7.0 through
27 attempt to use detected CUDA SDK it as if it were CUDA 11.5.
29 Before you build CUDA code, you'll need to have installed the CUDA SDK. See
30 `NVIDIA's CUDA installation guide
68 "CUDA driver version is insufficient for CUDA runtime version" errors when you
71 * ``<CUDA install path>`` -- the directory where you installed CUDA SDK.
[all …]
H A DNVPTXUsage.rst21 This document assumes a basic familiarity with CUDA and the PTX
23 language can be found in the `CUDA documentation
100 copy data to it by name with the CUDA Driver API.
117 generated PTX compatible with the CUDA Driver API.
119 Example: 32-bit PTX for CUDA Driver API: ``nvptx-nvidia-cuda``
223 map in the following way to CUDA builtins:
226 CUDA Builtin PTX Special Register Intrinsic
657 // CUDA initialization
664 std::cout << "Using CUDA Device [0]: " << name << "\n";
766 system location by the driver, not the CUDA toolkit.
[all …]
/llvm-project-15.0.7/mlir/lib/Dialect/GPU/
H A DCMakeLists.txt91 # Configure CUDA language support. Using check_language first allows us to
94 check_language(CUDA)
96 enable_language(CUDA)
99 "Building mlir with cuda support requires a working CUDA install")
108 # Add CUDA headers includes and the libcuda.so library.
/llvm-project-15.0.7/openmp/libomptarget/cmake/Modules/
H A DLibomptargetGetDependencies.cmake17 # CUDA : required to control offloading to NVIDIA GPUs.
143 # Looking for CUDA...
148 find_package(CUDA QUIET)
170 # Looking for CUDA Driver API... (needed for CUDA plugin)
254 # Looking for CUDA libdevice subdirectory
/llvm-project-15.0.7/polly/lib/External/ppcg/
H A DREADME49 Using PPCG to generate CUDA or OpenCL code
51 To convert a fragment of a C program to CUDA, insert a line containing
59 after the fragment. To generate CUDA code run
115 Compiling the generated CUDA code with nvcc
141 corresponding CUDA code runs fine.
157 in the CUDA or OpenCL code, but for now it is left to the user
195 the PPCG generated code using nvcc since CUDA does not support VLAs.
198 CUDA and function overloading
200 While CUDA supports function overloading based on the arguments types,
207 In the transformed (CUDA) code, however, overloading will cause the
[all …]
/llvm-project-15.0.7/clang/lib/Frontend/
H A DFrontendOptions.cpp22 .Case("cui", InputKind(Language::CUDA).getPreprocessed()) in getInputKindForExtension()
33 .Cases("cu", "cuh", Language::CUDA) in getInputKindForExtension()
/llvm-project-15.0.7/clang/test/Driver/
H A Dhip-inputs.hip21 // CHECK-NOT: error: mixed CUDA and HIP compilation is not supported
22 // MIX: error: mixed CUDA and HIP compilation is not supported
/llvm-project-15.0.7/clang/test/OpenMP/
H A Dreduction_implicit_map.cpp26 #if defined(CUDA)
88 #if defined(CUDA) in main()
/llvm-project-15.0.7/clang/lib/Sema/
H A DSemaCUDA.cpp38 assert(getLangOpts().CUDA && "Should only be called during CUDA compilation"); in PushForceCUDAHostDevice()
43 assert(getLangOpts().CUDA && "Should only be called during CUDA compilation"); in PopForceCUDAHostDevice()
663 assert(getLangOpts().CUDA && "Should only be called during CUDA compilation"); in maybeAddCUDAHostDeviceAttrs()
729 assert(getLangOpts().CUDA && "Should only be called during CUDA compilation"); in CUDADiagIfDeviceCode()
759 assert(getLangOpts().CUDA && "Should only be called during CUDA compilation"); in CUDADiagIfHostCode()
787 assert(getLangOpts().CUDA && "Should only be called during CUDA compilation"); in CheckCUDACall()
898 assert(getLangOpts().CUDA && "Should only be called during CUDA compilation"); in CUDASetLambdaAttrs()
907 assert(getLangOpts().CUDA && "Should only be called during CUDA compilation"); in checkCUDATargetOverload()
/llvm-project-15.0.7/mlir/lib/ExecutionEngine/
H A DCMakeLists.txt153 # Configure CUDA support. Using check_language first allows us to give a
156 check_language(CUDA)
158 enable_language(CUDA)
161 "Building the mlir cuda runner requires a working CUDA install")
/llvm-project-15.0.7/clang/lib/Basic/
H A DLangOptions.cpp173 Opts.CUDA = Lang == Language::CUDA || Opts.HIP; in setLangDefaults()
184 } else if (Opts.CUDA) { in setLangDefaults()
/llvm-project-15.0.7/clang/test/Driver/Inputs/CUDA_102/usr/local/cuda/
H A Dversion.txt1 CUDA Version 10.2.333
/llvm-project-15.0.7/clang/test/Driver/Inputs/CUDA_90/usr/local/cuda/
H A Dversion.txt1 CUDA Version 9.0.103
/llvm-project-15.0.7/clang/test/Frontend/
H A Dstdlang.c25 #if defined(CUDA)
/llvm-project-15.0.7/polly/include/polly/CodeGen/
H A DPPCGCodeGeneration.h21 enum GPURuntime { CUDA, OpenCL }; enumerator
/llvm-project-15.0.7/openmp/runtime/cmake/
H A DLibompCheckLinkerFlag.cmake45 # Ignore the warning about the newer or unknown CUDA version.
