Can a Q6600 with VID 1.325 hit 3.6GHz?

[eldredpe]

My setup:

Q6600
Abit IP-35 Pro mobo
4 GB (2x2GB) G-Skill DDR 800 RAM


My VID is, as stated in the topic, 1.325. I've got it to 3GHz without any problems, but can't seem to push past that with stable results. I've tried for 3.2 and 3.6 at voltages as high as 1.4925 and though I can boot into XP, when starting up Prime95 the 4th core instantly drops out and the system is prone to blue-screening.

I'm using auto-voltages in the BIOS for my current stable 3GHz OC, but bumping the clock up past it with auto settings still produces unstable results (this includes running it at 3.15Ghz with the ram underclocked, 3.2Ghz with 8x400fsb, and 3.6Ghz with 9x400fsb.)

Temperatures have not been an issue as of yet. I'm using a xigmatek SD-1283 (I think that's the model) cooler, and at 3Ghz my core temps idle at 20-25 degrees and hit no more than 45 degrees under prime 95 load.

My goal is to hit 3.6Ghz, but at this point, I'd just like to be able to push past the wall of 3.0Ghz I seem to be at.

Any suggestions?

Also, when manually playing with voltages, I've increased voltages on the processor only. I didn't feel I was knowledgeable enough to mess around with mobo/ram voltages.

 

[outlw6669]

Yes, it is possible. My Q6600 with the same VID hits 3.6Ghz 100% stable with 1.55V. Any less and I will loose a core under Prime95 after 1-3min. Of course not all CPU/Mb's are the same. Your CPU or Mb may not be up to the task of running past 333Mhz on the FSB.

Here is where I would start.
Check to see if a newer BIOS for your Mb supports 1600FSB. If it dose, the auto voltage settings should supply just as much power as the manufacture recommends at that speed. If not, leave it on auto and check the voltage from windows. After your CPU is 100% stable, start taking it down a notch at a time. Re test every adjustment until you find the minimum stable voltage.

Add more voltage. The commonly accepted max voltage for the 65nm Core 2 architecture is around 1.5V. Depending on your cooling, it is possible to add some more but be careful with your temps. I run my Q6600 at 1.55V only when I am playing a demanding game but never seem to have to much of a temp spike compared to my normal speed of 3Ghz.

You can run your RAM at 1:1, stock voltage and timings when you push to 3.6Ghz.

If you are still unable to reach 3.6Ghz, raise your CPU and Nb voltage, drop your RAM to 1:1 and start raising your FSB 5-10Mhz at a time. Give a quick 10min Prime95 test every bump. If it passes, raise it 5-10Mhz more until you find your hardware's wall.

Best luck.
.

 

[lucuis]

Without looking it up, i'm going to assume your mobo natively supports a rated FSB of 1333mhz. So by increasing your FSB past that you've gone into the realm of unknown. Where things are chaotic, and evil, or just plain wrong.... Oh i mean, you probably just need to bump the voltage up on the Northbridge

 

[FHDelux]

is your CPU a g0 stepping? I'm assuming it is based on the temps you listed but if its a B3 im not sure you will make 3.6 at all. I agree with the above poster about updating the bios, a lot of those p35 motherboards have come out with new bios revs that officially support high fsb settings. One thing you could try, lower the multiplier of your chip to 8, 7, 6, or whatever, and then raise up the fsb of the motherboard to get the stock clock speed of your chip. That way you can find the max fsb of your motherboard, so you know what your limiting factor is. Right now you are working with two unknowns which is tough, how high fsb does the motherboard support, how high can the chip go. Typically i have gotten G0 Q6600s to 3.6Ghz 100% stable with 9X400 1.425 - 1.45 vcore. Good luck in you quest!

 

[Verillion]

Don't assume that your VID has anything to do with overclocking.

VID is a lot more complicated than everyone thinks, and has absolutely nothing to do with how a cpu will oc as far as anyone has been able to conclusively show. You can get different VID readings on a cpu in different motherboards, depending on what voltage regulation chip it uses. This alone is enough to make VID useless for predicting oc'ability.

The VID is hardcoded into each individual cpu, but it's not just one number; it's a range of 6 different VID levels and these levels are specific to each individual cpu. VID is not static and changes with power requirements. VID can fluctuate almost constantly to try to maintain the level programed into it during manufacturing.

I was just curious what CT actually reports, is it reporting the VID being used at the time, max VID, etc.. VID is read from a sense line off one of the cpu pads, it seems like CT would have to be reading whatever VID level is in use at that time. :coocoo:

Here is an explaination of VID on C2D processors, this is the short version, the whole thing explaining all the levels, how it's read, etc.. is available in the data sheets for the cpu's on Intel's site..

