I see your reindexing and raise you:
We have a reindexing process that last August with lots fewer records that we have now that took
some 14 hours. We ran it this weekend on the N-Class in 3:12!
--
Leonard S. Berkowitz
Perot Health Care Systems
(Harvard Pilgrim Health Care account)
voice: 617-509-1212
fax: 617-509-1955
pager: 781-226-2431
Craig Lalley
<[log in to unmask] To: [log in to unmask]
com> cc:
Sent by: HP-3000 Subject: Re: [HP3000-L] Tune command flakes out
Systems
Discussion
<[log in to unmask]
TC.EDU>
07/26/2003 01:37
PM
Please respond to
Craig Lalley
BTW, Do I get an award,... I could go looking for more, I bet Glance is full of
this stuff. :-)
But seriously it seems to me that a lot of the performance tools invented did
not account for high speed raid arrays, 16GB of memory etc.
As for the N-Class 750-400 with 16GB of memory, it rocks. We had a re-indexing
job that was seriously CPU bound on the 997, it took about 6 hours to run. On
the N-Class, it took 1:23 minutes.
I had a repacking job, very I/O bound that I could only estimate between 35 and
70 hours. We never really got a chance to run it to completion on the 997. It
did run for 8 hours straight once and then we had to back it out. It took
3hours and 55 minutes.
Those were the only two benchmarks I had time to run. I plan on running a
whole lot more!
Cheers,
-Craig
--- Bill Cadier <[log in to unmask]> wrote:
> Craig writes:
>
> > Has anyone seen this before? It's an N-Class running MPE/ix 7.5 all
> patches.
> >
> > :showq;status
> >
> > ------QUANTUM-------
> > QUEUE BASE LIMIT MIN MAX ACTUAL BOOST TIMESLICE
> > ----- ---- ----- --- --- ------ ----- ---------
> > CQ 152 200 1 2000 27 DECAY 200
> > DQ 202 238 2000 2000 2000 DECAY 200
> > EQ 240 253 2000 2000 2000 DECAY 200
> >
> > :tune ;dq=,,4000,4000,,400
> > :showq;status
> >
> > ------QUANTUM-------
> > QUEUE BASE LIMIT MIN MAX ACTUAL BOOST TIMESLICE
> > ----- ---- ----- --- --- ------ ----- ---------
> > CQ 152 200 1 2000 19 DECAY 200
> > DQ 202 238 -1726 -1726 -1726 DECAY 400
> > EQ 240 253 2000 2000 2000 DECAY 200
> >
> >
> > What gives?
> >
> > -Craig
>
> I had a few moments this afternoon to look into this. Since it hasn't been
> reported
> so far, I filed SR 8606-321773 for you. This doesn't appear to be a problem
> other
> than reporting an odd value so I expect that this would be handled
> opportunistically
> the next time we have a need to change that area of the code.
>
> The negative values are the result of storing the values internally in ticks
> as opposed
> to milliseconds as they're displayed. On faster systems the internal value
> can be large
> enough to make the value appear to be negative if treated as a signed value.
> I can cause
> this on an N-class, not on a 997.
>
> hth,
>
> Bill
> hp/vCSY
__________________________________
Do you Yahoo!?
Yahoo! SiteBuilder - Free, easy-to-use web site design software
http://sitebuilder.yahoo.com
* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *
* To join/leave the list, search archives, change list settings, *
* etc., please visit http://raven.utc.edu/archives/hp3000-l.html *
|
