Sign in with
Sign up | Sign in
Your question

Q6600 VID 1.3250 currently @3.2 Ghz (400*8)

Last response: in Overclocking
Share
July 27, 2010 8:31:05 PM

Hi guys,

I have searching and reading for the past 2 weeks about overclocking options. I have an Asus P5QLD Pro with a Q6600 G0 processor and Coolermaster Hyper TX3 cooler in an Antec 1200 case.

This is my CPU-Z validation: http://valid.canardpc.com/show_oc.php?id=1317338

I have some questions I couldn't find an answer to:

My system is currently stable at 1.3125 Vcore which Cpu-z shows as 1.294 as idle and 1.304 at load during Prime95 blend test. I have tested it for a couple of hours and it remains stable. Vista shows my speed as 3.6 Ghz because it assumes that I am using the x9 multiplier while the speed actually is 400*8. The weird thing is, if I don't change the Vcore and just change the multiplier to 9 and fsb to 356 which makes 3.2 Ghz again, the Vcore is absolutely not sufficient anymore! How is that possible, it's the same speed, just different multiplier. I have read that a Vid of 1.3250 is not very well at overclocking.

Next question would be: how far can I safely overclock with this Vid and processor?

I thought that Vid meant the minimal voltage need to run the processor, how come I only need about 1.3 to overclock to 3.2 while the Vid is 1.3250 which should be the minimum needed for 2.4 Ghz?

Thank you very much in advance for your help!

More about : q6600 vid 3250 ghz 400

July 28, 2010 9:52:49 AM

That's not 356 fsb, that's 356 base frequency you meant.

356 base frequency x9 = 3.2 Ghz clock speed. But,
356 base frequency x4 = 1474 Mhz FSB. <-- this one requires more Volt to the motherboard.

Compare this to initial stock Q6600: 2.40 Ghz, 1066 FSB.
266 base frequency x9 = 2398.5 Ghz (as near as written to the 2.40 Ghz label, this is your actual clock speed.)
266 base frequency x4 = 1066 Mhz Fsb

So the default power is for this setting. Of course, to achieve higher Fsb, we need more power, and that 1474 > 1066, so we need more power than those used in 1066, in this case, 1066 uses 1.3 V. So if you run 1474, you need > 1.3 V.

So basically, the more base frequency, more power you need to power Fsb flow to the motherboard, if the processor clock is the same.
July 28, 2010 10:05:16 AM

Please recognize also that an increase in voltage to your motherboard will decrease its lifespan.
Related resources
Can't find your answer ? Ask !
a c 172 à CPUs
a c 197 K Overclocking
July 28, 2010 5:37:25 PM

andrern2000 said:
That's not 356 fsb, that's 356 base frequency you meant.

356 base frequency x9 = 3.2 Ghz clock speed. But,
356 base frequency x4 = 1474 Mhz FSB. <-- this one requires more Volt to the motherboard.

Not exactly accurate. The problem is that different manufacturers define these terms slightly differently. Both of your statements above are different ways of describing the same thing.

Definition time (attention purists, I'm talking about DDR2 and I'm simplifying a little :) ):
Core2 CPU's use a frontside bus (FSB). The FSB is a thing with two main characteristics: speed which is usually defined in MHz and width which in the Core2's is 64 bits wide. We are concerned with the speed.

Using the Q6600 as an example, the FSB frequency is 266 MHz. The matching DDR2 memory clock for that frequency is 533 MHz (266 X 2). DDR2 memory transfers two chunks of data for each bus cycle, hence double the frequency. So, to run 1:1 at an FSB of 266 MHz, we need DDR2-533 RAM.

The FSB clock is 1066 MHz (266 X 4). The bus is "quad pumped". It transfers 4 chunks of data into and out of the CPU each cycle. So each FSB cycle generates 4 FSB clocks.

Now, if you increase the FSB frequency to 333 MHz, the corresponding memory clock is 667 MHz and the FSB clock is 1333 MHz.

Next - VID

The best plain language description of VID that I have seen is from Computronix, one of our old, former regulars:
http://www.tomshardware.co.uk/forum/248414-11-vcore-bio...

The thing to keep in mind about VID is that a CPU with a lower VID tends to be a better overclocker because you have more room to push the CPU voltage before you bump into Intel's recommended limits. It doesn't necessarily mean that the chip won't operate at a lower voltage.

Case in point, my G0 Q6600 has a VID of 1.2625 volts - about average. I can operate at stock frequency with stability (24 hour Prime95 test) at 1.200 volts. And I can operate at 3.0 GHz with the stock cooler at 1.300 volts, 3.3 GHz with a mediocre AC7FP at 1.32 volts, and 3.6 GHz with a TRUE at 1.42 volts (load temps 61 - 66 C).

