Hi Alan Thanks. I'll look at this later (away from the office currently). Does your tar cover score-p? If not, could you comment on whether score-p should be a dependency?
Thanks again Joachim Sent from my nanoPad On 19 Dec 2016, at 13:19, Alan O'Cais <[email protected]<mailto:[email protected]>> wrote: Dear Joachim, We do have these available within JSC, but we use minimal toolchains and customisations at JSC, porting these back to EasyBuild is an awful lot of effort (that we don't have time for). This is a topic I'm addressing for the upcoming User Meeting. Attached is a set of easyconfigs that can be untarred and used as follows (with EB 3.x): eb --include-easyblocks=Scalasca/jhbuild.py --robot=./Scalasca -D Scalasca/Scalasca-2.3.1-* Kind regards, Alan On 19 December 2016 at 12:42, Joachim Hein <[email protected]<mailto:[email protected]>> wrote: Markus, Thanks for the detailed answer. > On 16 Dec 2016, at 15:24, Markus Geimer > <[email protected]<mailto:[email protected]>> wrote: > > Joachim, > >> Thanks for the update. I assume moving foss/2016a to foss/2016b >> would not be an issue. > > As I said: Change the toolchain version number and you're done. > I have now a configuration set for Scalasca 2.3.1 based on foss/2016b - currently in my own GitHub, see https://github.com/jhein32/easybuild-easyconfigs/commit/1ed43293efeb84f856bac8d45f26e02da462bad9 if interested. I make a PR once I have a good answer to the below: When fixing this up, I was surprised that I didn’t need to fix-up a score-p. And now, when I am testing it on one of my OpenMP teaching examples (using skin to instrument) it complains no score-p (I wasn’t surprised to be honnest). I think Score-P should be a dependency. Could you please comment? If a score-p is fixed up, which version should one go? Score-p support in the current EB devel branch is also not that fresh (v 2.0.1). Anything in the pipeline here from your end? >> Any comments on the intel toolchains (e.g. intel/2016b or 2017/01)? >> Are there any known issues around intel compilers, intel MPI why >> there have been no easy configs for this in the past? > > Lack of resources. Basically, I'm trying to maintain the easyconfigs > for the tools of the Score-P universe (OPARI2, OTF2, Cube, Score-P, > Scalasca, Vampir). But this is only a side project, as my main job > is to develop (some of) those tools rather than system administration. > And we (my boss and I) decided to only maintain easyconfigs for the > foss toolchain. > I understand the resource issue to some extent. On the other hand, more and more centres are using or looking into easybuild. Many of the users I’s like to promote Scalasca to are using the intel toolchains (their codes do not always build with gcc, but that is another issue). I think support for intel and foss would be effort well spend to promote your tools. I agree that with the zillion of pre-requisites you require, fixing up e.g. a scalasca takes more effort than I thought. Though that effort might be well spend. Just for a comment. Please get back to me on the score-p. Best wishes Joachim > We do have those tools installed for Intel on our systems, but our > setup differs enough from upstream that we can't simply create a PR > from the easyconfigs we're using without some care. > >> It is my understand that it is supported in principle. > > Yes. It should be as easy as changing the toolchain in the easyconfig. > > Markus > > >>> On 16 Dec 2016, at 15:01, Markus Geimer >>> <[email protected]<mailto:[email protected]>> wrote: >>> >>> Joachim, >>> >>>> May I ask on the situation around scalasca? Since FZJ is very active >>>> here, I prefer the experts doing it than me hacking configs for a >>>> version push. >>>> >>>> Currently Scalasca-2.3-foss-2016a.eb is the latest there is. There is >>>> now a scalasca 2.3.1 and I would like to see support for a newer foss >>>> and the intel (or iomkl) toolchains. Any plans a FZJ to do this? >>> >>> Scalasca 2.3.1 was just a really minor update to address some >>> very specific build problems. Thus we didn't feel that it was >>> worthwhile adding an easyconfig for it. If you take the 2.3 >>> easyconfig, update the toolchain, the Scalasca version, and >>> the checksum, you should be fine as the dependencies haven't >>> changed. >>> >>> If you want to go ahead and prepare a PR, I'm happy to take a >>> look. >>> >>> Hth, >>> Markus >>> >>> -- >>> Dr. Markus Geimer >>> Juelich Supercomputing Centre >>> Institute for Advanced Simulation >>> Forschungszentrum Juelich GmbH >>> 52425 Juelich, Germany >>> >>> Phone: +49-2461-61-1773 >>> Fax: +49-2461-61-6656 >>> E-mail: [email protected]<mailto:[email protected]> >>> WWW: http://www.fz-juelich.de/jsc/ >>> >>> >>> >>> ------------------------------------------------------------------------------------------------ >>> ------------------------------------------------------------------------------------------------ >>> Forschungszentrum Juelich GmbH >>> 52425 Juelich >>> Sitz der Gesellschaft: Juelich >>> Eingetragen im Handelsregister des Amtsgerichts Dueren Nr. HR B 3498 >>> Vorsitzender des Aufsichtsrats: MinDir Dr. Karl Eugen Huthmacher >>> Geschaeftsfuehrung: Prof. Dr.-Ing. Wolfgang Marquardt (Vorsitzender), >>> Karsten Beneke (stellv. Vorsitzender), Prof. Dr.-Ing. Harald Bolt, >>> Prof. Dr. Sebastian M. Schmidt >>> ------------------------------------------------------------------------------------------------ >>> ------------------------------------------------------------------------------------------------ >>> >> > > > -- > Dr. Markus Geimer > Juelich Supercomputing Centre > Institute for Advanced Simulation > Forschungszentrum Juelich GmbH > 52425 Juelich, Germany > > Phone: +49-2461-61-1773<tel:%2B49-2461-61-1773> > Fax: +49-2461-61-6656<tel:%2B49-2461-61-6656> > E-mail: [email protected]<mailto:[email protected]> > WWW: http://www.fz-juelich.de/jsc/ > -- Dr. Alan O'Cais E-CAM Software Manager Juelich Supercomputing Centre Forschungszentrum Juelich GmbH 52425 Juelich, Germany Phone: +49 2461 61 5213 Fax: +49 2461 61 6656 E-mail: [email protected]<mailto:[email protected]> WWW: http://www.fz-juelich.de/ias/jsc/EN <Scalasca.tgz>
