gaspersic/triton
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

Driver: bugfixes in CUDA dynamic loading

Browse Source
This commit is contained in:
Philippe Tillet
2015-08-25 19:08:31 -04:00
parent 67a35a62bd
commit 7b77d5ae4b
4 changed files with 19 additions and 18 deletions
Download Patch File Download Diff File Expand all files Collapse all files

6
include/isaac/driver/external/CL/cl.h vendored
View File

@@ -24,11 +24,7 @@
#ifndef __OPENCL_CL_H
#define __OPENCL_CL_H
#ifdef __APPLE__
#include <OpenCL/cl_platform.h>
#else
#include <CL/cl_platform.h>
#endif
#include "isaac/driver/external/CL/cl_platform.h"
#ifdef __cplusplus
extern "C" {

7
include/isaac/wrap/clBLAS.h
View File

@@ -29,12 +29,7 @@
* the AMD GPU hardware.
*/
#if defined(__APPLE__) || defined(__MACOSX)
#include <OpenCL/cl.h>
#else
#include <CL/cl.h>
#endif
#include "isaac/driver/external/CL/cl.h"
#include "clBLAS-complex.h"
#ifdef __cplusplus

22
lib/driver/dispatch.cpp
View File

@@ -142,21 +142,21 @@ OCL_DEFINE1(cl_int, clReleaseKernel, cl_kernel)
CUDA_DEFINE1(CUresult, cuCtxDestroy_v2, CUcontext)
CUDA_DEFINE2(CUresult, cuEventCreate, CUevent *, unsigned int)
CUDA_DEFINE2(CUresult, cuDeviceGet, CUdevice *, int)
CUDA_DEFINE3(CUresult, dispatch::cuMemcpyDtoH_v2, void *, CUdeviceptr, size_t)
CUDA_DEFINE3(CUresult, cuMemcpyDtoH_v2, void *, CUdeviceptr, size_t)
CUDA_DEFINE2(CUresult, cuStreamCreate, CUstream *, unsigned int)
CUDA_DEFINE3(CUresult, cuEventElapsedTime, float *, CUevent, CUevent)
CUDA_DEFINE1(CUresult, dispatch::cuMemFree_v2, CUdeviceptr)
CUDA_DEFINE4(CUresult, dispatch::cuMemcpyDtoHAsync_v2, void *, CUdeviceptr, size_t, CUstream)
CUDA_DEFINE1(CUresult, cuMemFree_v2, CUdeviceptr)
CUDA_DEFINE4(CUresult, cuMemcpyDtoHAsync_v2, void *, CUdeviceptr, size_t, CUstream)
CUDA_DEFINE1(CUresult, cuDriverGetVersion, int *)
CUDA_DEFINE3(CUresult, cuDeviceGetName, char *, int, CUdevice)
CUDA_DEFINE4(CUresult, dispatch::cuMemcpyHtoDAsync_v2, CUdeviceptr, const void *, size_t, CUstream)
CUDA_DEFINE4(CUresult, cuMemcpyHtoDAsync_v2, CUdeviceptr, const void *, size_t, CUstream)
CUDA_DEFINE2(CUresult, cuModuleLoad, CUmodule *, const char *)
CUDA_DEFINE11(CUresult, cuLaunchKernel, CUfunction, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, unsigned int, CUstream, void **, void **)
CUDA_DEFINE1(CUresult, cuModuleUnload, CUmodule)
CUDA_DEFINE5(CUresult, cuModuleLoadDataEx, CUmodule *, const void *, unsigned int, CUjit_option *, void **)
CUDA_DEFINE3(CUresult, cuDeviceGetAttribute, int *, CUdevice_attribute, CUdevice)
CUDA_DEFINE1(CUresult, cuDeviceGetCount, int *)
CUDA_DEFINE3(CUresult, dispatch::cuMemcpyHtoD_v2, CUdeviceptr, const void *, size_t )
CUDA_DEFINE3(CUresult, cuMemcpyHtoD_v2, CUdeviceptr, const void *, size_t )
CUDA_DEFINE1(CUresult, cuInit, unsigned int)
CUDA_DEFINE2(CUresult, cuEventRecord, CUevent, CUstream)
CUDA_DEFINE3(CUresult, cuCtxCreate_v2, CUcontext *, unsigned int, CUdevice)
@@ -175,8 +175,18 @@ NVRTC_DEFINE2(nvrtcResult, nvrtcGetProgramLog, nvrtcProgram, char *)
void dispatch::release()
{
if(opencl_)
if(opencl_){
dlclose(opencl_);
opencl_ = nullptr;
}
if(cuda_){
dlclose(cuda_);
cuda_ = nullptr;
}
if(nvrtc_){
dlclose(nvrtc_);
nvrtc_ = nullptr;
}
}
void * dispatch::opencl_;

