Skip to content

Allows to configure CUDA sm level for Cling CUDA tests #7190

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
vgvassilev merged 1 commit into from
Feb 18, 2021
Merged

Allows to configure CUDA sm level for Cling CUDA tests #7190

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
2 changes: 1 addition & 1 deletion interpreter/cling/test/CUDADeviceCode/CUDADefineArg.C
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,7 +9,7 @@

// The Test checks whether a define argument (-DTEST=3) is passed to the PTX
// compiler. If it works, it should not throw an error.
// RUN: cat %s | %cling -DTEST=3 -x cuda --cuda-path=%cudapath -Xclang -verify 2>&1 | FileCheck %s
// RUN: cat %s | %cling -DTEST=3 -x cuda --cuda-path=%cudapath %cudasmlevel -Xclang -verify 2>&1 | FileCheck %s
// REQUIRES: cuda-runtime

#include <iostream>
Expand Down
2 changes: 1 addition & 1 deletion interpreter/cling/test/CUDADeviceCode/CUDAHostPrefix.C
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,7 +9,7 @@

// The Test checks if a function with __host__ and __device__ prefix available
// on host and device side.
// RUN: cat %s | %cling -x cuda --cuda-path=%cudapath -Xclang -verify 2>&1 | FileCheck %s
// RUN: cat %s | %cling -x cuda --cuda-path=%cudapath %cudasmlevel -Xclang -verify 2>&1 | FileCheck %s
// REQUIRES: cuda-runtime

.rawInput 1
Expand Down
2 changes: 1 addition & 1 deletion interpreter/cling/test/CUDADeviceCode/CUDAInclude.C
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,7 +9,7 @@

// The test checks whether setting a new include path at runtime also works for
// the PTX compiler.
// RUN: cat %s | %cling -DTEST_PATH="\"%/p/\"" -x cuda --cuda-path=%cudapath -Xclang -verify 2>&1 | FileCheck %s
// RUN: cat %s | %cling -DTEST_PATH="\"%/p/\"" -x cuda --cuda-path=%cudapath %cudasmlevel -Xclang -verify 2>&1 | FileCheck %s
// REQUIRES: cuda-runtime

#include <iostream>
Expand Down
2 changes: 1 addition & 1 deletion interpreter/cling/test/CUDADeviceCode/CUDAKernelArgument.C
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,7 +9,7 @@

// The Test checks if a CUDA kernel works with a arguments and built-in
// functions.
// RUN: cat %s | %cling -x cuda --cuda-path=%cudapath -Xclang -verify 2>&1 | FileCheck %s
// RUN: cat %s | %cling -x cuda --cuda-path=%cudapath %cudasmlevel -Xclang -verify 2>&1 | FileCheck %s
// REQUIRES: cuda-runtime

// Test, if a simple kernel with arguments works.
Expand Down
2 changes: 1 addition & 1 deletion interpreter/cling/test/CUDADeviceCode/CUDAKernelTemplateComplex.C
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8,7 +8,7 @@
//------------------------------------------------------------------------------

// The Test checks if templated CUDA kernel in some special cases works.
// RUN: cat %s | %cling -x cuda --cuda-path=%cudapath -Xclang -verify 2>&1 | FileCheck %s
// RUN: cat %s | %cling -x cuda --cuda-path=%cudapath %cudasmlevel -Xclang -verify 2>&1 | FileCheck %s
// REQUIRES: cuda-runtime

// Check if templated CUDA kernel works, without explicit template type declaration.
Expand Down
2 changes: 1 addition & 1 deletion interpreter/cling/test/CUDADeviceCode/CUDAKernelTemplateSimple.C
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8,7 +8,7 @@
//------------------------------------------------------------------------------

// The Test checks if templated CUDA kernel works.
// RUN: cat %s | %cling -x cuda --cuda-path=%cudapath -Xclang -verify 2>&1 | FileCheck %s
// RUN: cat %s | %cling -x cuda --cuda-path=%cudapath %cudasmlevel -Xclang -verify 2>&1 | FileCheck %s
// REQUIRES: cuda-runtime

// Check if template device side resoultion works.
Expand Down
2 changes: 1 addition & 1 deletion interpreter/cling/test/CUDADeviceCode/CUDARegression.C
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,7 +9,7 @@

// The test checks if the interface functions process(), declare() and parse()
// of cling::Interpreter also work in the cuda mode.
// RUN: cat %s | %cling -DTEST_PATH="\"%/p/\"" -x cuda --cuda-path=%cudapath -Xclang -verify 2>&1 | FileCheck %s
// RUN: cat %s | %cling -DTEST_PATH="\"%/p/\"" -x cuda --cuda-path=%cudapath %cudasmlevel -Xclang -verify 2>&1 | FileCheck %s
// REQUIRES: cuda-runtime

