On Thu, Aug 15, 2019 at 2:34 PM Smith, Barry F. <bsm...@mcs.anl.gov> wrote:

>
>   Mark,
>
>    I don't know how one uses it; we don't yet have an option in hypre.py
> to turn it on.
>

Any ideas about when this might get done?


>
>    You should just use regular PETSc matrices and vectors, not CUDA ones.
> Hypre manages all that stuff internally for itsetl.
>
>    I don't know how one knows if hypre is using the GPU or not, there are
> some Nvidia profiling tools for tracking GPU usage, perhaps you can use
> those to see if it says the GPU is being used.
>

I have just been timing it. It is very slow. 3x slower than GAMG/CPU and
20x slower than GAMG/GPU. But Hypers parameters tend to be optimized for 2D
and I have not optimized parameters. But clearly its not using GPUs.


>
>    Barry
>
>
> > On Aug 15, 2019, at 10:47 AM, Mark Adams via petsc-dev <
> petsc-dev@mcs.anl.gov> wrote:
> >
> > I have configured with Hypre on SUMMIT, with cuda, and it ran. I'm now
> trying to verify that it used GPUs (I doubt it). Any ideas on how to verify
> this? Should I use the cuda vecs and mats, or does Hypre not care. Can I
> tell hypre not to use GPUs other than configuring an non-cude PETSc? I'm
> not sure how to run a job without GPUs, but I will look into it.
> >
> > Mark
>
>

Reply via email to