Computer for sequence analysis

[William T. Rankin]

In article <362a67ce.3846840 at netnews.upenn.edu>,
Al Wang <alwang at doubt.com> wrote:

>On 15 Oct 1998 10:56:07 -0700, Bent.Nagstrup at kba.sks.aau.dk (Bent
>Nagstrup) wrote:
>
>>K6's are generally not good due to their rather poor floating point 
>>performance.

Again, this depends.  I have seen non-toy benchmarks where the K6 ran
pretty much neck-n-neckwith some of the pentium class machines (for a
given clockspeed).

In the end, the only good benchmark for processor/system performance is to
run your own application.

>>>  I understand most software is for Unix, isn't it?
>>
>>Yes, but which unix flavour? That's the real problem, if it runs under Linux 
[...]
>As another option to keep in mind, it is possible to run Solaris on an
>Intel box...  Not terribly fast, but if you need economic hardware and
>the compatibility of Solaris, it's an option.

Um, that comparability only extends as far as the operating systems
interface and the API.  Executable programs that run on a SparcStation
will *not* run on an Intel box running Solaris x86.

Yes, often this is just a matter of a recompile.  If you are running
commercial software, however, you need to be aware that a lot of
commercial software that runs under "Solaris" is not available for the
Intel version of Solaris, only the Sparc.

-bill

-- 
bill rankin ...................................... philosopher/coffee-drinker
wrankin at ee.duke.edu ........................................ doctoral wannabe
duke university dept. of electrical engr ......... scientific computing group