Q6600 Overclock thread.

I'm also finding out that not RMA'ing my 680i when they had the offer was a big mistake. 680i AR boards suck for OC'ing Q6600s and apparently 1333 is as fast as that board will take me, which explains all the weird behaviors I've been seeing when trying to OC this chip. I can't get into windows reliably at any FSB above 1333 with this chip. I couldn't get it to do 3.1Ghz. I can POST up to 4GHZ and even get partially into safe mode under 1.5v. I'm waiting for my place in the EVGA trade up queue for the 780i. I hear those boards aren't bad for OC'ing Q6600s. The board is back ordered and I'm 1071 in line so I have a bit of a wait. 3Ghz *4 ain't too bad (Who am I kidding, it's great!).

I feel your pain. I had a 122-CK-NF63-AR which f***ing blew for oc'ing. At least for me it did. Wouldn't get my Q6600 past 3.1 without burpin, fartin, cussin, shitin n pissin everywhere.
 
Does anyone do other task like web browsing, watching dvd, emails while doing the prime95 test? Can you or is it best not to, with facts.
 
I feel your pain. I had a 122-CK-NF63-AR which f***ing blew for oc'ing. At least for me it did. Wouldn't get my Q6600 past 3.1 without burpin, fartin, cussin, shitin n pissin everywhere.

beautifully said.

i too had a 680i a1 revision that sucked balls for oc'ing. My maximus formula, however, is a whole different story.
 
Runing SMP Folding and playing games is all I care about.. And its has been stable for 48 hours on these settings so far.

If prime 95 says these settings are not stable and normal use and Folding says they are, then what you you think I am going to trust? A program that I am never going to run all the time, or things that I do all the time. Pretty clear to me.

Problem is F@H will run on a bad overclock and just submit bad work units, so it should matter if it is stable or not.

even though you dont play it, it would be nice to know your not wasting time folding and submitting bad WU.
 
Here you are.

Though I suspect the problem is more your motherboard than anything else. It was my problem, then I upgraded to the one I have now, which allowed me to get the current clock.

Thanks for posting this, i am having issues getting mine over 3.2 stable in an ip35-e :(
 
I have the same problem...my 680i is terrible for ocing my cpu and incredibly unstable too. That's why I plan to sell it and pick up a DFI P35 Dk :D. When/if I get a new mobo I'll post back with my oc.
 
Got mine running at 3Ghz not Prime stable. Fails after about 10-15 minutes on one core randomly. Everything else runs fine. Huge increase in performance. Like I spent 300+ dollars for a new chip when all I did was change a few settings. Gotta see what I can do over the next few days to get it Prime stable for hours of testing. Gonna tweak voltages and maybe re-seat my cooler to try for 6 hour Prime stability.

I have the feeling 3Ghz is the most I'll get outta this chip, being a B3, without making some major changes. Like a new case, better cooling, etc..

Fun stuff though, this overclocking. Pretty much my first serious attempt.
 
Well, to add to my 680i misery, it fried one of my Ballistix sticks. I ordered a set of 2x2GB Geil pc6400 4-4-4-4-12 sticks so I could cold boot with 4GB of memory since running 2GB sucks in comparison. The board has always shown difficulty with four sticks after being powered off. Wierd, anyway, I called EVGA and they offered to RMA my 680i for a new revision but I would lose my place in the queue for the 780i. I would be allowed to join again even though the program is now closed but I would be at the end of the queue. I was hoping they offer to just RMA my board and charge me for the 780i and ship it out ASAP. Oh well, I have to wait another month or so. Hopefully this board doesn't fry the new memory in the meantime. I'm just glad I'm back to 4GB, I was hurting with 2GB.
 
I just started to OC my system yesterday and ran Prime95 to stress the comp. I simply raised the FSB to 333 (3.0GHZ OC) at stock voltages. Everthing seems fine even with other apps running but then I let the system run overnight and at about 8hrs in threads 3 and 4 caught errors while 1 and 2 were still running fine. Temps peak at 58. Is this acceptable or if it's stable should there be absolutely no errors?
 
