Equalizer
1.10.1
Parallel Rendering Framework
|
The interface definition for API-specific GPGPU handling. More...
#include <computeContext.h>
Public Member Functions | |
EQ_API | ComputeContext (Pipe *parent) |
Create a new ComputeContext for the given accelerator. | |
virtual EQ_API | ~ComputeContext () |
Destroy the ComputeContext. More... | |
Data Access | |
Pipe * | getPipe () |
const Pipe * | getPipe () const |
Methods forwarded from Pipe | |
virtual EQ_API bool | configInit ()=0 |
Initialize the ComputeContext. More... | |
virtual EQ_API void | configExit ()=0 |
De-initialize the ComputeContext. More... | |
Protected Member Functions | |
Error information. | |
EventOCommand | sendError (const uint32_t error) |
Send an error event to the application node. More... | |
The interface definition for API-specific GPGPU handling.
The ComputeContext abstracts all GPGPU API-system specific code for handling a GPU for computing purposes. Each Pipe uses one ComputeContext, which is initialized in Pipe::configInit.
Definition at line 35 of file computeContext.h.
|
virtual |
Destroy the ComputeContext.
|
pure virtual |
De-initialize the ComputeContext.
Implemented in eq::CUDAContext.
|
pure virtual |
Initialize the ComputeContext.
Implemented in eq::CUDAContext.
|
inline |
Definition at line 49 of file computeContext.h.
|
inline |
Definition at line 52 of file computeContext.h.
|
protected |
Send an error event to the application node.
error | the error message. |