• We’re currently investigating an issue related to the forum theme and styling that is impacting page layout and visual formatting. The problem has been identified, and we are actively working on a resolution. There is no impact to user data or functionality, this is strictly a front-end display issue. We’ll post an update once the fix has been deployed. Thanks for your patience while we get this sorted.

Finally B3 Q6600 overclocking success @ 3400 w IP35-E

R

Regalk

Golden Member
Using Crucial ballistix it was all about memory timings and I struggled with it for days
Now running at 378 X 9 1.40 vid set in bios
My system
Q6600 B3 at 3402 1.40v with vdroop 1.352
Scythe Rev B Ninja NOT Thermalright U120e AS5 (edited)
Abit IP35-E (maybe I may still like this board after all)
Sapphire HD3870
Crucial Ballistix 2G PC6400 4 4 4 12 at 2.05 and at 1.25 = 946 at 5 5 5 16 BUT the other memory settings were key to running at 378 FSB === tRFC 42, tWR 10, tWTR 5 tRRD 5 tRTP 7
All the voltages are at default other than CPU and DDR
OCZ gamestream 700W
1 TerraG WD
Creative XFI

Just running some of my battery of tests fror stability and it ran them all
3Dmark 2003 and 2005
UT 2004, Aquamark 3
SuperPi
Ran prime 95 for 15 mins
memtest
all fine so far
looks like I am good to go
Running the memory at AUTO was not working (i.e. AUTO for tRRD tRFC etc)
 
Wow. That's a good vcore setting for that stepping and speed.

I assume you're going to let it run PRIME95 until it fails or your moving toward a half'day's running? 15 minutes doesn't prove a lot . . . . yet. Still, if you put it through all those other software tests . . . .

Somewhere, there's a review for the Kentsfield (B3) making all sorts of whoopee-do about running it at 3.4-something Ghz with VCORE at 1.46V.
 
bump the voltage any higher and you know the system will randomly reboot.

Thats the one thing i hate about this board.

Change your settings to 425x8. That will yield better performance.

Keep your voltage the same. The B3's should be able to pull that off on 1.4

Thats what mine is at.
 
Yeah, eight hours minimum with Prime. Otherwise, with that B3 chip 3.4 is as good as it gets and more than satisfactory - 1 GHz overclock on a B3 chip is unusual. Aigomorla's right, a lower multiplier and FSB speed might work better. Nice to have the 9 multi, though.
 
Originally posted by: aigomorla
bump the voltage any higher and you know the system will randomly reboot.

Thats the one thing i hate about this board.

Change your settings to 425x8. That will yield better performance.

Keep your voltage the same. The B3's should be able to pull that off on 1.4

Thats what mine is at.
Click to expand...

The reason I am in the 3XX FSB range is that it refuses to go beyond 400 FSB (last time I tried it did not go past bios screen). I was hoping for 425 X 8 when I got the quad over the E4500/E2160 since those have FSB walls at around 417 to 420. I can go to 1.425 (so some room yet).

The Prime95 thing - Really depends - I have run it on several occasions before and it holds at 71/72C using other settings - will do so later. First I am going to swap out the Scythe Ninja Rev B for the Thermalright I have on the DS3L (my error above I forgot that I was using the Thermalrights on my DS3L and P5K-E) - the push pins may not be making the contact i want
Anyway I have to run Quake 4 and Halo for a few hours to see how it is going to hold up. 1.40 may be a tad low as i just noticed (tried 1.38 and it bluescreened after a few minutes in prime)
The point is the ABIT behaves differently that the DS3L/P5K-E boards - have to be specific with memory timings not auto. This wa sthe reason for my post since a lot of people who have issues may want to look at this area of their OC
Will update later as i continue testing
 
Originally posted by: LOUISSSSS
gotta prime way more than 15 mins. i once detected an error after 8 hours...
Click to expand...