Individual processor VID values may be calibrated during manufacturing such that two devices at the same core speed may have different default VID settings. This is reflected by the VID Range values provided in Table 5.

The processor uses six voltage identification signals, VID[6:1], to support automatic selection of power supply voltages.

The processor provides the ability to operate while transitioning to an adjacent VID and its associated processor core voltage (VCC). This will represent a DC shift in the load line. It should be noted that a low-to-high or high-to-low voltage state change may result in as many VID transitions as necessary to reach the target core voltage.
Transitions above the specified VID are not permitted. Table 5 includes VID step sizes and DC shift ranges. Minimum and maximum voltages must be maintained as shown in Table 6 and Figure 1 as measured across the VCC_SENSE and VSS_SENSE lands.

 

[Evilonigiri]

Lupi's VID didn't change in Coretemps when he tested his 10 Q6600 on 3 different motherboards. After doing some OCing on all of them, he saw that there was negative correlation between VID and OCing potential. Of course he ran into 1 or 2 sample that had low VID but didn't OC well. I suppose from this you can conclude that a low VID chip has a better chance at getting a better OC.

Anyhow, 3.6GHz should be possible for most Q6600s. How much voltage you need is a different issue. With Lupi's ultra low 1.200VID Q6600, he only needs 1.3125v to get Prime95 stable at 3.6Ghz.

 

[Shadow703793]

lol, lucky lupi. Any ways, has any one ever gotten a .85v VID CPU? The Intel website lists VIDs from .85v-1.5v ( http://processorfinder.intel.com/details.aspx?sSpec=SL9UM ) for the Q6600. Most other CPUs have a minimum of .85v, so I am wondering if any one has a .85v VID CPU?

 

[Evilonigiri]

Quite doubtful. Lowest I've seen was near 1.125V I think.

 

[Shadow703793]

Also another thing I have noticed is that the website reviews for CPUs NEVER tell you the VID. All they show are the CPUZ screen shots,etc. To me VID is one of the biggest OCing factors, ranking in the same range of heat, CPU architecture, cooling,etc.

@Evilonigiri: Don't you have a job? Most ppl. are at jobs at the current time. :lol:

 

[jsc]

lol, lucky lupi. Any ways, has any one ever gotten a .85v VID CPU? The Intel website lists VIDs from .85v-1.5v ( http://processorfinder.intel.com/details.aspx?sSpec=SL9UM ) for the Q6600. Most other CPUs have a minimum of .85v, so I am wondering if any one has a .85v VID CPU?

I did not realize that there was so much variation. Mine is 1.2625 volts. Using a TRUE, I can reach 3.0 GHz with stock voltage, 3.3 GHz at 1.300 volts, and 3.6 GHz at 1.45 volts (BIOS setting, droops to 1.400 volts on a GA-EP35-DS3P motherboard, all Prime 95 stable after 24 hours.

eldredpe:
I try to avoid "AUTO" settings whereever I can.

 

[Evilonigiri]

Well, it could very well be something that's accompanying low VID that helps OC, so perhaps it's not the VID that's affecting the OC but the part that's accompanying low VIDs. See, in some cases low VID doesn't translate into good OC.

But then again, what do I know?

 

[roadrunner197069]

You could have issues with not disableing speedstep and other features that help overclock when disabled.

 

[eldredpe]

I do believe my mobo (at least from what the box says) has only a 1333 fsb. I'll check for an update.

Also, how much extra voltage is safe to give to the northbridge?

And to the Q6600? It's rated for up to 1.5 I know, if I push a bit past that is it going to fry the chip?

 

[Evilonigiri]

I do believe my mobo (at least from what the box says) has only a 1333 fsb. I'll check for an update.

Also, how much extra voltage is safe to give to the northbridge?

And to the Q6600? It's rated for up to 1.5 I know, if I push a bit past that is it going to fry the chip?

Nope, but it's not exactly a good thing to do. People has run them at 1.6v or even higher, but those chips aren't expected to last more than 2years.

For your NB voltages, increase it by .1v or so. If it's not enough just increase it.

 

[htoonthura]

Nope, but it's not exactly a good thing to do. People has run them at 1.6v or even higher, but those chips aren't expected to last more than 2years.

For your NB voltages, increase it by .1v or so. If it's not enough just increase it.

what vcore value should we use. bios or cpuz.

 

[Qwakrz]

My Q6600 B3 with 1.3v VID can only do 3.4GHz at 1.55v (1.6v in BIOS)

As with all overclocks what you get from the chip will change depending on Mobo, BIOS and luck.

