<div dir="ltr"><div class="gmail_default" style="font-family:verdana,sans-serif;font-size:small;color:#000000">Regarding MKL, superlu only uses some BLAS routines in MKL. If you have alternative BLAS, you don't need to use MKL.</div><div class="gmail_default" style="font-family:verdana,sans-serif;font-size:small;color:#000000"><br></div><div class="gmail_default" style="font-family:verdana,sans-serif;font-size:small;color:#000000">But, in my experience, on an Intel system, the compiler can correctly include the path with mkl.h, that is, I don't need to do anything special. </div><div class="gmail_default" style="font-family:verdana,sans-serif;font-size:small;color:#000000">I think mpiicc should work.</div><div class="gmail_default" style="font-family:verdana,sans-serif;font-size:small;color:#000000"><br></div><div class="gmail_default" style="font-family:verdana,sans-serif;font-size:small;color:#000000">Sherry</div></div><br><div class="gmail_quote"><div dir="ltr" class="gmail_attr">On Mon, Aug 3, 2020 at 8:46 AM <<a href="mailto:flw@rzg.mpg.de">flw@rzg.mpg.de</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">Hi Barry,<br>
Thanks for the branch (and thanks to Sherry as well). I tried to use <br>
the configure example " arch-ci-linux-intel-mkl-single.py" (adding the <br>
--with-batch flag, since I am running on a cluster), but I get the <br>
following error message:<br>
TESTING: check from <br>
config.libraries(config/BuildSystem/config/libraries.py:157) <br>
<br>
<br>
<br>
*******************************************************************************<br>
UNABLE to CONFIGURE with GIVEN OPTIONS (see configure.log <br>
for details):<br>
-------------------------------------------------------------------------------<br>
Downloaded ptscotch could not be used. Please check install in <br>
/u/flw/petsc-hash-pkgs/073aec050<br>
<br>
(see configure1.log)<br>
<br>
Next I tried a minimalistic version using Intel MPI compilers:<br>
./configure --download-superlu_dist --download-metis <br>
--download-parmetis --download-ptscotch --with-precision=single <br>
--with-batch --with-fc=mpiifort --with-cc=mpiicc --with-cxx=mpiicpc<br>
<br>
There I got the following error:<br>
=============================================================================== Compiling and installing SUPERLU_DIST; this may take several minutes =============================================================================== <br>
*******************************************************************************<br>
UNABLE to CONFIGURE with GIVEN OPTIONS (see configure.log <br>
for details):<br>
-------------------------------------------------------------------------------<br>
Error running make on SUPERLU_DIST<br>
<br>
(see configure2.log)<br>
<br>
<br>
Next I tried<br>
./configure --download-superlu_dist --download-metis <br>
--download-parmetis --download-ptscotch --with-precision=single <br>
--with-batch <br>
--with-mpi-dir=/mpcdf/soft/SLE_12/packages/x86_64/intel_parallel_studio/2018.4/impi/2018.4.274/intel64<br>
<br>
Same error message (see configure3.log)<br>
<br>
<br>
It seems that there is something going on with mkl in case I want to <br>
use Intel compilers for C and C++ (compiling with gcc and g++ seems to <br>
work )<br>
<br>
Do you know what is going on there? (I am running on the Draco <br>
cluster, if that helps <br>
