pxtask Namespace Reference
[Physics]
Classes | |
class | CpuDispatcher |
A CpuDispatcher is responsible for scheduling the execution of tasks passed to it by the SDK. More... | |
struct | CudaInteropMode |
Possible graphic/CUDA interoperability modes for context. More... | |
class | CudaContextManagerDesc |
Descriptor used to create a CudaContextManager. More... | |
class | CudaContextManager |
Manages memory, thread locks, and task scheduling for a CUDA context. More... | |
class | ScopedCudaLock |
Convenience class for holding CUDA lock within a scope. More... | |
struct | AllocId |
ID of the Feature which owns/allocated memory from the heap. More... | |
struct | CudaBufferMemorySpace |
memory type managed by a heap More... | |
class | AllocInfo |
class to track allocation statistics, see PxgMirrored More... | |
struct | AllocIdStats |
statistics collected per AllocationId by HeapManager. More... | |
struct | CudaBufferFlags |
Hint flag to tell how the buffer will be used. More... | |
struct | CudaMemoryManagerStats |
Memory statistics struct returned by CudaMemMgr::getStats(). More... | |
struct | CudaBufferType |
Buffer type: made of hint flags and the memory space (Device Memory, Pinned Host Memory, ...). More... | |
class | NvCudaBuffer |
Buffer which keeps informations about allocated piece of memory. More... | |
class | CudaMemoryManager |
Allocator class for different kinds of CUDA related memory. More... | |
struct | GpuCopyDesc |
Input descriptor for the GpuDispatcher's built-in copy kernel. More... | |
class | GpuCopyDescQueue |
Container class for queueing GpuCopyDesc instances in pinned (non-pageable) CPU memory. More... | |
class | GpuDispatcher |
A GpuTask dispatcher. More... | |
struct | GpuTaskHint |
Define the 'flavor' of a GpuTask. More... | |
class | GpuTask |
Task implementation for launching CUDA work. More... | |
class | SpuDispatcher |
A SpuDispatcher. More... | |
class | SpuTask |
A task to be executed on one or more SPUs. More... | |
class | BaseTask |
Base class of all task types. More... | |
class | Task |
A BaseTask implementation with deferred execution and full dependencies. More... | |
class | LightCpuTask |
A BaseTask implementation with immediate execution and simple dependencies. More... | |
struct | TaskType |
Identifies the type of each heavyweight Task object. More... | |
class | TaskManager |
The TaskManager interface. More... | |
Typedefs | |
typedef size_t | CudaBufferPtr |
typedef unsigned long | TaskID |
Functions | |
int | getSuggestedCudaDeviceOrdinal (PxErrorCallback &errc) |
Ask the NVIDIA control panel which GPU has been selected for use by PhysX. Returns -1 if no PhysX capable GPU is found or GPU PhysX has been disabled. | |
CudaContextManager * | createCudaContextManager (PxFoundation &foundation, const CudaContextManagerDesc &desc, physx::PxProfileZoneManager *mgr) |
Allocate a CUDA Context manager, complete with heaps and task dispatcher. You only need one CUDA context manager per GPU device you intend to use for CUDA tasks. If mgr is NULL, no profiling of CUDA code will be possible. | |
void * | loadPhysxGPUModule () |
get handle of physx GPU module |
Typedef Documentation
typedef size_t pxtask::CudaBufferPtr |
typedef unsigned long pxtask::TaskID |
Function Documentation
CudaContextManager* pxtask::createCudaContextManager | ( | PxFoundation & | foundation, | |
const CudaContextManagerDesc & | desc, | |||
physx::PxProfileZoneManager * | mgr | |||
) |
Allocate a CUDA Context manager, complete with heaps and task dispatcher. You only need one CUDA context manager per GPU device you intend to use for CUDA tasks. If mgr is NULL, no profiling of CUDA code will be possible.
int pxtask::getSuggestedCudaDeviceOrdinal | ( | PxErrorCallback & | errc | ) |
Ask the NVIDIA control panel which GPU has been selected for use by PhysX. Returns -1 if no PhysX capable GPU is found or GPU PhysX has been disabled.
void* pxtask::loadPhysxGPUModule | ( | ) |
get handle of physx GPU module
Copyright © 2008-2012 NVIDIA Corporation, 2701 San Tomas Expressway, Santa Clara, CA 95050 U.S.A. All rights reserved. www.nvidia.com