I’m doing a test with library cuSolver, i would like to do a batch linear solver but my data is already in devive memory. Don’t have a function like cusolverRfBatchSetupDevice to avoid do a copy from device to host memory?
Related topics
| Topic | Replies | Views | Activity | |
|---|---|---|---|---|
| cusolverRfSetupHost X cusolverRfSetup | 1 | 755 | June 12, 2015 | |
| New cuBlas function "getrfBatched" | 5 | 8596 | January 21, 2014 | |
| CuSolver refactorization | 1 | 1269 | February 3, 2015 | |
| Multi-Thread or Batched cuSolver | 2 | 2033 | May 14, 2015 | |
| Interfacing OpenACC with cublasDgetrfBatched/cusparseDgtsvStridedBatch in Fortran | 0 | 586 | April 10, 2014 | |
| Does cusolver and cublas have multiple shifting of memory between host and device?? | 2 | 494 | July 19, 2017 | |
| Solve AX=B With cuSolver | 10 | 5867 | August 18, 2015 | |
| cusolverRf: LU factorization pivoting parameters | 1 | 1741 | June 10, 2015 | |
| Solve one dense linear system in one thread block | 4 | 733 | January 20, 2017 | |
| cuSolver device API so that I can call from kernel? | 1 | 1313 | March 10, 2015 |