6 hours stable is fine. Some will say 12-48 hours but come on...... Who plays a game or uses an app that stresses all your cores 100%?
 
I went up to 3.96GHz yesterday on my Q6600 with 1.45V set in the BIOS. My computer was loud with the Tri-Cools all running max speed. Idle temperature was about 45C, and I didn't dare prime it. It was PI 1M stable and 3DMark06 stable. It did hit 68C during the 3DMark test so I don't even want to know what temperatures I would get trying to run prime. I then tried 4.05GHz with 1.475V and got a BSOD a few seconds after windows loaded. Called it a day after that. :D
 
6 hours stable is fine. Some will say 12-48 hours but come on...... Who plays a game or uses an app that stresses all your cores 100%?
If you plan on folding, don't bother. Apparently you can turn in bad units and then Steve & Kyle will come to your house to put their collective foot in your ass. :eek:

If you don't then your ass is safe, no harm done and enjoy your games. I've never had a problem with an unstable system while gaming except for the odd crash here or there. That alone is my motivation to make sure my rig is stable. I like to see Prime95 run for 24+ hours before I consider it truly stable.
 
Does anyone do other task like web browsing, watching dvd, emails while doing the prime95 test? Can you or is it best not to, with facts.

I do. I find that it's better to use the machine while stress testing in order to throw as much at the system as I can to get it to mess up. Prime is stressful on the system but it's basically using one type of calculation over and over. I like throwing different things into the mix to see how the system will handle it.

Since you are testing an unknown overclock I do not suggest doing anything that you can't afford to start doing over. I'll do web browsing, copy useless files around all over the place (if the overclock isn't stable it can lead to data corruption, the reason I just copy files I don't care about) and other things similar to that. If I'm going to be away from the machine I'll start up a copy of 3DMark to run looped which also helps to test how the system runs with the CPU maxed out and the case heating up due to the video card.

Basically, I find it very useful to run other software while Prime is going on in the background. I do not suggest installing any software or anything like that while stress testing an overclock. If it's not stable you could easily end up with a bad installation which can cause you trouble down the road.

 
I do. I find that it's better to use the machine while stress testing in order to throw as much at the system as I can to get it to mess up. Prime is stressful on the system but it's basically using one type of calculation over and over. I like throwing different things into the mix to see how the system will handle it.

Since you are testing an unknown overclock I do not suggest doing anything that you can't afford to start doing over. I'll do web browsing, copy useless files around all over the place (if the overclock isn't stable it can lead to data corruption, the reason I just copy files I don't care about) and other things similar to that. If I'm going to be away from the machine I'll start up a copy of 3DMark to run looped which also helps to test how the system runs with the CPU maxed out and the case heating up due to the video card.

Basically, I find it very useful to run other software while Prime is going on in the background. I do not suggest installing any software or anything like that while stress testing an overclock. If it's not stable you could easily end up with a bad installation which can cause you trouble down the road.
+1

I usually spend my time here and maybe play Peggle Extreme or an emulator. Since going quad, you can get away with it just about.
 
Well, i just installed CPUz and was suprised to see that the multiplier was shown at 6 x with my 333mhz FSB. My bios is showing 9 x 333mhz so i dont know what's going on.
 
I'm no expert, but that would sound like Intel's "Speedstep" which lowers the Q6600's multiplier from x9 to x6 during "idle moments".
 
I agree, my multiplier drops to 6 when idle. I really like that, since it gives everything a chance to cool down when I'm not pushing it.

I love the q6600's ability to multitask. I can play UT3 and record video on my tuner and I don't even feel it. Its also beautiful to watch 4 movies convert to ipod format at once without increasing encode time.
 
I do. I find that it's better to use the machine while stress testing in order to throw as much at the system as I can to get it to mess up. Prime is stressful on the system but it's basically using one type of calculation over and over. I like throwing different things into the mix to see how the system will handle it.

Since you are testing an unknown overclock I do not suggest doing anything that you can't afford to start doing over. I'll do web browsing, copy useless files around all over the place (if the overclock isn't stable it can lead to data corruption, the reason I just copy files I don't care about) and other things similar to that. If I'm going to be away from the machine I'll start up a copy of 3DMark to run looped which also helps to test how the system runs with the CPU maxed out and the case heating up due to the video card.

