[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

Re: [ccp4bb]: question about Linux-based crystallographic computing(fwd)



***  For details on how to be removed from this list visit the  ***
***          CCP4 home page http://www.ccp4.ac.uk         ***

To stir up a bit:

If you are talking abt individual workstations,by all means, get the latest
AMD or P4 and run Linux.

I find a good choice for labs with more than 3-4 active crystallographers that
ain't computer gurus, to have a strong multi-proc machine:

1. Put there a RAID array and sleep well at nights
2. No-overhead in execution or thinking time to submit short jobs, or multiple
jobs
3. Common software distributions
4. Lower maintanace overhead
5. Easier backups
6. Eventually something might get even parallelised in xtallography (well,
thats a joke)

I also have to add that since we got our ES40 last June (10 months now) I did
not have to reboot it a single time  (also thanx to the UPS, worked nicely
during 4-5 power-cuts). Really, not oncerebooting in a year! I can't say the
same for the average PC, even when running Linux, not to mention Vindoz.

So, if you opt for that choice (as we do) then you really get much more from
the Alpha architectures, in any respect I can think of.
rem also that PC with multi-procs tend to cost as much as Alphas.

Last but not least, no matter what the benchmark sites say, in my experience
Refmac is ~3 times faster in my ES40/667MhZ than in a PIII-850Mhz. Simply, I
guess, the Alpha compilers for F77 are way better than g77. So unless you also
get the pg or intel compilers and spend some (...) time to compile everything
with them, at the end of the day the Alphas are faster for the
usual-day-in-crystallography.

But just to contradict my previous statements, back to the start, for a lab
based mostly on the power of machines on people's desks Linux AMD seems best,
and as David says OSX is so tempting (waiting for a grant to get approved to
buy one soon to play with)

        Tassos

--
NKI, Department of Molecular Carcinogenesis
Plesmanlaan 121 1066 CX Amsterdam NL
Phone +31-20-512-1951 Fax +31-20-512-1954
SMS: +31-6-28597791 WWW: http://den.nki.nl