<div dir="ltr">Understood, thanks for your answers,<div><br></div><div><br></div></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Tue, Jul 23, 2019 at 1:01 PM Smith, Barry F. <<a href="mailto:bsmith@mcs.anl.gov">bsmith@mcs.anl.gov</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0px 0px 0px 0.8ex;border-left:1px solid rgb(204,204,204);padding-left:1ex"><br>
Manuel,<br>
<br>
We don't have support for this yet for GPUs, it is on our list but not a priority, just utilizing GPUs well is our first priority.<br>
<br>
Barry<br>
<br>
<br>
Note what the people who give these types of presentations forget to emphasis enough is that though this techniques for some algorithms can give good benefits on GPUs utilizing lower precision on CPUs doesn't generally benefit you much.<br>
<br>
<br>
> On Jul 23, 2019, at 2:17 PM, Manuel Valera via petsc-users <<a href="mailto:petsc-users@mcs.anl.gov" target="_blank">petsc-users@mcs.anl.gov</a>> wrote:<br>
> <br>
> Hello,<br>
> <br>
> I was wondering if PETSc had some form of a low precision linear solver algorithm like seen in:<br>
> <br>
> <a href="https://www.dropbox.com/s/rv5quc3k72qdpmp/iciam_lowprec19.pdf?dl=0" rel="noreferrer" target="_blank">https://www.dropbox.com/s/rv5quc3k72qdpmp/iciam_lowprec19.pdf?dl=0</a><br>
> <br>
> I understand this treatment is coming from one of the NAG library developers, <br>
> <br>
> Thanks,<br>
> <br>
> <br>
<br>
</blockquote></div>