[CP2K:10705] Performance of cube file writing with MPI I/O

Ondrej Marsalek ondrej.... at gmail.com
Sun Sep 9 19:04:05 UTC 2018


Hi Nico,

thanks for looking into this. I thought I could at least disable it
for now by hardcoding `parallel_write = .FALSE.` (not letting it be
overwritten based on `mpi_io`) in `pw_to_cube`, but that apparently
does not work, it fails later at `CPASSERT(checksum == 1)` (line 150
in 6.1). I'd be happy to provide benchmark numbers for a suitable test
case that show the slowdown, let me know if you'd like those.

Best regards,
Ondrej

On Sun, Sep 9, 2018 at 9:08 AM Nico Holmberg <holmbe... at gmail.com> wrote:
>
> Hi,
>
> My experiences were totally opposite, which is why I made it the default option. Give me a couple of days and I'll add an input option to disable it.
>
> BR,
>
> Nico
>
> --
> You received this message because you are subscribed to the Google Groups "cp2k" group.
> To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns... at googlegroups.com.
> To post to this group, send email to cp... at googlegroups.com.
> Visit this group at https://groups.google.com/group/cp2k.
> For more options, visit https://groups.google.com/d/optout.



More information about the CP2K-user mailing list