• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

Interesting...

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.

Net-trix

Registered
Joined
Aug 17, 2002
Location
Vancouver, Canada
A few weeks ago I posted a thread about not being able to get over a 206 FSB with a multiplier of 11:( I have an ABIT NF7-S v. 2.0...512MB Mushkin 3200 2,2,2 RAM and the system is watercooled and well ventilatd. I had tried a 10 multiplier but no luck posting:mad:
Someone suggested that I disable the CPU enhancement option in the BIOS. Success:D I am now running the system on a 230FSB with a 10 multiplier. Stock vcore...chipet voltage at 1.7...vdimm at 2.8...and 2, 2, 2, 11 memory timings! Thank you to all that gave suggestions to my last thread in the forum:cool:

Would love to post the WCPUID image...but dun know how...
 
Have you run any benchmarks with CPU enhancement disabled?

If you do I think you will see that your performance will not be much better than a slower FSB with CPU enhancement enabled.
 
Net-Trix - you got up to 230mhz with your memory at 2-2-2-11?! That's amazing. I take it you have the Mushkin Black Level 2 Pc3200 as that's the only one rated 2-2-2. I also have this ram, and anything higher than 220mhz, I have to lower the timings to 2-3-3-7. It does bench pretty close to the 2-2-2-11 at 220 though, so that's alright. You must've gotten a real sweetheart of a chip.

Also, as cphastings as mentioned, disabling CPU Interface, though it may allow you to achieve a higher fsb, your perfomance is usually much worse. CPU Interface sets your ram to 1T command rate, disabling it sets it to 2T (it probably does other things as well, but the command rate is well documented), which will considerably lower your performance. As cphastings mentions, you should probably run some benches with both and see what gives you the best performance. 220 for me with CPU interface benches considerably higher than 227 (the limit I've found with my stock nb cooler) with CPU interface disabled...
 
Hmmm....new experiment

I will have to run some benchmarks. Thank you for the advice. However, I have noticed that with 1.8 BIOS I am unable to boot with the 10 multi...that is why I have it disabled. Awww...and I was just getting excited too...:(
 
with bios 10 i was able to use the CPU interface at 230Mhz. So why don't you try it ?
 
Welcome to the forums! (dunno how you have an avatar as a new member, but oh well :) )

This is how you make a screenshot. Just hit the print screen key when you have what you want to show in the foreground. Hitting this key will capture whatever is on your monitor anc copy it to the clipboard. Next open photoshop or paint. Then just paste into a blank file. If you are savy and your app can do this, safe the image a a GIF or JPG. This will save a lot in file size since the default is usually a BMP.

HTH
 
Well...I enabled CPU interface...and went back up to 11 multiplier. I am presently at 11*217 and this produces SANDRA benchmarks that are way better than the 10*232 (CPU interface-disabled) setup. Thank you all for your help. I will be keeping my fingers crossed for a BIOS update that will allow me to have CPU interface enabled...a 10 multiplier...and 230 FSB (I can dream....).
 
At 11*217 I got artifacts in 3d gaming. At 11*215 I have no artifacts at all. Any suggestions as to how I can prevent artifacting while increasing FSB? Thanx again all!
 
Any suggestions as to how I can prevent artifacting while increasing FSB?

Asking a question like that begs the question of whether or not you have run any stability test programs like prime95 or 3Dmark loops.

You are talking about being on the edge of stability with artifacts in games. If you haven't run prime95 or 3dmark yet, you might be in for a suprise with stability.
 
Well..after benchmarking for days....I have decided to stick with my 11 multiplier...CPU interface enabled.

In experimenting I got my 9.5 multiplier to 230FSB with CPU interface on and 2,2,2,11 memory timings. However, I noted that although my SANDRA memory scores were at their highest...the arithmetic score was low. So now I am at 11*210 with 2,2,2,11 and have almost as high memory scores as 9.5*230 and much higher arimetic scores.

I want to go higher...and have gotten to 217FSB...but I notice artifacts in games...and it reboots sometimes. I am still on a 1.7 vcore....

Do these 2500+ love high vcore? Could this be my problem?
Thanks again for the help
 
The system was stable running Folding @home for a 24 hour period...so the system was rock solid in that regard.

That is good. But folding doesn't show the same errors as prime95 would. I'd recommend running prime95 to see if your machine is stable in that regard.

As for the vcore, I am at 1.8. I have read here that folks do between 1.8 and 1.9 as necessary to go higher. But I messed with my machine again this weekend and went up to 1.85, and still can't get past my 10x215 sweetspot. (i was testing with prime95). Plus my core temp shot up a whole lot too.
 
I must admit that I never used Prime95. I usually use [email protected] stress it with SANDRA running CPU and memory burn-ins in the background. I will try Prime 95. I have noticed that increasing my vcore causes lock ups with an 11 multiplier. I am not sure if this is because my 12v rail gets low (11.64) or if it is a heat problem. The whole thing seems to be a mystery....I can get to 9.5*230 (CPUI enabled) with 2, 3, 3,11 timings...no prob. I move to 11 multi and I am stuck at 11*210...2,2,2,11 timings....looser timings do not help. I guess I should just be content.. :) Maybe a later BIOS upgrade will cure my ails.
 
Intertesting finding about the higher vcore and your stability. I wonder if that is what is happening with mine. I have an enermax 350W PSU. Perhaps that is my problem with me not being able to go higher.

When you run Prime95, use the defaults, and run the torture test. Hopefully you won't get any errors. If your system is "flaky" it will throw an error almost immediately. The optimal amount of time to run it is at least 12 to 18 hours.

Good luck and let me (us) know how it goes! And I will fill you in on my investigation of your PSU/vcore theory. Does anyone else have any advice on this?
 
Using the Winbond tool...I have noticed that increasing my vcore causes my 12v rail to drop at least a tenth. I also notice that when the processor is not stressed the 12v rail can be almost 2 tenths higher (11.86 as opposed to 11.66). I was thinking of getting a better power supply, as my power supply is over three years old. Although it is an Enermax...back then the emphasis was on the +5v rail...as this was the main line to the processor. Now the P4s and XPs use the 12v rail to power the core....so I am thinking a new power supply probably has this in mind when engineered. Who knows...guess I will be able to post that soon.

Thanx again!
 
Back