Unless you are going to run your machine 24/7 gaming non-stop I would say you are fine. I went to a few local shops to see how they burn in machines - not even close to doing any real burn (they just run a few 3D game demos continuously). Our testing is much more rigid
I will try later just have to change the HS since the Ninja does not do justice
 
Hmmm - after a cold boot overnight it reverted to default? I noticed the memory speeds at 756 and immediately did a warm reboot but still at 266FSB!!!?. Restarted and went into BIOS but everything was OK so I just Saved and rebooted (did not change anything) and all OC settings are back. Ofc ourse it double booted this time which is my signal that it is holding the OC. Weirdnes.
I do not like this - more testing
 
Meh 8x425 is MARGINALY better then 9x378, it has been rigorously tested and posted about on this forum. It is only in sisandra, a synthetic benchmark, that you see any real gains. 9x378 will be easier on the mobo and the nb, and I'd def stick with that to get things stable first. Then perhaps, if you're bored and have time to waste, you could try 8x425.
 
With unlimited resources, I'd be buying more motherboards and processors, and the "experience" would leap ahead.

What I did notice, though, is that the C2D and C2Q have a whole different regime in comparison for FSB settings.

With my 680i board, I've had similar quirks in how the board holds a different multiplier settings. It seemed to be pretty stable with multi- of 8 and 360 Mhz, and I could get it to run stable at 372. But the board would reset to default when rebooted at those settings. No -- worse than that: it would keep the FSB and DDR settings, and the multiplier would reset to 9 -- a guarantee for instability, failure to boot into windows, etc. I think I tried one notch lower multi- or 7, pushing up close to 400, but decided to go back to the stock multiplier.

Anyway, MarcVenice is probably right about the FSB and Northbridge . . .

I really think you should do your testing this way: For a given VCORE and FSB, run the multi-core Prime95 for an hour. Change the FSB and try again to see if it will last that long, if not, bump the VCORE a notch -- tweaking it until it goes past an hour. If you get to a point where you think the vcore is not too high, temperatures are good, then test for 6 to 8 hours -- even longer. If it fails, bump the vcore up a notch, or drop the FSB a tad. The temptation is to just assume that such a slight tweak will assure perpetual running without failure, but it's best to test again, even if it's going to take another 8 or 10 hours.

I think the failure rate for any given setting combination is Poisson distributed. If the vcore is insufficient or FSB too high, it will be more likely to fail early, and for each hour there is no failure, the probability of failure for the next hour is less. So if it fails after several hours, it's worth moving the settings in the right direction a tad more until you could theoretically run it day and night with no problem.

Unless you switch to peak over-clock settings just for demonstration purposes, it just pays to take the time and patience to assure rock-solid stability.

 
Originally posted by: BonzaiDuck
With unlimited resources, I'd be buying more motherboards and processors, and the "experience" would leap ahead...

Missed the point?!

Anyway, MarcVenice is probably right about the FSB and Northbridge . . .

Already knew that .. not really questionng this logic. However I like to know that I have the flexibility as i have on the P5K-e which is at 463FSB

I really think you should do your testing this way: For a given VCORE and FSB, run the multi-core Prime95 for an hour. Change the FSB and try again to see if it will last that long, if not, bump the VCORE a notch -- tweaking it until it goes past an hour. If you get to a point where you think the vcore is not too high, temperatures are good, then test for 6 to 8 hours -- even longer. If it fails, bump the vcore up a notch, or drop the FSB a tad. The temptation is to just assume that such a slight tweak will assure perpetual running without failure, but it's best to test again, even if it's going to take another 8 or 10 hours.

Know all of this already -I use more than prime95 when burning in

I think the failure rate for any given setting combination is Poisson distributed. If the vcore is insufficient or FSB too high, it will be more likely to fail early, and for each hour there is no failure, the probability of failure for the next hour is less. So if it fails after several hours, it's worth moving the settings in the right direction a tad more until you could theoretically run it day and night with no problem.

The above was well put - excellent stuff (I forgot about poisson distribution from my high school days)

Unless you switch to peak over-clock settings just for demonstration purposes, it just pays to take the time and patience to assure rock-solid stability.

yeah we all know that - took >10 years from my life
Click to expand...

 
Is anyone benefitting from the memory settings - should be equally applicable to other boards. I have since dropped to 3350 at 1.38vcore and now prime95 is running (at 1.4 cpu was getting up to 73C a little too hot - the Thermalright will soon take care of that).
Besides thermalright anyone has any good ideas on coolers
Been down the road with Zalman 9700 LED, Big Typhoon, Thermalright Ultra 120-e, Scythe Ninja, Scythe Mine - what else?

