 08/02/2004 07:07 AM
|
Hornet3 Senior Member

Posts: 237
Joined: 01/09/2004
|
I've been doing a few benchmarks on my system and it appears to be a rather poor performer compared to the 64 3000+ in many reviews.
For starters, in SiSoft Sandra 2004, my memory bandwidth is 2760MB/s, as opposed to the 3000MB/s in every other review.
In Aquamark 3, my CPU score is about 8266, as opposed to 8700+ in many other systems.
I'm not sure why I'm not getting full performance. The only thing I've noticed, but I don't think this is the reason, is that my VCore is 1.472V (and it should be 1.50V). I'm not sure if voltage has anything to do with this issue, but I don't know what it could be.
-------------------------
Current System:
- AMD 64 3000+ @ 2.01GHz
- Gigabyte K8T800
- 1024MB Corsair PC3200XL
- Gecube X800XT 500/1000 256MB
- Creative Audigy 2 ZS
- Logitech Z-680 5.1 500W
- Mitsubishi 19" 1995GB
- Antec TrueBlue 480W PSU
|
|
|
|
 08/02/2004 09:28 AM
|
Hornet3 Senior Member

Posts: 237
Joined: 01/09/2004
|
I hit Ctrl-Alt-Delete and I just want to check if these numbers sound right.
Physical Memory Total - 523760 (sounds right) Available - 322360 System cache - 261900
Does Windows XP usually use up 200MB or RAM? Also, while I'm on about RAM, would another 512MB influence the CPU benchmarks much. I know they will improve performance, but will they greatly improve performance?
-------------------------
Current System:
- AMD 64 3000+ @ 2.01GHz
- Gigabyte K8T800
- 1024MB Corsair PC3200XL
- Gecube X800XT 500/1000 256MB
- Creative Audigy 2 ZS
- Logitech Z-680 5.1 500W
- Mitsubishi 19" 1995GB
- Antec TrueBlue 480W PSU
|
|
|
|
 08/02/2004 09:31 AM
|
randy2 Senior Member

Posts: 303
Joined: 12/01/2003
|
The differences in your scores as compared to the reference systems are so small, if they matched, you wouldn't know the difference. Sisoft Sandra is usually worthless for benchmark comparisons from one machine to another. It is a good tool to measure changes within your own system - software tweaks etc, but worthless in comparing two separate machines. In other words, you can build 2 identical machines, and have different scores on each with Sandra.
|
|
|
|
 08/02/2004 09:35 AM
|
Hornet3 Senior Member

Posts: 237
Joined: 01/09/2004
|
Thanks for the reply, I'll take that fact into account. I can understand the 200MB/s difference in SiSoft Sandra, but the differences in Aquamark are huge. In a CPU comparison at Anandtech, their AMD 64 3000+ scored 8900.
Anyway, I'll look into it and see if there are any tweaks I can make to BIOS. I want that extra performance - I might have to settle for a 3500+.
-------------------------
Current System:
- AMD 64 3000+ @ 2.01GHz
- Gigabyte K8T800
- 1024MB Corsair PC3200XL
- Gecube X800XT 500/1000 256MB
- Creative Audigy 2 ZS
- Logitech Z-680 5.1 500W
- Mitsubishi 19" 1995GB
- Antec TrueBlue 480W PSU
|
|
|
|
 08/02/2004 11:42 AM
|
randy2 Senior Member

Posts: 303
Joined: 12/01/2003
|
You are only talking about 400 points there, nothing noticable. An antivirus program running or a slighlty fragmented disk can alter the score by that much.
|
|
|
|
 08/02/2004 05:07 PM
|
GripS Senior Member

Posts: 1443
Joined: 03/25/2004
|
Your memory bandwidth seems a little off. Especially since you have higher rated memory. You don't have cool and quiet enabled do you? If so, shut it off(in power management options and/or in bios). Run the bench again and i bet you'll see a higher score. For your memory you'll only reach it's full potential when you overclock the Hyper transport(or fsb). At default it will only run at pc3200 speed(400 mhz). I score 3,090/3,089 with my setup. When I overclock to 2.2 ghz my memory runs at 440mhz giving me a score of 3,374/3,370.
-------------------------
AMD Opteron 170 \ DFI UT nf4 Ultra-D \ evga 7600gt \ 2gb (2x1gb) G.Skill pc4000 HZ \ NEC 3550a DVDRW \ WD Raptor 150gb \ Seagate barracuda 80gb \ Mitsumi floppy \ OCZ powerstream 520watt
|
|
|