Chip a) with 1.35v VID may be a very high overclocker but needs high voltage to be stable whereas Chip b) may only need low volts to work but a FAB error means it wont overclock more than 200MHz and chip c) needs low volts and can overclock beyond beliefe, you just never know till you try it.

 

[Shadow703793]

what vcore value should we use. bios or cpuz.

I use CPUZ cause it shows the effect of vdrop or vdroop.

 

[Evilonigiri]

what vcore value should we use. bios or cpuz.

Well that depends. For a moderate OC for the Q6600 G0 (3Ghz - 3.4GHz), I'd set it to 1.4v or so for the time being. Then I'd minimize it when everything is truly stable.

 

[iluvgillgill]

it really bepend what cooler you using.i use the zalman 9700LED i only got to 3.4@1.5v

http://45s0cg.bay.livefilestore.com/y1pedVwQhARuXFPOFPV3yhBqcVx_euWYefPg7s4diAvx522es6picEY7wRj6Muyu7gupXkJ7aK1fqp3gJbILBk23A/final%20oc.JPG

but 1 thing this CPU i was testing got a bum core inside which doesnt matter how much voltage i apply.even 1.65V i couldnt get it last 1min in prime.

you took the risk in OC.

 

[knotknut]

What is the best method to determine what a C2D default VID is ?

 

[Lupiron]

Core Temp will show only the VID, and it should never change with voltage. Old old core temp used to change. New one will never change. Your VID CAN NOT change. Its a totally fixed element.

Since the voltage required as you OC is a sliding scale, that turns sharply upwards at around 3.6 Ghz... the lower the VID means the Lower over all voltage required.

As in, 1.2000 VID gets 3.6 @ 1.312 volts, loaded. A 1.3000 would not need simply the .1 extra between the VID, it nees almost half of that, extra. So 1.3000 would need like 1.43 minimum, up to x.xxx dependant of other factors. NB, for instance, can cause a single core to fail, no matter the VCore. And a 1.3250 needs 1.45-1.46.

I have used other things to stabilize my 1.3250 VID chip near 1.42 volts, but thats as low as it can go!

I will add more later!

--Lupi

 

[Lupiron]

A B3 will automatically be harder to OC.

And listen up! YOU all are the ones (some of you.) Who are scared to use high voltage! Obviously for all those scared people, VID should mean everything!

The q6600 series limit within spec is near 3.8 Ghz. Only good chips, and mid or lower VIDs will get it.

Not all will get it, even my super low VID, because that is the techs limits. Sure, at over 1.5000 you can squeeze out a few more Mhz, but hell... OCing is somewhat about being safe. So what if I could prolly be 4.2 stable at 1.6000 volts loaded. How long do you want your computer to last?

I assume OCing is not only finding the balance where everything works together at a higher speed. Its getting that balance, and having it last as long as a normal computer would!!!

Anywho, look for my old VID threads for the Data, its all there.

--Lupi

 

[eldredpe]

After upping the north bridge voltage by about .1 I've been able to get a stable 3.4ghz out of my Q6600 (378x9 with the RAM multiplier at 1.2, so it's overclocked to ~900 mhz). I've got cores dropping out on me in prime95 when this number is upped to 3.6ghz still. Voltage is currently 1.45.

Upping the voltage as far as 1.5 still has a core dropping out quickly at 3.6ghz. Does the northbridge need even more power? How much is it safe to give to the northbridge without a dedicated NB cooler? Should I just resign myself to 3.4ghz with good temps at 1.45 and stop overclocking, or is there still hope for a fairly safe 3.6?

Also, Is the RAM OC'd to 900ish alright? The voltages were left at stock, and I did a pass with memtest86+ with no errors. Should I be bumping up the voltages, or is it fine as it is? The RAM is 2x2gb G-Skill. Is the extra 100mhz actually making my RAM operate more quickly?

 

[Lupiron]

Whats your VDrop and droop? That would help, what do you get in windows at 1.xxx volts? As in, set 1.xxx in the Bios, remember what it is, load windows and look with CPUz, and see what the now lower Core Voltage is listed at, save this as well, and run Prime 95 small ffts on all the cores, after 30 seconds of watching CPUz Core voltage, remember what the LOWEST value it flickers to, and then list those here.

You'll need that 1.45 Volts while loaded!

--Lupi

 

[Lupiron]

Well, I can tell you, you may fail prime simply because of the ram, so I'd place the ram on stock, via 400 x 9 FSB.

What happens when you run Primes Blend test? if it just drops the threads one after the other as well, you know its the ram!

--Lupi