OCZ 3200 Plat Rev 2 @ 275fsb Holy Cow!

[Clevor]

Arguably the reign of the mighty BH-5 might be coming to an end. I just beat eva2000's bh-5 unbuffered score. Well sorta. I beat him on Sandra Integer and he beat me on Float. He did it at a lower fsb and tighter timings, but I did it on 2.8v. So I guess yo might call it a tie. Check it out..

BigStan, how did you test for stability?

Also, you won't beat BH-5 even with more bandwidth, not at 3D benches anyway. CAS timings make up for a LOT. 2.5-3-3-6 versus 2-2-2-5? No contest, you will lose big time. See below:

292, 5:4, 2-3-3-6, partial PAT:

3DMark2003=12,750
Aquamark=72,876
3DMark2001=21900

292, 5:4, 2.5-2-2-5, no PAT:

3DMark2003=12,760
Aquamark=72,976
3DMark2001=22,200

Morever, Far Cry at extreme settings: 1600x1200x32, 4xFSAA, 16xAniso, Very High settings is playable at 2.5-2-2-5, but have to drop down to 1280x1024 at 2-3-3-6. How's that for a seat-of-the-pants difference? If I had to run no PAT at 2-3-3-6, the numbers wouldn't even be close.

BTW, the 2-3-3-6 is with PQI PC3200 and the 2.5-2-2-5 is with Ballistix PC4000. And from the numbers, you can see the video card is a 6800 Ultra .

 

[BigStan]

Those benchmarks are all graphics card heavy. eva2000 and I were looking at just the pure speed of the mem as rated by Sandra Unbuffered score. I won't get anywhere near those scores because I have an average ~non oc'd~ vid card

 

[flapperhead]

hey stan that is oc scores and sure id like to see some benchies

 

[Clevor]

Those benchmarks are all graphics card heavy. eva2000 and I were looking at just the pure speed of the mem as rated by Sandra Unbuffered score. I won't get anywhere near those scores because I have an average ~non oc'd~ vid card

Yes, but compare 2.5-3-3-6 versus 2-2-2-5 on the same video card, whether it be a Geforce2 32MB or 6800 Ultra 256MB, and you will see the same spread .

 

[BigStan]

Yes, but compare 2.5-3-3-6 versus 2-2-2-5 on the same video card, whether it be a Geforce2 32MB or 6800 Ultra 256MB, and you will see the same spread .

I'm one of those dummies that ordered that supposed Kingston BH-5 so I might have an opportunity to do that next week. I'll keep you posted.

 

[Reefa_Madness]

I'm one of those dummies that ordered that supposed Kingston BH-5 so I might have an opportunity to do that next week. I'll keep you posted.

Remember the possibilities KHX PC3200 with 2-2-2-6 timings (with BH-5) if defective, then maybe RMA for KHX PC3200U with 2-2-2-6 timings (with TCCD). Maybe all is not bad, huh?

That new TCCD based ram would be the closest product currently carried by Kingston, wouldn't it?

 

[BigStan]

Remember the possibilities KHX PC3200 with 2-2-2-6 timings (with BH-5) if defective, then maybe RMA for KHX PC3200U with 2-2-2-6 timings (with TCCD). Maybe all is not bad, huh?

That new TCCD based ram would be the closest product currently carried by Kingston, wouldn't it?

That thought is what tipped me off the fence to order it. I'm really hoping it is BH-5. That would be fun to play with.

 

[flapperhead]

cant go wrong for the price. and it really gives one the opportunity to campare chips

 

[BigStan]

hey stan that is oc scores and sure id like to see some benchies

Here is 260fsb with my 2.4 proc

Here is 260fsb with my 3.0 proc

 

[flapperhead]

thnx guy, i noticed the 3.0 scores a little higher. do u think that is from the higher cpu speed? or just maybe a natural variation in the software??

 

[BigStan]

thnx guy, i noticed the 3.0 scores a little higher. do u think that is from the higher cpu speed? or just maybe a natural variation in the software??

I seem to consistantly get a little higher scores when my 3.0 is in. But I also can get a 50 point spread just from running Sandra back to back. It's tough to tell how much is from the proc and how much is from software variation. But I think the proc does have a positive impact. I know it impacts other stuff. The fastest I got SuperPi to run with the 2.4 in was 40 secs. I can get the 3.0 to do it in 34sec. I just think that has to improve the way it handles the mem I/O.