Basically, I find it very useful to run other software while Prime is going on in the background. I do not suggest installing any software or anything like that while stress testing an overclock. If it's not stable you could easily end up with a bad installation which can cause you trouble down the road.


Thanks. I thought it shouldn't throw off the stress test. Thats what I usually do at work, have prime95 running while doing other task. They're not data critical computers so I don't mind doing it and if it happens to completely die, I'll just build another one using the company left over parts. It makes me sick that we throw away hard drives, ram, cpu, and motherboards that are still functional. Granted they are not the best but still good, the hard drives especially because those are from servers. Yea I wouldn't want to do any kind of install or data sensitive task just in case it does crash on me. Thanks again.
 
I'm at 2.8ghz right now (ocing today) with no voltage changes and with my Zalman 9500 at 37.5-50% of its speed (silent). 53 degrees when running a CPU stress test. At stock speeds, the temp stays at 45 degrees during tests but with my fan speed at 2500 which is a bit loud for my taste. I wanna see how far I can go while still keeping things cool AND quiet.

At what point did you guys need to change your voltage?

My friend with a similar system and an Ultra-120 cooler + 120mm fan at 1400rpm had to take it up at 3.4 and 3.6 ghz to keep it stable during stress tests. His temps go slightly over 70 during those times though. He doesn't mind it, but I think that's a bit high.
 
Right when I first went to overclock I had to change mine from auto to 1.2v after I saw my chip go to 72°C on 1 core @ the stock 2.4ghz. I almost sheet ma pants. My motherboard wants to make the kids fat and slow by feedin em to much.

I wouldn't worry about 70°C on the G0 chips. I'm pushing 67°C when I stress my chip for 1/2 an hour with Prime. And its a B3. :eek: No crashes or abnormal behavior so far so I'm rolling with it.
 
I have a very picky Q6600, it idles at 38c, loads at max of 54c and at stock voltage 1.22v I am stable at 3.1ghz. I've had i up to 3.6ghz but at 1.4v it is hitting 70c as read by core temp and never would get 100% stable... So I put it back to stock volts and will be happy with 3.1 until I can get a better cooling solution... Hell that is 700mhz free.. I can't complain too much. I'll be curious what this box can fold. I'm going to get it started with just the normal smp client at first, while I'm working on doing dual VM's.

Well figured out my stability issue, it was a failing Hard drive... I had 3 in my spare parts and I guess one of them was bad and I've ordered a new one to ensure I don't spend a day rebuilding a machine to have it go down less than 24hrs later...



 
Whats the fastest on the stock intel cooler anyone has seen or heard of?
 
Whats the fastest on the stock intel cooler anyone has seen or heard of?

Too vague of a question. One person's "OK" might be another person's "Unacceptable".

FWIW: the highest clock rate I achieved with the stock cooler before getting into the 70's was 3.1. I also didn't let the AS5 cure and simply put the stock cooler on to have temps to reference against.

I'm sure others could go higher if they don't mind extremely high temps.
 
4.1ghz on stock cooler is the highest I've seen but I think the guy lived in the arctic and had sub zero ambient temps!
 
i've planted mine at 3ghz, on stock volts, (1.07 thanks to droop.) no real need to go faster.
idles at 32, never gets past 60 when stressing, and orthos prime and occt stable for 8 hours each
 
Nice overclocks so far. I have QX9650 and sometimes I wish I kept my old Q6600, but then again I can do 3.6ghz like you guys but with only 1.1875 volts needed to do so.
 
I can hit 3.6Ghz with my Zalman 9700 stable. run it at 3.0 to keep it cooler thou.
 
