<div dir="auto">I confess to being rather doubtful that gfortran+... is comparable to ifort+... for Intel cpu, it might be for AMD. While the mkl vector libraries are useful in a few codes such as aim, they are minor for the main lapw[0-2].</div><br><div class="gmail_quote"><div dir="ltr">On Wed, Apr 4, 2018, 10:55 Pavel Ondračka <<a href="mailto:pavel.ondracka@email.cz">pavel.ondracka@email.cz</a>> wrote:<br></div><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex">Rui Costa píše v St 04. 04. 2018 v 14:21 +0100:<br>
> I will see what I can do about the Intel compilers. I've had a<br>
> question about this, supposedly the intel compilers are the fastest<br>
> [<a href="https://urldefense.proofpoint.com/v2/url?u=https-3A__www.mail-2D&d=DwIGaQ&c=yHlS04HhBraes5BQ9ueu5zKhE7rtNXt_d012z2PA6ws&r=U_T4PL6jwANfAy4rnxTj8IUxm818jnvqKFdqWLwmqg0&m=_NY1hgzBYHxYTFCc6OKMrgmlmLPdt3b__vsdXAoQC14&s=M8g1ileELz5aImqKU7vYPAV3_-54VZ7qLYURtbQMg2o&e=" rel="noreferrer noreferrer" target="_blank">https://urldefense.proofpoint.com/v2/url?u=https-3A__www.mail-2D&d=DwIGaQ&c=yHlS04HhBraes5BQ9ueu5zKhE7rtNXt_d012z2PA6ws&r=U_T4PL6jwANfAy4rnxTj8IUxm818jnvqKFdqWLwmqg0&m=_NY1hgzBYHxYTFCc6OKMrgmlmLPdt3b__vsdXAoQC14&s=M8g1ileELz5aImqKU7vYPAV3_-54VZ7qLYURtbQMg2o&e=</a><br>
> <a href="http://archive.com/wien@zeus.theochem.tuwien.ac.at/msg13021.html" rel="noreferrer noreferrer" target="_blank">archive.com/wien@zeus.theochem.tuwien.ac.at/msg13021.html</a>], but how<br>
> much faster are they than the others? I expect this to vary from case<br>
> to case but on average, how much faster are they?<br>
<br>
In fact the compiler (e.g. ifort vs gfortran) hardly makes a difference<br>
. The important part are the algebra libraries. The opensource OpenBLAS<br>
should be almost identical to Intels MKL (see <a href="https://urldefense.proofpoint.com/v2/url?u=https-3A__www.mail-2Darchive&d=DwIGaQ&c=yHlS04HhBraes5BQ9ueu5zKhE7rtNXt_d012z2PA6ws&r=U_T4PL6jwANfAy4rnxTj8IUxm818jnvqKFdqWLwmqg0&m=_NY1hgzBYHxYTFCc6OKMrgmlmLPdt3b__vsdXAoQC14&s=7D6kz5IB8jHQBz1mDaZyKAY1Qi7mPGB6quKsNFRB0o4&e=.com/wien@zeus.theochem.tuwien.ac.at/msg15106.html" rel="noreferrer noreferrer" target="_blank">https://urldefense.proofpoint.com/v2/url?u=https-3A__www.mail-2Darchive&d=DwIGaQ&c=yHlS04HhBraes5BQ9ueu5zKhE7rtNXt_d012z2PA6ws&r=U_T4PL6jwANfAy4rnxTj8IUxm818jnvqKFdqWLwmqg0&m=_NY1hgzBYHxYTFCc6OKMrgmlmLPdt3b__vsdXAoQC14&s=7D6kz5IB8jHQBz1mDaZyKAY1Qi7mPGB6quKsNFRB0o4&e=.<br>
com/wien@zeus.theochem.tuwien.ac.at/msg15106.html</a> for comparison of<br>
OenBLAS vs MKL). However in this old benchmark the opensource stack is<br>
still quite slower since the MKL also provides the VML library for<br>
vectorized math functions, which did not had any open source<br>
alternative for a long time. Recently there is the libmvec library<br>
which provides such functions (you need recent glibc), but there is no<br>
official Wien2k support for this. However it is actually quite easy to<br>
