G.Skill GBHZ 2.0-4-4-5 @ 273.5 in P4P800

Printable View

05-05-2006, 01:31 PM
Axylone

1 Attachment(s)

G.Skill GBHZ 2.0-4-4-5 @ 273.5 in P4P800

I'm thinking that the motherboard is holding up the memory, allowing such a low CAS. The tRCD and tRP (x-4-4-x) cannot be lowered without losing stablilty, and I have to drop it down to ~230. The board is a P4P800e-Dlx (no mods yet).
This is OCCT and BF2 stable.

32m 12.750s w/ 2.5-4-4-5 @ 273.5 earlier, so not much difference in super-pi times.
05-05-2006, 02:44 PM
_damien_

CAS 2.0 is really tight for that speed. Not sure if I've seen 2.0-4-4 in combination before.

BTW, have you compared the performance difference between 1:1 and 4:5? I don't think running the memory faster than the FSB yields any gain at all when running in dualchannel. The NB-bufring will hurt latency, and the FSB will limit the effective throughtput anyway.
05-05-2006, 03:12 PM
Axylone

It's at 4:5 because the CPU only goes to about 250. I can't compare 1:1 to 4:5 because a locked multiplier, but the ratio would probably explain the weird timings.
05-05-2006, 03:17 PM
_damien_

Quote:

Originally Posted by Axylone

It's at 4:5 because the CPU only goes to about 250. I can't compare 1:1 to 4:5 because a locked multiplier, but the ratio would probably explain the weird timings.

What? I know the CPU is locked, but if the FSB-speed remains constant, the 1:1 ratio would make the RAM run slower. The setting is called FSB : DRAM ratio, i.e. 4:5 makes the RAM run 25% faster than the FSB, not the other way around. I was suggesting that you run the RAM and FSB at 218 MHz.
05-05-2006, 03:19 PM
quent

hum..

Could you post a Cpu-z validation, please.

quentin.
05-05-2006, 06:13 PM
Axylone

sure: http://valid.x86-secret.com/show_oc?id=91632

Damien: Ah, I misread your post. I assumed you meant the difference between 1:1 and 4:5 at the same memory speed.
05-06-2006, 02:25 AM
farksy

Hmmm nice clocking ! Agree with_damien_ , 2-4-4-5 sure is a odd combination !
How about 3-4-4-8 ?

To me honest, when i first read the timings i was unsure whether this was DDR or DDR2.
05-06-2006, 02:26 AM
quent

Thank you ;)
05-06-2006, 02:42 AM
giorgioprimo

may you please test them @ cas 3?

Thanks ;)
05-06-2006, 04:47 AM
Repoman

See, it's so weird how intel has all these fantastic memory results on DDR1.. wonder what the difference is that AMD can't?

The only really amazing tight timing stuff, like 2-2-2-5 on micron at 250, 2-3-2-5 on infineon at 246 with cheap ZX at that, and this, have been on intel.
05-06-2006, 10:33 AM
filmbot

I didnt even think it was possible for UCCC to run at CAS 2 at any speeds.
But I guess Im wrong.
Nice results, this reassures me that I should have no problem when I set up my P4P800 SE with a Dothan and my HZ.