#include <cuda_runtime_api.h>
Go to the source code of this file.
|
namespace | olb |
| Top level namespace for all of OpenLB.
|
|
namespace | olb::gpu |
| Implementations of GPU specifics.
|
|
namespace | olb::gpu::cuda |
| Implementations of Nvidia CUDA specifics.
|
|
namespace | olb::gpu::cuda::device |
| Basic wrappers of common CUDA functions.
|
|
|
int | olb::gpu::cuda::device::getCount () |
| Return number of available devices.
|
|
void | olb::gpu::cuda::device::check () |
| Check errors.
|
|
void | olb::gpu::cuda::device::synchronize () |
| Synchronize device.
|
|
int | olb::gpu::cuda::device::get () |
| Get current device.
|
|
void | olb::gpu::cuda::device::copyToHost (void *src, void *dst, std::size_t count) |
| Copy data from device to host.
|
|
void | olb::gpu::cuda::device::copyToDevice (void *src, void *dst, std::size_t count) |
| Copy data from host to device.
|
|
template<typename T > |
T * | olb::gpu::cuda::device::malloc (std::size_t size) |
| Allocate data on device.
|
|
std::size_t | olb::gpu::cuda::device::getDevicePageSize () |
| Returns device memory page size.
|
|
template<typename T > |
std::size_t | olb::gpu::cuda::device::getPageAlignedCount (std::size_t count) |
| Returns count rounded up to be a multiple of getDevicePageSize
|
|
void | olb::gpu::cuda::device::asyncCopyToHost (Stream &stream, void *src, void *dst, std::size_t count) |
| Copy data from device to host (async)
|
|
void | olb::gpu::cuda::device::asyncCopyToDevice (Stream &stream, void *src, void *dst, std::size_t count) |
| Copy data from host to device (async)
|
|