Abit IP35-Pro Rev 1.1 BIOS v14 (BoltModded) || [email protected] (Thermalright Ultra-120 Extreme [lapped] w/ Scythe S-FLEX SFF21F) || G.Skill F2-6400CL5D-4GBPQ 2x2GB DDR2 888 5-5-5-15 || MSI NX8600GT || OS: WD Raptor 150GB || Storage: Hitachi 400GB {Eventually will add 4x750gb Raid5 for storage} || CoolerMaster Real Power Pro 750w || Antec P182
Abit IP35 Pro BIOS Settings: 370x9 | Dram 1:1 | vcore 1.315v | DDR2 1.85v | CPU VTT 1.2v | MCH 1.25v | ICH 1.05v | ICHIO 1.5v | DDR2Ref 0% | GTLREF 67%
OC: Voltage & Temps Spreadsheet
 
I run my quad at 2.0ghz... it's a Q6600 and at stock was running 55/55/50/50 idle, and 69/69/65/65 on load during games. So now it's at 46/46/43/43 idle and haven't tried games since under clocking it.

I OC'd it at one point to up around 2.7... but that just wasn't a good experience.

I could ramble on, but I don't want to get off topic.
 
Ouch Pous. That's sad. Have you tried anything to lower your temps? Different cooler? Cooler reseat? Change of thermal compound? Change of voltage? Etc?
 
Ouch Pous. That's sad. Have you tried anything to lower your temps? Different cooler? Cooler reseat? Change of thermal compound? Change of voltage? Etc?

I'm really really new at this kind of thing. Rather than doing any of that, I purchased P180 case and AF7pro. They'll arrive tomorrow along with some MX-2 compound and arctic cleaner 2 step solution.

At that point I'm going to attempt to clean the stock stuff off the CPU and put on MX-2, then mount up the AF7 and put everything into the P180.

I couldn't get it to boot properly at stock speed with lower voltage.

With my old Centurion 534 case I'm planning to learn about modding, I've posted a thread in the Case/Mod section with my proposed mods.

Do you think I'll notice good improvement with the AF7 and the P180? Should I get different fans than the tri-cools? Should I add in the optional front and side fan to have a total of 5? There's so many options and configurations and such that it gets overwhelming because I want to learn everything all at once, and to try it all as well.

Thanks for your help.
 
....

Do you think I'll notice good improvement with the AF7 and the P180?

I believe you will net an improvement in your temps. I have a P180 and a Tuniq Tower that is using a Tri Cool instead of the stock fan. I get 35/34/30/30 idle and 62/61/60/60 on Prime95. My setup is not exactly quiet and I am sure the airflow is disrupted by cable mismanagement.

Should I get different fans than the tri-cools?

Most definitely. There are many fans such as the Yate Loon D12SL that output more CFM and are quieter than the stock Tri Cools. Search in this subforum to find exact models that can suit your cooling needs.

Should I add in the optional front and side fan to have a total of 5?

Well without getting into much detail you want to have as much intake fans as exhaust fans. Try to experiment and see what results you get.
 
Might be splitting hairs but if you are using fan filters to catch dust/pet hair you'll want slightly more intake fans than exhaust fans in regards to fan layout -- so you can control where the air comes in. IE You want the air pressure slightly higher inside the case so you aren't picking up dust and crud from every single joint or hole in the case. In this setup you'll be blowing air out all the tiny cracks and joints in the case and all the air coming in is coming through your fan filters and being cleaned.
 
When people report their vcore, are they giving their BIOS setting or are they using the measurement after vdrop and vdroop under load?
 
my vcore with vdroop is 1.20-1.22, my bios is set to 1.225, which means I'm losing .005 regardless and .02 on vdroop which is not horrible. My EVGA 680 loses .05. if they say VID then that is usually what the bios is set at...
 
Hey dudes, here are my results.

Q6600 with TRUE (thermalright ultra extreme)
Arctic silver 5
Maximus Formula mobo
Corsair XMS2 DHX 4x2gb ram modules


At 1,344 vcore I'm on 3,5ghz using Cooler master RC690 case with all fans installed except the top side and back 80mm one.
TRUE has two Noctuas P12 attached in push and pull mode. Temps are 58c for one pair and 54c for second pair of cores (uneven paste spread).
Idle temps are 34 for one pair and 31 for another.

Need to work on this machine :D
 
Back
Top