> On Feb 26, 2016, at 11:12 AM, Brent Krueger <kruegerb.hope.edu> wrote:
>
> As someone who is responsible for a multi-user facility that currently
> operates CentOS 5.6 and uses version 12.0 of the Intel compilers, I'd like
> to second Ross's notes about working to support older (AND STABLE)
> systems. If I'm the only one running something this old, then I will admit
> that the problem is mine and we will work to upgrade, but I suspect that is
> not the case.
There's a difference between what we support and what we can set up to test through CI. As long as a CentOS 5 box has the necessary software, it will work. You'll just need to not use GCC 4.1.2 anymore :).
CI setups have to be balanced against hardware availability constraints and what allocation of resources will best help limit bugs and detect problems early. The right balance is likely to be something we have to try multiple things to find the best solution.
>
>
> Also, it is probably relevant for me to note that somewhere between Jan 14
> and today things have stopped building nicely for me. It appears that my
> libfftw is out of date, which seems strange to me since this is something
> that I believe configure checks. I wonder if the code requirements
> increased without someone updating configure accordingly? Anyway, the most
> obvious error I am seeing is that I do not have libfftw3.so.3 This
> happens at the point when the build is working on cpptraj.
>
> I'm attaching my build output for anyone who wishes to wade through it.
>
>
> Cheers,
> Brent
>
>
>
>> On Fri, Feb 26, 2016 at 10:44 AM, Daniel Roe <daniel.r.roe.gmail.com> wrote:
>>
>> On Fri, Feb 26, 2016 at 8:37 AM, Ross Walker <ross.rosswalker.co.uk>
>> wrote:
>>> Recent means CentOS 7? Or also CentOS 6? - We should at least be testing
>> on all OS versions that still under official support. I think this also
>> means CentOS 5 although I don't see many systems running this anymore. But
>> 6 is very ubiquitous. A lot of people still install 6 on new machines given
>> how much they screwed things up in 7.
>>>
>>> We should probably add Suse to that list as well.
>>>
>>> Do github offer CI hardware that includes various GPUs? (and KNC Xeon
>> Phi would be useful too).
>>
>> I don't think they do (although someone correct me if I'm wrong here).
>> I think we would have to set up our own test server for things like
>> that, but GitHub (and probably similar things like GitLab) have hooks
>> you can set up to do this automatically (see e.g.
>> https://developer.github.com/guides/building-a-ci-server/).
>>
>> Having our own servers might actually be the best way to go in the
>> long run since we should certainly cover as many of the common
>> environments as we possibly can, which may be a bit too much for
>> Travis, especially considering the size of the test cases for e.g.
>> sander. Speaking of that - after this release it may be worth seeing
>> if we can streamline the test cases somewhat...
>>
>> -Dan
>>
>>>
>>> All the best
>>> Ross
>>> _______________________________________________
>>> AMBER-Developers mailing list
>>> AMBER-Developers.ambermd.org
>>> http://lists.ambermd.org/mailman/listinfo/amber-developers
>>
>>
>>
>> --
>> -------------------------
>> Daniel R. Roe, PhD
>> Department of Medicinal Chemistry
>> University of Utah
>> 30 South 2000 East, Room 307
>> Salt Lake City, UT 84112-5820
>> http://home.chpc.utah.edu/~cheatham/
>> (801) 587-9652
>> (801) 585-6208 (Fax)
>>
>> _______________________________________________
>> AMBER-Developers mailing list
>> AMBER-Developers.ambermd.org
>> http://lists.ambermd.org/mailman/listinfo/amber-developers
>
>
>
> --
> _______________________________________________
> Brent P. Krueger.....................phone: 616 395 7629
> Professor.................................fax: 616 395 7118
> Hope College..........................Schaap Hall 2120
> Department of Chemistry
> Holland, MI 49423
> <amber16_install_2016-02-26-0811.gz>
> _______________________________________________
> AMBER-Developers mailing list
> AMBER-Developers.ambermd.org
> http://lists.ambermd.org/mailman/listinfo/amber-developers
_______________________________________________
AMBER-Developers mailing list
AMBER-Developers.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber-developers
Received on Fri Feb 26 2016 - 08:30:10 PST
