Dear PETSc team, I want to call the multithreading sparse direct solver SuperLU_MT in PETSc, Could I download by "--download-superlu-mt"? Or what is a good way to support calling SuperLU_MT interface in PETSc? Thank you. Best Regards, Gang
On Sun, Oct 23, 2022 at 12:48 AM 赵刚 <[email protected]> wrote:
Dear PETSc team,
I want to call the multithreading sparse direct solver SuperLU_MT in PETSc, Could I download by "--download-superlu-mt"? Or what is a good way to support calling SuperLU_MT interface in PETSc? Thank you.
Is it a separate package, or only configure arguments to SuperLU? Thanks, Matt
Best Regards, Gang
-- What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead. -- Norbert Wiener https://www.cse.buffalo.edu/~knepley/ <http://www.cse.buffalo.edu/~knepley/>
On 23 Oct 2022, at 12:34 PM, Matthew Knepley <[email protected]> wrote:
On Sun, Oct 23, 2022 at 12:48 AM 赵刚 <[email protected] <mailto:[email protected]>> wrote: Dear PETSc team,
I want to call the multithreading sparse direct solver SuperLU_MT in PETSc, Could I download by "--download-superlu-mt"? Or what is a good way to support calling SuperLU_MT interface in PETSc? Thank you.
Is it a separate package,
Yes, one would need to add it: https://github.com/xiaoyeli/superlu_mt <https://github.com/xiaoyeli/superlu_mt> Gang, it would be a good exercise to make your first contribution to PETSc if you feel like it. You should be able to copy/paste most of what is in config/BuildSystem/config/packages/SuperLU.py into a new file config/BuildSystem/config/packages/SuperLU_MT.py Thanks, Pierre
or only configure arguments to SuperLU?
Thanks,
Matt
Best Regards, Gang
-- What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead. -- Norbert Wiener
https://www.cse.buffalo.edu/~knepley/ <http://www.cse.buffalo.edu/~knepley/>
Thank you, Pierre, I'll try to add it. Best Regards, Gang -----原始邮件----- 发件人:"Pierre Jolivet" <[email protected]> 发送时间:2022-10-23 18:55:12 (星期日) 收件人: "Matthew Knepley" <[email protected]> 抄送: "赵刚" <[email protected]>, [email protected] 主题: Re: [petsc-users] Calling SuperLU_MT in PETSc On 23 Oct 2022, at 12:34 PM, Matthew Knepley <[email protected]> wrote: On Sun, Oct 23, 2022 at 12:48 AM 赵刚 <[email protected]> wrote: Dear PETSc team, I want to call the multithreading sparse direct solver SuperLU_MT in PETSc, Could I download by "--download-superlu-mt"? Or what is a good way to support calling SuperLU_MT interface in PETSc? Thank you. Is it a separate package, Yes, one would need to add it: https://github.com/xiaoyeli/superlu_mt Gang, it would be a good exercise to make your first contribution to PETSc if you feel like it. You should be able to copy/paste most of what is in config/BuildSystem/config/packages/SuperLU.py into a new file config/BuildSystem/config/packages/SuperLU_MT.py Thanks, Pierre or only configure arguments to SuperLU? Thanks, Matt Best Regards, Gang -- What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead. -- Norbert Wiener https://www.cse.buffalo.edu/~knepley/
Yes, it's a separate multithreading-based solver package. I noticed that PETSc support SuperLU (sequential package) and SuperLU_DIST through the option "--download-superlu" and "--download-superlu_dist", but SuperLU_MT doesn't seem to exist. I'm gonna try to add it according to Pierre's suggestions and push into PETSc. Thanks, Matthew. Best Regards, Gang -----原始邮件----- 发件人:"Matthew Knepley" <[email protected]> 发送时间:2022-10-23 18:34:47 (星期日) 收件人: "赵刚" <[email protected]> 抄送: [email protected] 主题: Re: [petsc-users] Calling SuperLU_MT in PETSc On Sun, Oct 23, 2022 at 12:48 AM 赵刚 <[email protected]> wrote: Dear PETSc team, I want to call the multithreading sparse direct solver SuperLU_MT in PETSc, Could I download by "--download-superlu-mt"? Or what is a good way to support calling SuperLU_MT interface in PETSc? Thank you. Is it a separate package, or only configure arguments to SuperLU? Thanks, Matt Best Regards, Gang -- What most experimenters take for granted before they begin their experiments is infinitely more interesting than any results to which their experiments lead. -- Norbert Wiener https://www.cse.buffalo.edu/~knepley/
participants (3)
-
Matthew Knepley -
Pierre Jolivet -
赵刚