2
python/setup.py
View File

@@ -124,7 +124,7 @@ def main():
libraries += ['gnustl_shared']
#Source files
src = 'src/lib/symbolic/preset.cpp src/lib/symbolic/execute.cpp src/lib/symbolic/io.cpp src/lib/symbolic/expression.cpp src/lib/array.cpp src/lib/value_scalar.cpp src/lib/driver/backend.cpp src/lib/driver/device.cpp src/lib/driver/kernel.cpp src/lib/driver/buffer.cpp src/lib/driver/platform.cpp src/lib/driver/check.cpp src/lib/driver/program.cpp src/lib/driver/command_queue.cpp src/lib/driver/dispatch.cpp src/lib/driver/program_cache.cpp src/lib/driver/context.cpp src/lib/driver/event.cpp src/lib/driver/ndrange.cpp src/lib/driver/handle.cpp src/lib/exception/unknown_datatype.cpp src/lib/exception/operation_not_supported.cpp src/lib/profiles/presets.cpp src/lib/profiles/profiles.cpp src/lib/profiles/predictors/random_forest.cpp src/lib/kernels/templates/gemv.cpp src/lib/kernels/templates/axpy.cpp src/lib/kernels/templates/gemm.cpp src/lib/kernels/templates/ger.cpp src/lib/kernels/templates/dot.cpp src/lib/kernels/templates/base.cpp src/lib/kernels/mapped_object.cpp src/lib/kernels/stream.cpp src/lib/kernels/parse.cpp src/lib/kernels/keywords.cpp src/lib/kernels/binder.cpp src/lib/wrap/clBLAS.cpp '.split() + [os.path.join('src', 'bind', sf) for sf in ['_isaac.cpp', 'core.cpp', 'driver.cpp', 'kernels.cpp', 'exceptions.cpp']]
src = 'src/lib/exception/operation_not_supported.cpp src/lib/exception/unknown_datatype.cpp src/lib/value_scalar.cpp src/lib/driver/check.cpp src/lib/driver/ndrange.cpp src/lib/driver/platform.cpp src/lib/driver/backend.cpp src/lib/driver/program.cpp src/lib/driver/command_queue.cpp src/lib/driver/event.cpp src/lib/driver/kernel.cpp src/lib/driver/handle.cpp src/lib/driver/device.cpp src/lib/driver/program_cache.cpp src/lib/driver/buffer.cpp src/lib/driver/context.cpp src/lib/driver/dispatch.cpp src/lib/kernels/templates/axpy.cpp src/lib/kernels/templates/gemv.cpp src/lib/kernels/templates/dot.cpp src/lib/kernels/templates/base.cpp src/lib/kernels/templates/ger.cpp src/lib/kernels/templates/gemm.cpp src/lib/kernels/stream.cpp src/lib/kernels/keywords.cpp src/lib/kernels/mapped_object.cpp src/lib/kernels/binder.cpp src/lib/kernels/parse.cpp src/lib/wrap/clBLAS.cpp src/lib/profiles/predictors/random_forest.cpp src/lib/profiles/presets.cpp src/lib/profiles/profiles.cpp src/lib/symbolic/execute.cpp src/lib/symbolic/expression.cpp src/lib/symbolic/io.cpp src/lib/symbolic/preset.cpp src/lib/array.cpp '.split() + [os.path.join('src', 'bind', sf) for sf in ['_isaac.cpp', 'core.cpp', 'driver.cpp', 'kernels.cpp', 'exceptions.cpp']]
boostsrc = 'external/boost/libs/'
for s in ['numpy','python','smart_ptr','system','thread']:
src = src + [x for x in recursive_glob('external/boost/libs/' + s + '/src/','.cpp') if 'win32' not in x and 'pthread' not in x]