I will check out some reviews of the Tuniq tower
 
regal i wasnt kidding about the board randomly rebooting at votlages higher then 1.4

if you seriously dont believe me PM PCTC2. Hes the one that brought this bug to my attention. If you keep on trying to debug that OC, your only going to pull your hair out like he did.


Also Marc, i WCG and 4XXx8 yields more points then 3XXx9. Thats how i know its faster. Maybe not to the average program, but if you randomly encode, the 4XXx9 will yield faster times.

Also P35's breaking 400fsb is pifft. Almost given... Except the IP35-E, unless you can do it with lower voltage then 1.4!!!!

GAH! 300 dollars in h2o cooling limited by a 89 dollar budget board. what the hell was i thinking...
 
could you please post some pics of your bios settings, every page if possible, thats an amazing OC and I would like to do it. I have a different board then you, but everything else is similar.
 
Originally posted by: Regalk
Hmmm - after a cold boot overnight it reverted to default? I noticed the memory speeds at 756 and immediately did a warm reboot but still at 266FSB!!!?. Restarted and went into BIOS but everything was OK so I just Saved and rebooted (did not change anything) and all OC settings are back. Ofc ourse it double booted this time which is my signal that it is holding the OC. Weirdnes.
I do not like this - more testing
Click to expand...
Are you running bios 14? It has a qlitch that when you completely power down and then back up it will reset your CPU to stock. Flash back to bios 12, or you may want to try the just released bios 15. Link
Thanks to Heidifirst for pointing out the new bios in the Abit IP35-E review thread.

Originally posted by: Regalk
Is anyone benefitting from the memory settings - should be equally applicable to other boards.
Click to expand...
This thread may be helpful. Keep in mind though that the benchmarks were synthetic and may not reflect real life application results. I'm still trying to get my son, the current computer's owner, to run some 3D rendering benchmarks at different ram settings.

 
Somewhere, there was a statistical distribution posted for B3 VIDs. I'd be interested in seeing how yours fell into that sieve. And that P35 board -- there may be something "better" about it -- can't tell.

The real question: How long it will run that way in days or months before there's some "problem." But if the REAL voltage is close to your "set" and "monitored" values -- a good pile of risk has vanished from the picture . . .
 
Originally posted by: aigomorla
regal i wasnt kidding about the board randomly rebooting at votlages higher then 1.4

if you seriously dont believe me PM PCTC2. Hes the one that brought this bug to my attention. If you keep on trying to debug that OC, your only going to pull your hair out like he did.


Also Marc, i WCG and 4XXx8 yields more points then 3XXx9. Thats how i know its faster. Maybe not to the average program, but if you randomly encode, the 4XXx9 will yield faster times.

Also P35's breaking 400fsb is pifft. Almost given... Except the IP35-E, unless you can do it with lower voltage then 1.4!!!!

GAH! 300 dollars in h2o cooling limited by a 89 dollar budget board. what the hell was i thinking...
Click to expand...

Hey Aigo, I am pretty sure it was me who brought the bug to your attention. Anyway, I did not come in here to correct that, but rather confirm it. If you look in my sig, I do run a higher Vcore, but it will shutdown on me if I prime it for 15 minutes... I found that with a B3 the highest Vcore I can set without it shutting down on small FFT is right around 1.46/1.47 BIOS. The higher I increase the Vcore, the faster the systems takes a crap when priming it. It has nothing to do with the cooling or chip itself, but rather, the IP35-E.
 
Originally posted by: aigomorla
regal i wasnt kidding about the board randomly rebooting at votlages higher then 1.4

if you seriously dont believe me PM PCTC2. Hes the one that brought this bug to my attention. If you keep on trying to debug that OC, your only going to pull your hair out like he did.


Also Marc, i WCG and 4XXx8 yields more points then 3XXx9. Thats how i know its faster. Maybe not to the average program, but if you randomly encode, the 4XXx9 will yield faster times.

Also P35's breaking 400fsb is pifft. Almost given... Except the IP35-E, unless you can do it with lower voltage then 1.4!!!!

GAH! 300 dollars in h2o cooling limited by a 89 dollar budget board. what the hell was i thinking...
Click to expand...

