[mira_talk] Re: Regarding Time-estimate of mira
- From: Bastien Chevreux <bach@xxxxxxxxxxxx>
- To: mira_talk@xxxxxxxxxxxxx
- Date: Tue, 21 Dec 2010 23:53:40 +0100
On Dienstag 21 Dezember 2010 Ganga Jeena wrote:
> I had just 270GB on home and log files generated exceeded this.
> So I had to run it on NFS mount (3TB ), for which I got the warning that it
> wud reduce the speed by factor of 10.
> It has already been 4 days and yet the program is in Pass 1.
>
> I have 454 data with avg size 380, input 8 million reads, how much faster
> wd running on home|headnode have been?
Factor 2 to 3 is realistic with MIRA versions >=3.2.0, i.e. a slow down of 3
to 5 compared to local disk.
> As far as I have been observing it, though given an option for 30
> processors, it had just used 25 for skim, for rest , it has been using max
> 1 processor and 25-65 % RAM, if thats the status, How should running on
> head node make a difference to speed and if it does by what time can i
> expect my assembly to complete.
As written in earlier mail: no time estimates possible.
B.
--
You have received this mail because you are subscribed to the mira_talk mailing
list. For information on how to subscribe or unsubscribe, please visit
http://www.chevreux.org/mira_mailinglists.html
Other related posts:
