Re: system usage stats (Was: Re: Why Not MySQL? )
Ross J. Reedstrom <reedstrm@wallace.ece.rice.edu>
From: "Ross J. Reedstrom" <reedstrm@wallace.ece.rice.edu>
To: Tom Lane <tgl@sss.pgh.pa.us>
Cc: pgsql-hackers@postgresql.org
Date: 2000-05-04T16:03:31Z
Lists: pgsql-hackers
On Thu, May 04, 2000 at 11:06:20AM -0400, Tom Lane wrote: > Don Baccus <dhogaza@pacifier.com> writes: > > So, what's the bottom line? The numbers don't tell us much, > > though I still think Tom's right that the PG7.0 one is really > > slower. You just can't say if how MUCH slower. > > Actually I was a lot more concerned about disk performance than CPU > speed. I notice no one's said anything about the relative speed > of Mitch's two different disk setups ... > Mitch wrote: > Ok, the production server is a Celeron 433, 512 MEgs of PC100 ECC RAM, 2 18 > Gig Ultra 160 SCSI drives (only running at 80 megs since we dono't have a > 64-bit PCI slot). > > There is a big upgrade planned for this box when the motherboard we're > waiting for comes out.. > > The development server is a PII450, 128 Megs of RAM and Ultra 2 SCSI > drives.. However it's running several other things other than the database > and webserver so it has a small load. > So, 6.5.3 is running on Ultra 160 drives, with the controller throttled to 80 MB/s, and 7.0 is running on Ultra 2 drives, which also has a controller maximum of 80 MB/s. However, the sustained transfer speed of the drives themselves are what should be limiting: if they're all relatively modern drives, 20MB/s is typical, so neither config will max out the controller. (2 drives each, right?) Ross -- Ross J. Reedstrom, Ph.D., <reedstrm@rice.edu> NSBRI Research Scientist/Programmer Computer and Information Technology Institute Rice University, 6100 S. Main St., Houston, TX 77005