p61743098
/
sagecal
1
0
Fork 0
Code Issues Pull Requests Releases Wiki Activity

original blocks=2*(..)/threads reverting to blocks=(..)/threads

This commit is contained in:
Sarod Yatawatta 2018-02-06 21:08:03 +01:00
parent 829b7ea072
commit 1c2cfa3f7b
1 changed files with 8 additions and 8 deletions
Show Stats Download Patch File Download Diff File Expand all files Collapse all files

16
src/lib/Dirac/clmfit.c
View File

@ -146,7 +146,7 @@ clevmar_der_single(
/* calculate no of cuda threads and blocks */ /* calculate no of cuda threads and blocks */
int ThreadsPerBlock=DEFAULT_TH_PER_BK; int ThreadsPerBlock=DEFAULT_TH_PER_BK;
int BlocksPerGrid= 2*(M+ThreadsPerBlock-1)/ThreadsPerBlock; int BlocksPerGrid=(M+ThreadsPerBlock-1)/ThreadsPerBlock;
err=cudaSetDevice(card); err=cudaSetDevice(card);
checkCudaError(err,__FILE__,__LINE__); checkCudaError(err,__FILE__,__LINE__);
@ -199,7 +199,7 @@ clevmar_der_single(
checkCudaError(err,__FILE__,__LINE__); checkCudaError(err,__FILE__,__LINE__);
/* memory allocation: different dirac */ /* memory allocation: different solvers */
int work_size=0; int work_size=0;
int *devInfo; int *devInfo;
int devInfo_h=0; int devInfo_h=0;
@ -709,7 +709,7 @@ clevmar_der_single_cuda(
/* calculate no of cuda threads and blocks */ /* calculate no of cuda threads and blocks */
int ThreadsPerBlock=DEFAULT_TH_PER_BK; int ThreadsPerBlock=DEFAULT_TH_PER_BK;
int BlocksPerGrid= 2*(M+ThreadsPerBlock-1)/ThreadsPerBlock; int BlocksPerGrid=(M+ThreadsPerBlock-1)/ThreadsPerBlock;
unsigned long int moff; /* make sure offsets are multiples of 4 */ unsigned long int moff; /* make sure offsets are multiples of 4 */
@ -742,7 +742,7 @@ clevmar_der_single_cuda(
checkCudaError(err,__FILE__,__LINE__); checkCudaError(err,__FILE__,__LINE__);
err=cudaMalloc((void**)&ed, N*sizeof(double)); err=cudaMalloc((void**)&ed, N*sizeof(double));
checkCudaError(err,__FILE__,__LINE__); checkCudaError(err,__FILE__,__LINE__);
/* memory allocation: different dirac */ /* memory allocation: different solvers */
if (solve_axb==1) { if (solve_axb==1) {
err=cudaMalloc((void**)&taud, M*sizeof(double)); err=cudaMalloc((void**)&taud, M*sizeof(double));
checkCudaError(err,__FILE__,__LINE__); checkCudaError(err,__FILE__,__LINE__);
@ -1256,11 +1256,11 @@ attach_gpu_to_thread2(int card, cublasHandle_t *cbhandle, cusolverDnHandle_t *
if (usecula) { if (usecula) {
status=cusolverDnCreate(solver_handle); status=cusolverDnCreate(solver_handle);
if (status != CUSOLVER_STATUS_SUCCESS) { if (status != CUSOLVER_STATUS_SUCCESS) {
fprintf(stderr,"%s: %d: CUSOLV create fail %d\n",__FILE__,__LINE__,status); fprintf(stderr,"%s: %d: CUSOLV create fail card %d, %d\n",__FILE__,__LINE__,card,status);
sleep(10); sleep(10);
status=cusolverDnCreate(solver_handle); status=cusolverDnCreate(solver_handle);
if (status != CUSOLVER_STATUS_SUCCESS) { if (status != CUSOLVER_STATUS_SUCCESS) {
fprintf(stderr,"%s: %d: CUSOLV create fail %d\n",__FILE__,__LINE__,status); fprintf(stderr,"%s: %d: CUSOLV create fail card %d, %d\n",__FILE__,__LINE__,card,status);
exit(1); exit(1);
} }
} }
@ -1378,7 +1378,7 @@ mlm_der_single_cuda(
/* calculate no of cuda threads and blocks */ /* calculate no of cuda threads and blocks */
int ThreadsPerBlock=DEFAULT_TH_PER_BK; int ThreadsPerBlock=DEFAULT_TH_PER_BK;
int BlocksPerGrid= 2*(M+ThreadsPerBlock-1)/ThreadsPerBlock; int BlocksPerGrid=(M+ThreadsPerBlock-1)/ThreadsPerBlock;
if (opts) { if (opts) {
@ -1435,7 +1435,7 @@ mlm_der_single_cuda(
/* we need coherencies for only this cluster */ /* we need coherencies for only this cluster */
err=cudaMalloc((void**) &cohd, Nbase*8*sizeof(double)); err=cudaMalloc((void**) &cohd, Nbase*8*sizeof(double));
checkCudaError(err,__FILE__,__LINE__); checkCudaError(err,__FILE__,__LINE__);
/* memory allocation: different dirac */ /* memory allocation: different solvers */
if (solve_axb==1) { if (solve_axb==1) {
/* QR solver ********************************/ /* QR solver ********************************/
err=cudaMalloc((void**)&taud, M*sizeof(double)); err=cudaMalloc((void**)&taud, M*sizeof(double));