#include "cling/Interpreter/Interpreter.h"
Expand Down
2 changes: 1 addition & 1 deletion interpreter/cling/test/CUDADeviceCode/CUDASharedMemory.C
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8,7 +8,7 @@
//------------------------------------------------------------------------------

// The Test checks if runtime shared memory works.
// RUN: cat %s | %cling -x cuda --cuda-path=%cudapath -Xclang -verify 2>&1 | FileCheck %s
// RUN: cat %s | %cling -x cuda --cuda-path=%cudapath %cudasmlevel -Xclang -verify 2>&1 | FileCheck %s
// REQUIRES: cuda-runtime

const unsigned int numberOfThreads = 4;
Expand Down
2 changes: 1 addition & 1 deletion interpreter/cling/test/CUDADeviceCode/CUDASimpleKernel.C
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,7 +9,7 @@

// The Test checks if a CUDA compatible device is available and checks, if simple
// __global__ and __device__ kernels are running.
// RUN: cat %s | %cling -x cuda --cuda-path=%cudapath -Xclang -verify 2>&1 | FileCheck %s
// RUN: cat %s | %cling -x cuda --cuda-path=%cudapath %cudasmlevel -Xclang -verify 2>&1 | FileCheck %s
// REQUIRES: cuda-runtime

// Check if cuda driver is available
Expand Down
2 changes: 1 addition & 1 deletion interpreter/cling/test/CUDADeviceCode/CUDAStreams.C
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -8,7 +8,7 @@
//------------------------------------------------------------------------------

// The Test checks if cuda streams works.
// RUN: cat %s | %cling -x cuda --cuda-path=%cudapath -Xclang -verify 2>&1 | FileCheck %s
// RUN: cat %s | %cling -x cuda --cuda-path=%cudapath %cudasmlevel -Xclang -verify 2>&1 | FileCheck %s
// REQUIRES: cuda-runtime

const unsigned int numberOfThreads = 4;
Expand Down
2 changes: 1 addition & 1 deletion interpreter/cling/test/CodeGeneration/CUDACtorDtor.C
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,7 +9,7 @@
// The Test checks, if the symbols __cuda_module_ctor and __cuda_module_dtor are
// unique for every module. Attention, for a working test case, a cuda
// fatbinary is necessary.
// RUN: cat %s | %cling -x cuda --cuda-path=%cudapath -Xclang -verify 2>&1 | FileCheck %s
// RUN: cat %s | %cling -x cuda --cuda-path=%cudapath %cudasmlevel -Xclang -verify 2>&1 | FileCheck %s
// REQUIRES: cuda-runtime

#include "cling/Interpreter/Interpreter.h"
Expand Down
4 changes: 2 additions & 2 deletions interpreter/cling/test/Driver/CUDAMode.C
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -9,8 +9,8 @@

// The Test starts the cling with the arg -xcuda and checks if the cuda mode
// is enabled.
// RUN: cat %s | %cling -x cuda --cuda-path=%cudapath -Xclang -verify 2>&1 | FileCheck %s
// RUN: cat %s | %cling -x cuda --cuda-path=%cudapath -fsyntax-only -Xclang -verify 2>&1
// RUN: cat %s | %cling -x cuda --cuda-path=%cudapath %cudasmlevel -Xclang -verify 2>&1 | FileCheck %s
// RUN: cat %s | %cling -x cuda --cuda-path=%cudapath %cudasmlevel -fsyntax-only -Xclang -verify 2>&1
// REQUIRES: cuda-runtime


Expand Down
6 changes: 6 additions & 0 deletions interpreter/cling/test/lit.cfg
View file
Open in desktop
Original file line number Diff line number Diff line change
Expand Up @@ -344,6 +344,12 @@ if libcudart_path is not None:
# https://developer.nvidia.com/blog/cuda-pro-tip-control-gpu-visibility-cuda_visible_devices/
if 'CUDA_VISIBLE_DEVICES' in os.environ:
config.environment['CUDA_VISIBLE_DEVICES'] = os.environ['CUDA_VISIBLE_DEVICES']
# set the CUDA sm level for the tests (e.g. "export CLING_TEST_CUDA_SM_LEVEL=35")
# if no sm level is set, the default of cling will be used
if 'CLING_TEST_CUDA_SM_LEVEL' in os.environ:
config.substitutions.append(('%cudasmlevel', '--cuda-gpu-arch=sm_' + os.environ['CLING_TEST_CUDA_SM_LEVEL']))
else:
config.substitutions.append(('%cudasmlevel', ""))

# Loadable module
# FIXME: This should be supplied by Makefile or autoconf.
Expand Down