1 |
Donnie Berkholz wrote: |
2 |
> Andrew D. Fant wrote: |
3 |
> |
4 |
>> I think the real solution here is to use modules ( |
5 |
>>http://modules.sourceforge.net ) to handle the different versions of |
6 |
>>mpi. I filed a bug about using this for switching alternate virtuals a |
7 |
>>couple years ago, but nobody seemed to be particularly interested. I've |
8 |
>>got a grad student working on something to let users switch between mpi |
9 |
>>for gcc and mpi for the portland group compilers. I am most certainly |
10 |
>>not the right person to do the heavy lifting on this, but I do have an |
11 |
>>ebuild for modules I could contribute to the cause. |
12 |
> |
13 |
> |
14 |
> That is an idea that's been brought up on the bug, but we need something |
15 |
> that integrates into the Gentoo framework for switching things around: |
16 |
> eselect [1]. |
17 |
> |
18 |
> Perhaps one could create an eselect wrapper around modules. It certainly |
19 |
> couldn't hurt to take a look at modules (via your ebuild) and see |
20 |
> whether we could adapt it to our needs. |
21 |
> |
22 |
> Thanks, |
23 |
> Donnie |
24 |
> |
25 |
> 1. http://www.gentoo.org/proj/en/eselect/ |
26 |
> |
27 |
|
28 |
I'll dig around and forward the ebuild to the list later tonight or tomorrow. |
29 |
At the risk of going off on a rant, I think that the /etc/env.d model is a core |
30 |
part of the problem. If two users want to use two different mpi |
31 |
implementations, will eselect allow them to both use them at the same time, or |
32 |
will it have to be one or the other at once? As I see it, something will have |
33 |
to be tweaking $LD_LIBRARY_PATH for each user's environment, or /etc/ld.so.cache |
34 |
will have to be changed for each user's login to reflect their prefered mpi |
35 |
environment. $PATH and $MANPATH can be handled by judicious checks in |
36 |
/etc/csh.cshrc and /etc/profile, but since /etc/env.d insists on manipulating |
37 |
the shared object cache directly, this has the potential to get really ugly |
38 |
without some cooperation from the core system team. I don't think that the |
39 |
cluster herd can fix it on it's own. |
40 |
|
41 |
I would love to be proven wrong and will gladly defer to anyone who does so, I |
42 |
just know that I've been trying to outthink this problem for a year or so on our |
43 |
cluster and haven't come up with a sane solution for a true multi-user environment. |
44 |
|
45 |
Andy |
46 |
|
47 |
-- |
48 |
Andrew Fant | The lion and the calf shall lie | Disclaimer: |
49 |
andrew.fant@×××××.edu | down together, but the calf won't | Do you REALLY |
50 |
TCCS/USG | get much sleep. | think I can |
51 |
Tufts University | W. Allen | speak for Tufts? |
52 |
-- |
53 |
gentoo-cluster@g.o mailing list |