Loading...
Searching...
No Matches
tf_cuda.h File Reference
#include "tfError.h"#include <cuda.h>#include <cuda_runtime.h>#include <nvrtc.h>#include <vector>#include <stdexcept>#include <string>
Include dependency graph for tf_cuda.h:
This graph shows which files directly or indirectly include this file:
Go to the source code of this file.
Classes | |
| struct | TissueForge::cuda::CUDARTSource |
| Convenience class for loading source from file and storing, here intended for CUDA. More... | |
| struct | TissueForge::cuda::CUDARTProgram |
| A JIT-compiled CUDA Tissue Forge program. More... | |
| struct | TissueForge::cuda::CUDAFunction |
| A CUDA kernel from a JIT-compiled Tissue Forge program. More... | |
| struct | TissueForge::cuda::CUDAContext |
| A convenience wrap of the CUDA context for JIT-compiled Tissue Forge programs. More... | |
| struct | TissueForge::cuda::CUDADevice |
| A simple interface with a CUDA device. More... | |
Namespaces | |
| namespace | TissueForge |
| Include Python header, disable linking to pythonX_d.lib on Windows in debug mode. | |
| namespace | TissueForge::cuda |
| Tissue Forge GPU acceleration on CUDA-supporting devices. | |
Macros | |
| #define | TF_CUDA_CALL(res) |
| #define | TF_NVRTC_CALL(res) |
| #define | TF_CUDART_CALL(res) |
Functions | |
| CUresult | TissueForge::cuda::cuda_errorchk (CUresult retCode, const char *file, int line) |
| nvrtcResult | TissueForge::cuda::nvrtc_errorchk (nvrtcResult retCode, const char *file, int line) |
| cudaError_t | TissueForge::cuda::cudart_errorchk (cudaError_t retCode, const char *file, int line) |
| void | TissueForge::cuda::init () |
| Initialize CUDA. | |
| void | TissueForge::cuda::setGLDevice (const int &deviceId) |
| std::string | TissueForge::cuda::getDeviceName (const int &deviceId) |
| Get the name of a device. | |
| size_t | TissueForge::cuda::getDeviceTotalMem (const int &deviceId) |
| Get the total memory of device. | |
| int | TissueForge::cuda::getDeviceAttribute (const int &deviceId, const int &attrib) |
| Get the attribute value of a device. | |
| int | TissueForge::cuda::getNumDevices () |
| Get number of available compute-capable devices. | |
| std::string | TissueForge::cuda::getDevicePCIBusId (const int &deviceId) |
| Get the PCI bus id of a device. | |
| int | TissueForge::cuda::getCurrentDevice () |
| Get the device id of the current context of the calling CPU thread. | |
| int | TissueForge::cuda::maxThreadsPerBlock (const int &deviceId) |
| Maximum number of threads per block. | |
| int | TissueForge::cuda::maxBlockDimX (const int &deviceId) |
| Maximum x-dimension of a block. | |
| int | TissueForge::cuda::maxBlockDimY (const int &deviceId) |
| Maximum y-dimension of a block. | |
| int | TissueForge::cuda::maxBlockDimZ (const int &deviceId) |
| Maximum z-dimension of a block. | |
| int | TissueForge::cuda::maxGridDimX (const int &deviceId) |
| Maximum x-dimension of a grid. | |
| int | TissueForge::cuda::maxGridDimY (const int &deviceId) |
| Maximum y-dimension of a grid. | |
| int | TissueForge::cuda::maxGridDimZ (const int &deviceId) |
| Maximum z-dimension of a grid. | |
| int | TissueForge::cuda::maxSharedMemPerBlock (const int &deviceId) |
| Maximum amount of shared memory available to a thread block in bytes. | |
| int | TissueForge::cuda::maxTotalMemConst (const int &deviceId) |
| Memory available on device for constant variables in a CUDA C kernel in bytes. | |
| int | TissueForge::cuda::warpSize (const int &deviceId) |
| Warp size in threads. | |
| int | TissueForge::cuda::maxRegsPerBlock (const int &deviceId) |
| Maximum number of 32-bit registers available to a thread block. | |
| int | TissueForge::cuda::clockRate (const int &deviceId) |
| The typical clock frequency in kilohertz. | |
| bool | TissueForge::cuda::gpuOverlap (const int &deviceId) |
| Test if the device can concurrently copy memory between host and device while executing a kernel. | |
| int | TissueForge::cuda::numMultiprocessors (const int &deviceId) |
| Number of multiprocessors on the device. | |
| bool | TissueForge::cuda::kernelExecTimeout (const int &deviceId) |
| Test if there is a run time limit for kernels executed on the device. | |
| bool | TissueForge::cuda::computeModeDefault (const int &deviceId) |
| Test if device is not restricted and can have multiple CUDA contexts present at a single time. | |