get it working (see <a href="https://urldefense.proofpoint.com/v2/url?u=https-3A__www.mail-2Darchive.com_wien-40zeus.theochem.tuw&d=DwIGaQ&c=yHlS04HhBraes5BQ9ueu5zKhE7rtNXt_d012z2PA6ws&r=U_T4PL6jwANfAy4rnxTj8IUxm818jnvqKFdqWLwmqg0&m=_NY1hgzBYHxYTFCc6OKMrgmlmLPdt3b__vsdXAoQC14&s=rdruSPlbzu87Wc6ZNL7iE8pWopEi2yuuvJSnKEzGxTQ&e=" rel="noreferrer noreferrer" target="_blank">https://urldefense.proofpoint.com/v2/url?u=https-3A__www.mail-2Darchive.com_wien-40zeus.theochem.tuw&d=DwIGaQ&c=yHlS04HhBraes5BQ9ueu5zKhE7rtNXt_d012z2PA6ws&r=U_T4PL6jwANfAy4rnxTj8IUxm818jnvqKFdqWLwmqg0&m=_NY1hgzBYHxYTFCc6OKMrgmlmLPdt3b__vsdXAoQC14&s=rdruSPlbzu87Wc6ZNL7iE8pWopEi2yuuvJSnKEzGxTQ&e=</a><br>
<a href="http://ien.ac.at/msg16159.html" rel="noreferrer noreferrer" target="_blank">ien.ac.at/msg16159.html</a> ).<br>
Hence if you use the gfortran + OpenBLAS + libmvec the performance is<br>
virtually identical to ifort + MKL + VML. The setup is somewhat more<br>
difficult though.<br>
<br>
Best regards<br>
Pavel<br>
<br>
> My objective is not to do simulations with mpi in the computer that<br>
> I'm trying to install but to figure out how to install wien2k with<br>
> mpi and then give some guidelines to the IT technician. I spent two<br>
> weeks telling them that the simulations were not running because the<br>
> packages were not compiled and in the end everything was poorly<br>
> installed.<br>
><br>
> Thank you for your help.<br>
><br>
> Best regards,<br>
> Rui Costa.<br>
><br>
<br>
_______________________________________________<br>
Wien mailing list<br>
<a href="mailto:Wien@zeus.theochem.tuwien.ac.at" target="_blank" rel="noreferrer">Wien@zeus.theochem.tuwien.ac.at</a><br>
<a href="https://urldefense.proofpoint.com/v2/url?u=http-3A__zeus.theochem.tuwien.ac.at_mailman_listinfo_wien&d=DwIGaQ&c=yHlS04HhBraes5BQ9ueu5zKhE7rtNXt_d012z2PA6ws&r=U_T4PL6jwANfAy4rnxTj8IUxm818jnvqKFdqWLwmqg0&m=_NY1hgzBYHxYTFCc6OKMrgmlmLPdt3b__vsdXAoQC14&s=q8GD6lLCuIa7NZly_UAheLzby1ihOwinBLQyO21N9hI&e=" rel="noreferrer noreferrer" target="_blank">https://urldefense.proofpoint.com/v2/url?u=http-3A__zeus.theochem.tuwien.ac.at_mailman_listinfo_wien&d=DwIGaQ&c=yHlS04HhBraes5BQ9ueu5zKhE7rtNXt_d012z2PA6ws&r=U_T4PL6jwANfAy4rnxTj8IUxm818jnvqKFdqWLwmqg0&m=_NY1hgzBYHxYTFCc6OKMrgmlmLPdt3b__vsdXAoQC14&s=q8GD6lLCuIa7NZly_UAheLzby1ihOwinBLQyO21N9hI&e=</a><br>
SEARCH the MAILING-LIST at:  <a href="https://urldefense.proofpoint.com/v2/url?u=http-3A__www.mail-2Darchive.com_wien-40zeus.theochem.tuwien.ac.at_index.html&d=DwIGaQ&c=yHlS04HhBraes5BQ9ueu5zKhE7rtNXt_d012z2PA6ws&r=U_T4PL6jwANfAy4rnxTj8IUxm818jnvqKFdqWLwmqg0&m=_NY1hgzBYHxYTFCc6OKMrgmlmLPdt3b__vsdXAoQC14&s=f-OQQ1fsIHEDXpL8Zy4Sy4C_4jyrDkZgdqtQvTOMYk8&e=" rel="noreferrer noreferrer" target="_blank">https://urldefense.proofpoint.com/v2/url?u=http-3A__www.mail-2Darchive.com_wien-40zeus.theochem.tuwien.ac.at_index.html&d=DwIGaQ&c=yHlS04HhBraes5BQ9ueu5zKhE7rtNXt_d012z2PA6ws&r=U_T4PL6jwANfAy4rnxTj8IUxm818jnvqKFdqWLwmqg0&m=_NY1hgzBYHxYTFCc6OKMrgmlmLPdt3b__vsdXAoQC14&s=f-OQQ1fsIHEDXpL8Zy4Sy4C_4jyrDkZgdqtQvTOMYk8&e=</a><br>
</blockquote></div>