You cannot select more than 25 topics Topics must start with a letter or number, can include dashes ('-') and can be up to 35 characters long.
psblas3/cuda/spgpu/kernels/zdot.cu

149 lines
3.4 KiB
Plaintext

/*
* spGPU - Sparse matrices on GPU library.
*
* Copyright (C) 2010 - 2012
* Davide Barbieri - University of Rome Tor Vergata
*
* This program is free software; you can redistribute it and/or
* modify it under the terms of the GNU General Public License
* version 3 as published by the Free Software Foundation.
*
* This program is distributed in the hope that it will be useful,
* but WITHOUT ANY WARRANTY; without even the implied warranty of
* MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
* GNU General Public License for more details.
*/
#include "stdio.h"
#include "cudalang.h"
#include "cudadebug.h"
#include "cuComplex.h"
extern "C"
{
#include "core.h"
#include "vector.h"
}
//#define USE_CUBLAS
#define BLOCK_SIZE 512
//#define ASSUME_LOCK_SYNC_PARALLELISM
static __device__ cuDoubleComplex ddotReductionResult[128];
__global__ void spgpuZdot_kern(int n, cuDoubleComplex* x, cuDoubleComplex* y)
{
__shared__ cuDoubleComplex sSum[BLOCK_SIZE];
cuDoubleComplex res = make_cuDoubleComplex(0.0, 0.0);
cuDoubleComplex* lastX = x + n;
x += threadIdx.x + blockIdx.x*BLOCK_SIZE;
y += threadIdx.x + blockIdx.x*BLOCK_SIZE;
int blockOffset = gridDim.x*BLOCK_SIZE;
while (x < lastX)
{
res = cuCfma(x[0], y[0], res);
x += blockOffset;
y += blockOffset;
}
if (threadIdx.x >= 32)
sSum[threadIdx.x] = res;
__syncthreads();
// Start reduction!
if (threadIdx.x < 32)
{
for (int i=1; i<BLOCK_SIZE/32; ++i)
{
res = cuCadd(res, sSum[i*32 + threadIdx.x]);
}
//useless (because inter-warp)
#ifndef ASSUME_LOCK_SYNC_PARALLELISM
}
__syncthreads();
if (threadIdx.x < 32)
{
#endif
cuDoubleComplex* vsSum = sSum;
vsSum[threadIdx.x] = res;
if (threadIdx.x < 16) vsSum[threadIdx.x] = cuCadd(vsSum[threadIdx.x], vsSum[threadIdx.x + 16]);
__syncthreads();
if (threadIdx.x < 8) vsSum[threadIdx.x] = cuCadd(vsSum[threadIdx.x], vsSum[threadIdx.x + 8]);
__syncthreads();
if (threadIdx.x < 4) vsSum[threadIdx.x] = cuCadd(vsSum[threadIdx.x], vsSum[threadIdx.x + 4]);
__syncthreads();
if (threadIdx.x < 2) vsSum[threadIdx.x] = cuCadd(vsSum[threadIdx.x], vsSum[threadIdx.x + 2]);
__syncthreads();
if (threadIdx.x == 0)
ddotReductionResult[blockIdx.x] = cuCadd(vsSum[0], vsSum[1]);
}
}
cuDoubleComplex spgpuZdot(spgpuHandle_t handle, int n, __device cuDoubleComplex* a, __device cuDoubleComplex* b)
{
#ifdef USE_CUBLAS
cuDoubleComplex res;
cublasDdot(n,x,1,y,1,&res);
cudaDeviceSynchronize();
return res;
#else
cuDoubleComplex res = make_cuDoubleComplex(0.0, 0.0);
#if 0
int device;
cudaGetDevice(&device);
struct cudaDeviceProp prop;
cudaGetDeviceProperties(&prop,device);
int blocks = min(128, min(prop.multiProcessorCount, (n+BLOCK_SIZE-1)/BLOCK_SIZE));
#else
int blocks = min(128, min(handle->multiProcessorCount, (n+BLOCK_SIZE-1)/BLOCK_SIZE));
#endif
cuDoubleComplex tRes[128];
spgpuZdot_kern<<<blocks, BLOCK_SIZE, 0, handle->currentStream>>>(n, a, b);
cudaMemcpyFromSymbol(tRes, ddotReductionResult,blocks*sizeof(cuDoubleComplex));
for (int i=0; i<blocks; ++i)
{
res = cuCadd(res, tRes[i]);
}
cudaCheckError("CUDA error on ddot");
return res;
#endif
}
void spgpuZmdot(spgpuHandle_t handle, cuDoubleComplex* y, int n, __device cuDoubleComplex* a, __device cuDoubleComplex* b, int count, int pitch)
{
for (int i=0; i<count; ++i)
{
y[i] = spgpuZdot(handle, n, a, b);
a += pitch;
b += pitch;
}
}