New Borg - Poor Performance

verbal

verbal

verbal

Associate
Joined
17 May 2006
Posts
1,189
Location
Manchester, UK
I've borged a Xeon 3050 system running at stock (2.13ghz). I'm running two standard clients on it and I am only getting exactly 100ppd from each core. Running just one client doesn't affect this.

This ppd is on project 892 which I get 130ppd from on a E6320 at stock (1.86ghz)

Can anyone shed any light on why this might be? The Xeon system is doing next to nothing else and is running Server2003, the Core2Duo system is running a few light apps all the time under XP.
 
The Xeon isn't running standard loops, is it? Go take a look at the beginning of the log and look for something about SSE Optimizations.
 
Hmm, there is no line mentioning "SSE" specifically just... "Assembly optimizations on if available"

This is the same on my E6320 box
 
You could try the SMP client on it for a performance boost.

I remember back in the day when Ambers were desirable and it was Tinkers that everybody hated.
 
Never had a Tinker, at least I don't think so. Maybe when I folded for CPC...

When I joined OcUK it was just before the fabled 500 and 600 pointers were in abundance. 600PPD was considered awesome :cool:

Funny really, I drop 600PPD if I go into Windows for an hour :p
 
Hmm doesn't explain why the E6320 is out-performing the Xeon. My P4 3.0ghz rigs get more ppd for this WU than the Xeon and the E6320 and so do my Celeron 2.4s!
 
Dual-Core Xeon 3050 - SL9TY (B2) - Conroe - 2133 MHz - 2048 KiB - 1066 MT/s - 8x - 65 W
Dual-Core Xeon 3050 - SL9VS - Allendale - 2133 MHz - 2048 KiB - 1066 MT/s - 8x - 65 W

Core 2 Duo E6320 - SLA4U (B2) - Conroe - 1866 MHz - 4096 KiB - 1066 MT/s - 7x - 65 W

[Guesswork] The extra L2 cache on the E6320? [/Guesswork]
 
That's a decent guess beserker...

Unfortunately now one core is getting around 44ppd on project 2527 and the other core is getting 52ppd on 2096. Surely too low for a core2duo based Xeon?
 
You must log in or register to reply here.
Back
Top Bottom