Dear Mathew,
I have forwarded this answer to the amber mailing list which is where such
queries should be posted so that other researchers can also benefit from
replies. See
http://amber.scripps.edu for information on subscribing to the
list.
> -----Original Message-----
> From: Mathew Varghese [mailto:mathewkv.gmail.com]
> Sent: 06 April 2005 21:06
> To: ross.rosswalker.co.uk
> Subject: system comparison for amber
>
> Dear sir,
> I wish to know about the comparison of two systems.A two Xeon
> processor system fromIBM(IBM Xseries) and a single processor SGI ALTIX
> with 2gb memory. Which one will be better for molecular dynamics
> calculations with AMBER(sander)
There are a series of benchmarks on these types of systems posted on
http://amber.scripps.edu/amber8.bench1.html you can check these for timings.
In "my opinion" you will be better off with the Dual Xeon machine. The
reason I say this is that the Intel P4 xeon chips are generally as fast, if
not faster in some cases, than the Itanium chips in the Altix. Now, the
Altix is a very well built machine and scales to large numbers of cpus very
well. For large systems (ca 128 cpu) the Altix is awesome. However, you pay
for this privilege. The Altix was essentially designed from the ground up to
work exceptionally well as a big SMP machine. Hence for machines / clusters
with less than 8 cpus the Altix does not provide good value for money. It
does support true 64 bit addressing, which the new xeon chips only "half
heartedly" support but 64 bit integers don't give you any performance
increase. The only advantage is that you can allocate more than 2GB of
memory, which you are unlikely to need to do with AMBER, and you can write
files bigger than 2GB. The large file support is the most useful with AMBER
as MDCRD files can be very large. However, large file support emulation is
pretty much ubiquitous on 32 bit Linux machines these days (the Intel
compilers definitely support it) and so this advantage is also negated.
Hence in your case you would be far better getting the Dual Xeon machine
which I expect is cheaper than the single processor Altix. You might also
want to consider other dual xeon machines from vendors other than IBM as I
suspect performance across the range of machines available will be pretty
similar, and not necessarily correlated with the price.
All the best
Ross
/\
\/
|\oss Walker
| Department of Molecular Biology TPC15 |
| The Scripps Research Institute |
| Tel:- +1 858 784 8889 | EMail:- ross.rosswalker.co.uk |
|
http://www.rosswalker.co.uk/ | PGP Key available on request |
Note: Electronic Mail is not secure, has no guarantee of delivery, may not
be read every day, and should not be used for urgent or sensitive issues.
-----------------------------------------------------------------------
The AMBER Mail Reflector
To post, send mail to amber.scripps.edu
To unsubscribe, send "unsubscribe amber" to majordomo.scripps.edu
Received on Thu Apr 07 2005 - 17:53:00 PDT