| bool | TissueForge::cuda::computeModeProhibited (const int &deviceId) |
| Test if device is prohibited from creating new CUDA contexts. | |
| bool | TissueForge::cuda::computeModeExclusive (const int &deviceId) |
| Test if device can have only one context used by a single process at a time. | |
| int | TissueForge::cuda::PCIDeviceId (const int &deviceId) |
| PCI device (also known as slot) identifier of the device. | |
| int | TissueForge::cuda::PCIDomainId (const int &deviceId) |
| PCI domain identifier of the device. | |
| int | TissueForge::cuda::clockRateMem (const int &deviceId) |
| Peak memory clock frequency in kilohertz. | |
| int | TissueForge::cuda::globalMemBusWidth (const int &deviceId) |
| Global memory bus width in bits. | |
| int | TissueForge::cuda::L2CacheSize (const int &deviceId) |
| Size of L2 cache in bytes. 0 if the device doesn't have L2 cache. | |
| int | TissueForge::cuda::maxThreadsPerMultiprocessor (const int &deviceId) |
| Maximum resident threads per multiprocessor. | |
| int | TissueForge::cuda::computeCapabilityMajor (const int &deviceId) |
| Major compute capability version number. | |
| int | TissueForge::cuda::computeCapabilityMinor (const int &deviceId) |
| Minor compute capability version number. | |
| bool | TissueForge::cuda::L1CacheSupportGlobal (const int &deviceId) |
| Test if device supports caching globals in L1 cache. | |
| bool | TissueForge::cuda::L1CacheSupportLocal (const int &deviceId) |
| Test if device supports caching locals in L1 cache. | |
| int | TissueForge::cuda::maxSharedMemPerMultiprocessor (const int &deviceId) |
| Maximum amount of shared memory available to a multiprocessor in bytes. | |
| int | TissueForge::cuda::maxRegsPerMultiprocessor (const int &deviceId) |
| Maximum number of 32-bit registers available to a multiprocessor. | |
| bool | TissueForge::cuda::managedMem (const int &deviceId) |
| Test if device supports allocating managed memory on this system. | |
| bool | TissueForge::cuda::multiGPUBoard (const int &deviceId) |
| Test if device is on a multi-GPU board. | |
| int | TissueForge::cuda::multiGPUBoardGroupId (const int &deviceId) |
| Unique identifier for a group of devices associated with the same board. | |
| void | TissueForge::cuda::test (const int &numBlocks, const int &numThreads, const int &numEls, const int &deviceId=0) |
| Tests JIT-compiled program execution and deployment. | |
| std::string | TissueForge::cuda::tfIncludePath () |
| Returns the current path to the installed Tissue Forge include directory. | |
| HRESULT | TissueForge::cuda::setTfIncludePath (const std::string &_path) |
| Set the current path to the installed Tissue Forge include directory. | |
| std::string | TissueForge::cuda::tfPrivateIncludePath () |
| Returns the path to the installed Tissue Forge private include directory. | |
| std::string | TissueForge::cuda::tfResourcePath () |
| Returns the current path to the installed Tissue Forge resource directory. | |
| HRESULT | TissueForge::cuda::setTfResourcePath (const std::string &_path) |
| Set the current path to the installed Tissue Forge resource directory. | |
| std::string | TissueForge::cuda::CUDAPath () |
| Returns the path to the installed Tissue Forge CUDA resources directory. | |
| std::string | TissueForge::cuda::CUDAIncludePath () |
| Returns the current path to the installed CUDA include directory. | |
| HRESULT | TissueForge::cuda::setCUDAIncludePath (const std::string &_path) |
| Set the current path to the installed CUDA include directory. | |
| std::string | TissueForge::cuda::CUDAResourcePath (const std::string &relativePath) |
| Returns an absolute path to a subdirectory of the install Tissue Forge CUDA resources directory. | |
| std::string | TissueForge::cuda::CUDAPTXObjectRelPath () |
| Returns the relative path to the installed Tissue Forge CUDA PTX object directory. | |
| std::vector< std::string > | TissueForge::cuda::CUDAArchs () |
| Returns the supported CUDA architectures of the installation. | |
Macro Definition Documentation
◆ TF_CUDA_CALL
| #define TF_CUDA_CALL | ( | res | ) |
Value:
cuda_errorchk(res, __FILE__, __LINE__)
◆ TF_CUDART_CALL
| #define TF_CUDART_CALL | ( | res | ) |
Value:
cudart_errorchk(res, __FILE__, __LINE__)
◆ TF_NVRTC_CALL
| #define TF_NVRTC_CALL | ( | res | ) |
Value:
nvrtc_errorchk(res, __FILE__, __LINE__)