Same here... I have that remorse right now... :-(

 
Originally posted by: ArchAngel777
Originally posted by: aigomorla
regal i wasnt kidding about the board randomly rebooting at votlages higher then 1.4

if you seriously dont believe me PM PCTC2. Hes the one that brought this bug to my attention. If you keep on trying to debug that OC, your only going to pull your hair out like he did.


Also Marc, i WCG and 4XXx8 yields more points then 3XXx9. Thats how i know its faster. Maybe not to the average program, but if you randomly encode, the 4XXx9 will yield faster times.

Also P35's breaking 400fsb is pifft. Almost given... Except the IP35-E, unless you can do it with lower voltage then 1.4!!!!

GAH! 300 dollars in h2o cooling limited by a 89 dollar budget board. what the hell was i thinking...
Click to expand...

Hey Aigo, I am pretty sure it was me who brought the bug to your attention. Anyway, I did not come in here to correct that, but rather confirm it. If you look in my sig, I do run a higher Vcore, but it will shutdown on me if I prime it for 15 minutes... I found that with a B3 the highest Vcore I can set without it shutting down on small FFT is right around 1.46/1.47 BIOS. The higher I increase the Vcore, the faster the systems takes a crap when priming it. It has nothing to do with the cooling or chip itself, but rather, the IP35-E.
Click to expand...

oh shoot was it you?

Blah... sorry archangel, for some wierd reason i thought it was pctc2.


 
Originally posted by: aigomorla
Originally posted by: ArchAngel777
Originally posted by: aigomorla
regal i wasnt kidding about the board randomly rebooting at votlages higher then 1.4

if you seriously dont believe me PM PCTC2. Hes the one that brought this bug to my attention. If you keep on trying to debug that OC, your only going to pull your hair out like he did.


Also Marc, i WCG and 4XXx8 yields more points then 3XXx9. Thats how i know its faster. Maybe not to the average program, but if you randomly encode, the 4XXx9 will yield faster times.

Also P35's breaking 400fsb is pifft. Almost given... Except the IP35-E, unless you can do it with lower voltage then 1.4!!!!

GAH! 300 dollars in h2o cooling limited by a 89 dollar budget board. what the hell was i thinking...
Click to expand...

Hey Aigo, I am pretty sure it was me who brought the bug to your attention. Anyway, I did not come in here to correct that, but rather confirm it. If you look in my sig, I do run a higher Vcore, but it will shutdown on me if I prime it for 15 minutes... I found that with a B3 the highest Vcore I can set without it shutting down on small FFT is right around 1.46/1.47 BIOS. The higher I increase the Vcore, the faster the systems takes a crap when priming it. It has nothing to do with the cooling or chip itself, but rather, the IP35-E.
Click to expand...

oh shoot was it you?

Blah... sorry archangel, for some wierd reason i thought it was pctc2.
Click to expand...


Trust me, I have the same problem at times. I forget who said something and state the wrong person. I think SerpentRoyal got on my case for that mistake last time and he then tried to invalidate what I said because I made a mistake on who actually mentioned it.
 
Originally posted by: ArchAngel777


Trust me, I have the same problem at times. I forget who said something and state the wrong person. I think SerpentRoyal got on my case for that mistake last time and he then tried to invalidate what I said because I made a mistake on who actually mentioned it.
Click to expand...

lol.. either way im an idiot for not double checking my requst page when you posted this. I should of...

Im chugging this board this weekend. I cant stand it anymore.
 
Originally posted by: aigomorla
Originally posted by: ArchAngel777


Trust me, I have the same problem at times. I forget who said something and state the wrong person. I think SerpentRoyal got on my case for that mistake last time and he then tried to invalidate what I said because I made a mistake on who actually mentioned it.
Click to expand...

lol.. either way im an idiot for not double checking my requst page when you posted this. I should of...

Im chugging this board this weekend. I cant stand it anymore.
Click to expand...

I hear you... I am too lazy and cheap to redo my computer right now though. Especially since it works fine for what I use it for. I know one thing, my next rebuilt is going to be quality everything and no skimping.
 
Originally posted by: boglwe
could you please post some pics of your bios settings, every page if possible, thats an amazing OC and I would like to do it. I have a different board then you, but everything else is similar.
Click to expand...

I have already posted the memory settings - this is all you really need the rest is nothing special
If you still want pm me and I can send yu some pics
 
You must log in or register to reply here.

TRENDING THREADS

Back
Top