(<a href="https://www.mpcdf.mpg.de/services/computing/draco" rel="noreferrer" target="_blank">https://www.mpcdf.mpg.de/services/computing/draco</a>))<br>
<br>
<br>
Best regards,<br>
Felix<br>
<br>
Zitat von Barry Smith <<a href="mailto:bsmith@petsc.dev" target="_blank">bsmith@petsc.dev</a>>:<br>
<br>
> Felix,<br>
><br>
> The branch is ready. Just use<br>
><br>
> git checkout barry/2020-07-28/superlu_dist-single<br>
> ./configure --download-superlu_dist --download-metis <br>
> --download-parmetis --download-ptscotch --with-precision=single <br>
> and whatever else you use<br>
><br>
> Barry<br>
><br>
> It will automatically get the needed branch of SuperLU_DIST that <br>
> Sherry prepared.<br>
><br>
><br>
>> On Jul 27, 2020, at 2:10 PM, <a href="mailto:flw@rzg.mpg.de" target="_blank">flw@rzg.mpg.de</a> wrote:<br>
>><br>
>> Hi Shery,<br>
>> Yes, ideally we would like to compile PETSc in single precision and <br>
>> simply run a single precision version of SUPERLU_DIST just like <br>
>> e.g. MUMPS.<br>
>><br>
>> Best regards and thanks,<br>
>> Felix<br>
>> Zitat von "Xiaoye S. Li" <<a href="mailto:xsli@lbl.gov" target="_blank">xsli@lbl.gov</a>>:<br>
>><br>
>>> Barry,<br>
>>><br>
>>> I have a branch 'Mixed-precision' working with single precision FP32. I<br>
>>> assume Felix wants to use superlu_dist from petsc. How do you want to<br>
>>> incorporate it in petsc?<br>
>>><br>
>>> <a href="https://github.com/xiaoyeli/superlu_dist" rel="noreferrer" target="_blank">https://github.com/xiaoyeli/superlu_dist</a><br>
>>><br>
>>> PS1: in this version, FP32 only works on CPU. FP64 and complex-FP64 all<br>
>>> work on GPU.<br>
>>><br>
>>> PS2: currently there is no mixed-precision yet, but it is the branch we are<br>
>>> adding mix-prec support. Will take a while before merging to master.<br>
>>><br>
>>> Sherry<br>
>>><br>
>>><br>
>>> On Wed, Jul 22, 2020 at 6:04 AM <<a href="mailto:flw@rzg.mpg.de" target="_blank">flw@rzg.mpg.de</a>> wrote:<br>
>>><br>
>>>> Hi Barry,<br>
>>>> for now I just want to run everything in single on CPUs only with<br>
>>>> SUPERLU_DIST. Maybe we will also incorporate GPUs in the future, but<br>
>>>> there are no immediate plans yet. So if you could provide the support,<br>
>>>> that would be awesome.<br>
>>>><br>
>>>> Best regards,<br>
>>>> Felix<br>
>>>><br>
>>>> Zitat von Barry Smith <<a href="mailto:bsmith@petsc.dev" target="_blank">bsmith@petsc.dev</a>>:<br>
>>>><br>
>>>> > Felix,<br>
>>>> ><br>
>>>> > What are your needs, do you want this for CPUs or for GPUs? Do<br>
>>>> > you wish to run all your code in single precision or just the<br>
>>>> > SuperLU_Dist solver while the rest of your code double?<br>
>>>> ><br>
>>>> > If you want to run everything on CPUs using single precision<br>
>>>> > then adding the support is very easy, we can provide that for you<br>
>>>> > any time. The other cases will require more thought.<br>
>>>> ><br>
>>>> > Barry<br>
>>>> ><br>
>>>> ><br>
>>>> >> On Jul 21, 2020, at 8:58 AM, <a href="mailto:flw@rzg.mpg.de" target="_blank">flw@rzg.mpg.de</a> wrote:<br>
>>>> >><br>
>>>> >> Dear PETSc support team,<br>
>>>> >> some time ago you told me that you are planning on releasing a<br>
>>>> >> version that supports SUPERLU_DIST in single-precision soon. Can<br>
>>>> >> you tell me roughly what time frame you had in mind?<br>
>>>> >><br>
>>>> >> Best regards,<br>
>>>> >> Felix<br>
>>>> >><br>
>>>><br>
>>>><br>
>>>><br>
>>>><br>
>><br>
>><br>
>><br>
<br>
<br>
</blockquote></div>