31.03.10, 18:19, "Peter Murray-Rust" <[email protected]>: 
> 
> 
> On Wed, Mar 31, 2010 at 5:43 PM, Geoffrey Hutchison <[email protected]> wrote:
> > > More accurately, however, I have watched while a compchem syadmin expert 
> > > has spent an hour compiling MPQC. This requires the BLAS/LAPACK library 
> > > which took him a considerable time to tame - and he's been through this 
> > > before. If I'd had to do this myself (and I do know how to make make and 
> > > configure in general terms) it would have taken me a week or probably 
> > > longer. It was NOT trivial.
> >  
> >  No offense, but if a "compchem sysadmin expert" has problems with 
> > BLAS/LAPACK and has done it before, they're not really an expert.
> >  BLAS/LAPACK (or ATLAS) are easy to install on every Linux workstation or 
> > server I have ever used. Indeed, on most compchem Linux machines, it's 
> > already installed.
> >  
> >  
> Well I won't go into it in detail, but the point is that there are people who 
> find this easy and people who don't. Among other things we have to compile it 
> on machine A in a static manner and transfer it to a grid where most of the 
> libraries have been stripped off.
>  
> But in any case *I* didn't find it easy. Firstly you have to *compile* the 
> program. I've done a lot of compiling (but not recently). I went to
> http://www.mpqc.org/mpqc-snapshot-html/compile.html
>  and tried to follow the instructions. I failed. I am not stupid or 
> incompetent, simply ignorant. There are a lot of people like me who do not 
> normally compile Linux programs and who get turned off very rapidly indeed by 
> the compilation problem.

It's common problem over QC programs  - many of them use custom build systems, 
all of them require accurate choice of compiler/libraries, etc. Pre-buit 
packages will always lack optimal performance, so, I think everyone starting to 
work in QC should pass through this way

>  
> If I had been wanting to run it on a Windows machine what would I have done? 
> There is no binary on the MPQC site and no indication of how to compile it in 
> Windows. It's not trivial.

Windows users are always out of luck in QC field :) Most of software was 
written in early times when there wasn't any Windows or Linux :)

> 
> So I reiterate my problem. Many people want to use compchem programs and 
> don't know where to start or how to proceed. There is currently a priesthood 
> of compchemists which makes it very difficult when you don't have one at hand.
>  
> I have no idea what apt-get is and it's not on the MPQC compile page.
> 
> The point I am trying to make is that I am suggesting the BO try to make it 
> easy for chemists, not Linux / Ubuntu / Debian experts ,  to run a small 
> number of Open programs. I don't want to compile them, only run them. I know 
> what the programs do. I do not know and I don't want to know what the 
> compilers do. Is there a way the BO can help? 
>   
> 
> 
> -- 
> Peter Murray-Rust
> Reader in Molecular Informatics
>  Unilever Centre, Dep. Of Chemistry
> University of Cambridge
> CB2 1EW, UK
> +44-1223-763069


-- 
Regards,
Konstantin

Яндекс.Почта. Письма есть. Спама - нет. http://mail.yandex.ru/nospam/sign

------------------------------------------------------------------------------
Download Intel&#174; Parallel Studio Eval
Try the new software tools for yourself. Speed compiling, find bugs
proactively, and fine-tune applications for parallel performance.
See why Intel Parallel Studio got high marks during beta.
http://p.sf.net/sfu/intel-sw-dev
_______________________________________________
Blueobelisk-discuss mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/blueobelisk-discuss

Reply via email to