46 if(("${OUTPUT}" MATCHES ${regex}) AND NOT ("${OUTPUT}" MATCHES "Unknown CUDA version"))
/llvm-project-15.0.7/clang/docs/
H A DClangOffloadPackager.rst94 | IMG_Cubin | 0x03 | The image is a CUDA object file |
96 | IMG_Fatbinary | 0x04 | The image is a CUDA fatbinary file |
98 | IMG_PTX | 0x05 | The image is a CUDA PTX file |
111 | OFK_CUDA | 0x02 | The producer was CUDA |
/llvm-project-15.0.7/clang/test/CodeGenCUDA/Inputs/
H A Ddevice-code-2.ll1 ; Simple bit of IR to mimic CUDA's libdevice.
/llvm-project-15.0.7/polly/
H A DCMakeLists.txt87 # Do not require CUDA/OpenCL, as GPU code generation test cases can be run
88 # without a CUDA/OpenCL library.
90 FIND_PACKAGE(CUDA)
/llvm-project-15.0.7/clang/include/clang/Basic/
H A DLangOptions.def238 LANGOPT(CUDA , 1, 0, "CUDA")
247 …eFullRuntime , 1, 0, "Force to use full runtime in all constructs when offloading to CUDA devices")
248 LANGOPT(OpenMPCUDANumSMs , 32, 0, "Number of SMs for CUDA devices.")
249 LANGOPT(OpenMPCUDABlocksPerSM , 32, 0, "Number of blocks per SM for CUDA devices.")
263 LANGOPT(CUDAIsDevice , 1, 0, "compiling for CUDA device")
264 LANGOPT(CUDAAllowVariadicFunctions, 1, 0, "allowing variadic functions in CUDA device code")
270 LANGOPT(GPUDeferDiag, 1, 0, "defer host/device related diagnostic messages for CUDA/HIP")
271 …gSideOverloads, 1, 0, "always exclude wrong side overloads in overloading resolution for CUDA/HIP")
H A DDiagnosticDriverKinds.td54 def err_drv_cuda_bad_gpu_arch : Error<"unsupported CUDA gpu architecture: %0">;
57 "cannot find CUDA installation; provide its path via '--cuda-path', or pass "
58 "'-nocudainc' to build without CUDA includes">;
60 "cannot find libdevice for %0; provide path to different CUDA installation "
84 "GPU arch %0 is supported by CUDA versions between %1 and %2 (inclusive), "
85 "but installation at %3 is %4; use '--cuda-path' to specify a different CUDA "
89 "CUDA version%0 is newer than the latest%select{| partially}1 supported version %2">,
92 "CUDA version %0 is only partially supported">,
97 "mixed CUDA and HIP compilation is not supported">;
326 "NVPTX target requires CUDA 9.2 or above; CUDA %0 detected">;
[all …]
/llvm-project-15.0.7/llvm/lib/Target/NVPTX/
H A DNVPTXSubtarget.cpp58 if (TM.getDrvInterface() == NVPTX::CUDA) in hasImageHandles()
/llvm-project-15.0.7/polly/include/polly/
H A DLinkAllPasses.h59 GPURuntime Runtime = GPURuntime::CUDA);
63 GPURuntime Runtime = GPURuntime::CUDA);
/llvm-project-15.0.7/clang/lib/Headers/cuda_wrappers/
H A Dnew1 /*===---- new - CUDA wrapper for <new> -------------------------------------===
31 // and CUDA-specific macros are not available yet.

123456