Hi folks,

I have a very strange performance problem. We have two machines, nearly identical, except that SQL server is delivering data from one of them at about 6 times the speed of the other. We're kind of stumped.

Naturally we thought of some type of obvious config setting or hardware difference, but cannot find any.

For instance, the faster machine is a Dell PowerEdge 2300, dual 450Mhz processors, 512MB RAM, Perc 2 controller w/144MB RAM, and some HDD's that are spinning at 10,000RPM and are rated at 29.5Mb/s to buffer, 80Mb/s to host.

The odd thing is that the slower machine is a Dell PowerEdge 4300, dual 500Mhz processors, 1Gig RAM, identical perc 2 controller, and some slower 7200 RPM drives also rated 30Mb/s to buffer, 80Mb/s to host.

Tried examining the SQL server setups for something obviously awry, but found nothing. They are both configured to use both CPU's, with the same settings for parallelism.

We do not think it is usage, because we noticed the new machine was slower when we still had 2 people on it and the other twenty were on the older box.

When we first suspected the newer machine was slower, I cooked up a small test. I generated a few tables on each machine, with varying numbers of records, all packed with 500-byte records filled with endless repititons of the foxpro SYS(2015) call, which gives you in reality about 6 repeating bytes of every 8, the last two being more random. So I've got tables with 50 fields of 10 bytes each, populated with randomized but compressible data.

Now matter how I run the test, simple SELECT * FROM Table from Foxpro pulls 7500 recs/second from the older machine, and about 1800 from the newer.

Anybody seen anything like this?

TIA,