[BiO BB] remove CTL-M and Buying a bioinformatics workstation
dmb at mrc-dunn.cam.ac.uk
Thu Sep 4 08:13:32 EDT 2003
On Thu, 4 Sep 2003, Yannick Wurm wrote:
> And Microsoft Word runs on it too :)
> Are the results shown surprising? The code for A/G Blast and HMMer seem
> to be optimized for the G5, whereas the standard vanilla versions where
> used on the linux boxes. Could optimization reduce the bias against the
I hope so!
Did they run blast with the --number_of_cpus option set to 2 or 4?
The kernel level hyperthreading on the Xeon simulates 2n cpu's, but
I am not sure if this 'really' works (i.e. 4*2Gh from a 2*2Gh board).
Also I found the --number_of_cpus to be sub optimal in terms of cpu
usage. I prefer to simply start 4 jobs at the same time, so maby
this would improve the benchmark considerably (I got from ~50%
average usage on each cpu to ~100%).
Did they write jobs to /dev/null ?
if not the IO could be a (hidden?) factor.
> Yannick Wurm
> \\ http://yannick.poulet.org icq: 22044361
> \\ idh at poulet.org tel: ++126.96.36.199.71.92
> BiO_Bulletin_Board maillist - BiO_Bulletin_Board at bioinformatics.org
More information about the BBB