View Full Version : The real "250"

10-27-04, 06:40 AM
Even though my scores were good on this FX build...too me, the OC'd scores still looked alittle on the weak side.
It didn't take me long to find the "problem". Asus bios are always strange places too me and I didn't have the RAM dividers set right for 1:1 (the above scores were running @ 250 HTT : 200 MEM), I was under the impression that the "Auto" DIMM clock set 1:1...but it actually locks the DIMM clock at the CPU default FSB...
After finding the right setting I was able to get 1:1 @ 250 2.5 3-3-7 T1 @2.8v and run some new scores.....

Sandra :
mem: 6600+
cpu: 12000+

Aqua3 73900+

3dm03: 13800+
3dm05: 5600+

:( now some sad news....I rebooted the system and was planning to run the rest of the scores and post the screenie shots I saved....but on the reboot I was greeted with *tick*tick*tick*tick*tick* yep...it was a sudden failure of my boot drive (36gb Raptor) .....that's why the scores arn't exact, I'll post that info when I get rebuilt.

10-27-04, 07:40 AM
Bummer... Are you going to call WD and get them to replace that dead Raptor? The timing on that failure... well that's some bad luck there.

Bill The Cat
10-27-04, 07:46 AM
That's a huge bummer. If I recall WD isn't known for their quality control.

10-27-04, 08:05 AM
5 year warranty on the Raptors....WD's Advanced RMA was a piece-o-cake :)

10-27-04, 08:07 AM
I had one of my WD drives in my sig fail on me, they cross-shipped the replacement and I had it in 2 days. Not too bad, if you ask me.

Pain - what SATA controller are you using, the VIA or the Promise? If I remember correctly, the VIA controller on our boards does not use the PCI bus and therefore is overclocked even when the PCI lock in on. The Promise controller does use the PCI bus. Maybe your drive was being run out of spec for a while and just finally gave up?

10-27-04, 04:15 PM
Nope, I run the boot drive and storage off the Promise, which is the one that failed....
I believe the VIA SATA controller is effected by the true HT bus, with the HT set to 250 and the multi set at 4x that runs it right on spec. Anyway I run two 74gb Raptors in RAID"0" off the VIA controllers and they have no issue what so ever when overclocked...
This was a simple mechanical failure of the drive at a bad time.....

10-27-04, 04:20 PM
That sucks :( Let us know how the O/C goes when you get your drive back. OCZ told me they'd replace my PC4000 with 3200EL 2-2-2-5 so I'm a little giddy :D

10-27-04, 04:22 PM
5 year warranty on the Raptors....WD's Advanced RMA was a piece-o-cake :)
yep, they offerd that to me also, however i find it werid that im seeing alot of people with rapors are dieing, hmm :o

10-27-04, 05:02 PM
OCZ told me they'd replace my PC4000 with 3200EL 2-2-2-5 so I'm a little giddy :D
I'd be giddy to. :)

Congradulations, I'm curious to find out if the situation changes when you try to overclock with that memory. Mine does 250 1:1 on only 2.7v (2.5-3-3-8).

10-27-04, 09:53 PM
Here's how things turned out......the GT was clocked at 400\1100 on all tests.


I can not get over what this memory...that would barely run on the MSI K8N Neo2 is doing on the Asus board...The bandwidth is huge :)

10-27-04, 10:02 PM
Pain could you publish 3dmark 2001 , score id like to compare :D

10-27-04, 10:39 PM
He did. Look at the first image in his post above.

10-27-04, 10:44 PM
Hmm i must be blind cose i don't see Linky to futuremark for compare .
Riptide you misunderstood my request *publish* is 3dmark term to save score online for others two compare.

10-27-04, 10:47 PM
You're right I did. My bad. :spank:

10-27-04, 11:14 PM
I haven't posted(published) any 3dmark scores with the FX yet, That link I was posting http://service.futuremark.com/compare?2k3=3117677 was with the P4EE and vid card maxed out maxed (not close to where I could run it every day)....I've got too "learn" this platfrom, before I do that. The scores above are about at what speeds I'll be running this setup at for everday use. And I'm liking it :D

10-28-04, 10:32 AM
It seems my K8N2 doesn't really support 0.5 multpliers :( 10.5 x 250 really = 10x262 MHz, which explains why my computer was crashing at 250x10.5 but is stable @ 11x238. I know due my memory running only 2.85v i can't do 266Mhz or near that. :(

And the only time i could run was without optimised setting which reduced my bandwith by 1GB/sec So my 3D mark results are useless due to that :(

10-28-04, 11:01 AM
Just an FYI, but half multipliers won't increase bandwidth. IOW, 250x10 = the same memory bandwidth as 250x10.5.

10-28-04, 02:22 PM
Nope that wasn't what i was on about, the MSI board had 0.5 multipliers added however they don't work correctly and rather then increasing the multipliers it increases the bus speed and lies about it being 10.5x whatever. As for the bandwith it is an option in the bios called optimised, not sure what it does but it gives a huge boost to bandwith. As it it's an option for T1.

10-28-04, 02:24 PM

Interesting read about Neo2.

10-28-04, 02:43 PM
Just an FYI, but half multipliers won't increase bandwidth. IOW, 250x10 = the same memory bandwidth as 250x10.5.

Thats correct.