Porting software to OpenVMS AXP platform

Adrian Goldman, BTK x8029 GOLDMAN at ALA.BTK.UTU.FI
Thu Mar 11 06:49:40 EST 1993


In <1993Mar10.103238.1 at px1> hudel at px1 writes:

> I was wondering who has already ported their favourite crystalloghraphic
> program/package to the DEC OpenVMS AXP platform.  I am particularly interested
> in things like PROLSQ, PROTEIN, XPLOR etc.
> 
> We just got our first Alpha this week and so far I've only ported MOSFLM and
> CCP4 to run on it...  With less than spectacular CPU time reductions, since
> MOSFLM is mostly i/o-bound.  Reducing 60 images took only 20% less CPU
> time than on a VAXstation 4000/60.  And a ROTAVATA/AGROVATA run was only
> about three times as fast as on a VS4000/60.  So let's hope that we'll get to
> see a factor of 10 for those long s.a. runs with XPLOR!!!

We have had our Alpha AXP 3000 model 500 for a little while (not 
clustered :() and I pulled Xplor 3.0 across onto the machine.  There was 
one line of code that had to be changed (Axel sent me the patch instructions)
and I compiled & linked it with no problems.  The only machine that's faster
through his benchmark suite is the Cray YMP (by a factor of 3)

If anyone cares, I can post the benchmark results.

                    Adrian

Adrian Goldman                         |  Internet:  Goldman at Ala.BTK.UTu.Fi
Centre for Biotechnology,              |  Bitnet:    AGoldman at Finabo
P. O. Box 123,                         |  Phone:     358-21-6338029   
Turku, SF-20521                        |  Fax:       358-21-6338000   
Finland.                               |                              




More information about the Xtal-log mailing list