pxtask::GpuCopyDesc Struct Reference
Input descriptor for the GpuDispatcher's built-in copy kernel. More...
#include <PxGpuCopyDesc.h>
Public Types | |
enum | CopyType { HostToDevice, DeviceToHost, DeviceToDevice, DeviceMemset32 } |
Input descriptor for the GpuDispatcher's built-in copy kernel. More... | |
Public Member Functions | |
PX_INLINE bool | isValid () |
Copy is optimally performed as 64bit words, requires 64bit alignment. But it can gracefully degrade to 32bit copies if necessary. | |
Public Attributes | |
size_t | dest |
the destination | |
size_t | source |
the source (32bit value when type == DeviceMemset) | |
size_t | bytes |
the size in bytes | |
CopyType | type |
the memory transaction type |
Detailed Description
Input descriptor for the GpuDispatcher's built-in copy kernel.All host memory involved in copy transactions must be page-locked. If more than one descriptor is passed to the copy kernel in one launch, the descriptors themselves must be in page-locked memory.
Member Enumeration Documentation
Member Function Documentation
PX_INLINE bool pxtask::GpuCopyDesc::isValid | ( | ) | [inline] |
Copy is optimally performed as 64bit words, requires 64bit alignment. But it can gracefully degrade to 32bit copies if necessary.
Referenced by pxtask::GpuCopyDescQueue::enqueue().
Member Data Documentation
size_t pxtask::GpuCopyDesc::bytes |
size_t pxtask::GpuCopyDesc::dest |
the destination
size_t pxtask::GpuCopyDesc::source |
the source (32bit value when type == DeviceMemset)
the memory transaction type
The documentation for this struct was generated from the following file:
Copyright © 2008-2012 NVIDIA Corporation, 2701 San Tomas Expressway, Santa Clara, CA 95050 U.S.A. All rights reserved. www.nvidia.com