|
Grid 0.7.0
|
Go to the source code of this file.
Functions | |
| NAMESPACE_BEGIN (Grid) | |
| template<class Iterator> | |
| unsigned int | nextPow2 (Iterator x) |
| template<class Iterator> | |
| int | getNumBlocksAndThreads (const Iterator n, const size_t sizeofsobj, Iterator &threads, Iterator &blocks) |
| template<class sobj, class Iterator> | |
| __device__ void | reduceBlock (volatile sobj *sdata, sobj mySum, const Iterator tid) |
| template<class vobj, class sobj, class Iterator> | |
| __device__ void | reduceBlocks (const vobj *g_idata, sobj *g_odata, Iterator n) |
| template<class vobj, class sobj, class Iterator> | |
| __global__ void | reduceKernel (const vobj *lat, sobj *buffer, Iterator n) |
| template<class vobj> | |
| vobj::scalar_objectD | sumD_gpu_small (const vobj *lat, Integer osites) |
| template<class vobj> | |
| vobj::scalar_objectD | sumD_gpu_large (const vobj *lat, Integer osites) |
| template<class vobj> | |
| vobj::scalar_objectD | sumD_gpu (const vobj *lat, Integer osites) |
| template<class vobj> | |
| vobj::scalar_object | sum_gpu (const vobj *lat, Integer osites) |
| template<class vobj> | |
| vobj::scalar_object | sum_gpu_large (const vobj *lat, Integer osites) |
| NAMESPACE_END (Grid) | |
Variables | |
| __device__ unsigned int | retirementCount |
| NAMESPACE_BEGIN | ( | Grid | ) |
| unsigned int nextPow2 | ( | Iterator | x | ) |
Definition at line 15 of file Lattice_reduction_gpu.h.
Referenced by getNumBlocksAndThreads().
| int getNumBlocksAndThreads | ( | const Iterator | n, |
| const size_t | sizeofsobj, | ||
| Iterator & | threads, | ||
| Iterator & | blocks ) |
Definition at line 26 of file Lattice_reduction_gpu.h.
References GridLogError, and nextPow2().
Referenced by sumD_gpu(), and sumD_gpu_small().
| __device__ void reduceBlock | ( | volatile sobj * | sdata, |
| sobj | mySum, | ||
| const Iterator | tid ) |
Definition at line 65 of file Lattice_reduction_gpu.h.
References acceleratorSynchronise(), and acceleratorSynchroniseAll().
Referenced by reduceBlocks(), and reduceKernel().
| __device__ void reduceBlocks | ( | const vobj * | g_idata, |
| sobj * | g_odata, | ||
| Iterator | n ) |
Definition at line 102 of file Lattice_reduction_gpu.h.
References COALESCE_GRANULARITY, extractLane(), and reduceBlock().
Referenced by reduceKernel().
| __global__ void reduceKernel | ( | const vobj * | lat, |
| sobj * | buffer, | ||
| Iterator | n ) |
Definition at line 146 of file Lattice_reduction_gpu.h.
References acceleratorFence(), acceleratorSynchroniseAll(), COALESCE_GRANULARITY, reduceBlock(), reduceBlocks(), and retirementCount.
Referenced by sumD_gpu_small().
|
inline |
Definition at line 201 of file Lattice_reduction_gpu.h.
References accelerator_barrier, acceleratorCopyFromDevice(), getNumBlocksAndThreads(), and reduceKernel().
Referenced by sumD_gpu(), and sumD_gpu_large().
|
inline |
Definition at line 227 of file Lattice_reduction_gpu.h.
References accelerator_for, and sumD_gpu_small().
Referenced by sum_gpu_large(), sumD_gpu(), and sumD_large().
|
inline |
Definition at line 253 of file Lattice_reduction_gpu.h.
References getNumBlocksAndThreads(), sumD_gpu_large(), and sumD_gpu_small().
|
inline |
Definition at line 275 of file Lattice_reduction_gpu.h.
References sumD_gpu().
|
inline |
Definition at line 284 of file Lattice_reduction_gpu.h.
References sumD_gpu_large().
Referenced by rankSumLarge().
| NAMESPACE_END | ( | Grid | ) |
| __device__ unsigned int retirementCount |
Definition at line 12 of file Lattice_reduction_gpu.h.
Referenced by reduceKernel().