Neoseeker : Articles : Motherboards : Socket 939 : Sapphire Pure CrossFire 3200
Hardware Newsletter:
Email:

News Headlines
New Articles
Compare Prices

Motherboards
Abit
ASUS
Gigabyte
MSI
DFI
Intel
Tyan
More...

Processors
AMD
Intel
More...

Memory
DDR
DDR2
SDRAM
More...

Video Cards
ATI
eVGA
XFX
Sapphire
More...

search for lowest prices

send article   hardware newsletter   article comments (1)
Sapphire Pure CrossFire 3200 - PAGE 13
J. Micah Grunert - Monday, October 23rd, 2006


Now then, let's see what kind of bandwidth we can squeeze out of this board. And as with any board, flashing to the latest BIOS update helps. I had tried some overclocking with the stock BIOS that our Sapphire board came flashed with, and the best I couldn't really push things very far. Stability issues were abound. The board had this nasty habit of absolutely hating restarts. It would more often than not, loose the FSB setting, reverting back to it's 200MHz stock bus rate. It also hung up on restarts. Nothing more than a black screen and not a single post code squeaked.

So, I flashed it. And once flashed, the overclocking went much, much smoother. But there was a hitch involved. For some reason, the board didn't seem to like those 4-4-4-12 memory timings we had set in the beginning. So, to get it running we reset the memory to 5-5-5-18, and that seemed to work wonders. In the end, our final score came out to be a 524.8 x 2MHz memory speed, for a total of 1049.6MHz memory. As for the processor itself, with the help of a BIOS update, lower memory timings, and a lot of hope, we managed a decent overclock. With an 8x CPU multiplier tied to a 333MHz bus, our final rating of 2,664MHz (2.6Ghz) wasn't half bad. But it obviously wasn't that good either.

Because of stability issues (during the overclocking), we had decided that a stable 333MHz bus was far better than some sketchy 340MHz bus. So that's where we'll be testing. We only have a few integral benchmark test we run for over clocked boards. Here they are.

  • RightMark memory
  • SiSoft Sandra memory
  • WinRAR
  • Doom 3

These should give use some good indications as to our overclocking efforts.






Well, it would seem that the Sapphire Pure CrossFire 3200 had a few instances of moderate success in the RightMark memory analysis during it's over clocked stage. And it would seem especially during the bandwidth testing. Even though the Sapphire fell behind with the other benches, it is possibly that a different board layout just happened to optimize the memory bandwidth. What about SiSoft Sandra memory testing?



A modest improvement. But we can see some better figures when it come to WinRAR.



But, again, it can't compete with its others when playing Doom 3.



So, what's an enthusiast to make of these numbers; both stock and over clocked? Perhaps there will be some enlightened thoughts within my conclusion.

Article Index

1.Introduction
2.A Chipset Revisited
3.Package and Impressions
4.The BIOS
5.Test Setup and Benchmarks
6.Business Winstone/Content Creation 2004
7.Right Mark and SiSoft Sandra Memory
8.HDTach and WinRAR
9.TMPGEnc, XVid, and LameMP3
10.Call of Duty and Doom 3
11.Comanche 4 and Halo
12.Jedi Knight and UT2K4
13.Overclocking
14.Final Thoughts

Submit our article to: diggDigg this! de.le.ciousdel.icio.us

Get updates when we publish new articles
Email Address:

(0.0342/d/barracuda)