wrong clm output on mistral – in #9: CCLM

  @ulrichschättler in #cdb4f53

<p> Hello David, <br/> our <span class="caps"> COSMO </span> colleague Davide from Bologna lately had a similar problem. And it is not a problem of the name “David”, but he traced it back to a possible compiler bug in the Intel compiler: </p> <p> Quoting from his Mail: </p> <p> “we noticed that recent version of Intel compiler <br/> introduces a bug in src_conv_tiedtke of <span class="caps"> COSMO </span> , producing strange <br/> convective precipitation bands at <span class="caps"> MPI </span> subdomain boundaries (see <br/> attachment). This has been noticd in operational version 5.0, but I <br/> assume the convection scheme did not change recently. </p> <p> This is the compiler version: <br/> [dcesari0@node166 ~]$ ifort —version <br/> ifort ( <span class="caps"> IFORT </span> ) 15.0.2 20150121 <br/> Copyright © 1985-2015 Intel Corporation. All rights reserved. </p> <p> Qnd these are the compiler flags used: <br/> -fno-math-errno -opt-subscript-in-range -align all -axavx -xavx <br/> -heap-arrays -O3 </p> <p> By disabling all the optimizations only on the Tiedtke module (-O0) the <br/> anomaly disappears. The colleagues at Cineca will try to better isolate <br/> the bug (which flag and which code region); it seems that there is no <br/> appreciable feedback on any other model field in atmosphere, so I think <br/> the problem is confined to the prr_conv array. Has anybody noticed this <br/> problem?” </p> <p> And he showed a figure from convective precipitation, which I attach. </p> <p> Up to now nobody else complained about that problem, but maybe you are the next one. Are you using the Intel compiler? </p> <p> Ciao <br/> Uli </p>