As far as reduced life is concerned, as long as you stay within the recommended voltage and temperature limits, short of a random failure, your PC is going to be obsolete long before it fails from overvolting.
----------
Overclocking since 1978 - Z80 (TRS-80) from 1.77 MHz to 2.01 MHz :) 
a b à CPUs
a b K Overclocking
July 28, 2010 5:47:36 PM

Try raising your FSB voltage by 0.05 -> 0.1 V

When you raise your clock (bus freq) it increases the loading which drops the amplitude.
When I OCed my E6400 -> 32, I was able to use stock Vcore, but I had to raise my FSB V by 0.1 and raise My (G)MCH voltage by 0.1V

Jsc - Excellent explaination!!
July 29, 2010 2:17:38 AM

Well, that's what I meant. :) 

Hey, I might asking a question too in this thread.

I have Q6600 2.4 Ghz which runs on 1066 Fsb (266x9 multiplier). So, I want to adjust with RAM PC-6400 so it has 1:1 ratio.
I will set it into base frequency 400 Mhz and down its multiplier to 7x. So it will be 2,8 Ghz 1600 Fsb.

Is this categorized as heavy overclocking? What about 400x8 which is more 'common' spoken by people? Is that heavy overclocking? The most important thing is to be not increasing voltage that I DON'T want to increase voltage for my processor, as it can reduce processor lifespan. Maybe for motherboard it's slightly ok. But I do want to achieve 1:1 ratio.
July 29, 2010 9:11:40 AM

Thank you all for the input. Great help from a great community. I have been playing some more and think I found something interesting. I thought I needed more Vcore for the 9x multiplier so after some searching and playing I enabled the Load Line Calibration in the Bios (supposed to be a software version of a pencil mod which reduces vDroop??). Anyway, I got the exact same Vcore I needed with the 8x multiplier now. So this is the new CPU-z validation:

http://valid.canardpc.com/show_oc.php?id=1319667

While I was at it, I went and bought myself a Noctua NH-D14. The thing is just unbelievable. I got pretty good ears and I can't hear a thing, case fans from my antec 1200 are MUCH louder. I have the two fans running at push pull configuration at 100% all the time and while overclocked at 3.2 Ghz with 1.3125 in Bios and 1.304 at load in CPU-Z the hottest core gets to a maximum of 48 C. Much better than my old crappy Coolermaster Hyper TX3.

The current overclock (3.2 Ghz) is stable. Haven't tried raising it much.

My PCIE is now at 100, Ram at 1.8 (recommendation company Hyundai Electronics) and vCore 1.3125 with Load Line Calibration ENABLED. So if I go any higher with my Ghz, do I need to raise the PCIE? or adjust any of the other stuff in the bios? I can't up my ram more than 1.8 according to the manufacturer.

So what to do if I want 3.6 Ghz? :bounce: 
July 29, 2010 9:15:57 AM

andrern2000 said:
Well, that's what I meant. :) 

Hey, I might asking a question too in this thread.

I have Q6600 2.4 Ghz which runs on 1066 Fsb (266x9 multiplier). So, I want to adjust with RAM PC-6400 so it has 1:1 ratio.
I will set it into base frequency 400 Mhz and down its multiplier to 7x. So it will be 2,8 Ghz 1600 Fsb.

Is this categorized as heavy overclocking? What about 400x8 which is more 'common' spoken by people? Is that heavy overclocking? The most important thing is to be not increasing voltage that I DON'T want to increase voltage for my processor, as it can reduce processor lifespan. Maybe for motherboard it's slightly ok. But I do want to achieve 1:1 ratio.


I think the lifespan depends alot on the vCore. If you don't go too crazy and higher than the maximum of 1.5 volt, you should be fine according to many. I think what I got currently, (356*9) 3.2 Ghz overclock isn't very heavy, specially not with a vCore of 1.3125 with LLC enabled. But there are many others on this forum who are much more experienced than me of course, and perhaps they can shed some light on the matter?
a c 172 à CPUs
a c 197 K Overclocking
July 29, 2010 10:06:48 AM

Light, medium, heavy overclock ...
I don't think it is so much a matter of frequency as the amount of work you need to do to get there. For a Q6600 (probably the easiest chip I have ever oc'd), I'd say up to 3.0 GHz is a light overclock. I define "medium" as up to 3.3 - 3.4 GHz. "Heavy" is anything above that where you need serious cooling in a good, free flowing case, and lots of voltage tweaking and testing.

Keep in mind what I am saying. I'm defining an easy overclock as a 25% :o  increase in CPU speed. Admittedly, if I had had a chip with a higher VID where I had to work much harder, my perspective would be different.

Lifetime ...
As I said above.

RAM ...
RAM timing and frequency with the Core2's just don't have much effect on overall performance. Theoretically, maximizing bus speed for a given core speed yields increasing performance. Unfortunately, performance does not scale linearly with FSB frequency. For a given core speed, say 3.0 GHz, (375 KHz X 8) yields an insignificant increase in performance compared to (333 X 9). Likewise, overclocking RAM doesn't do much.

1:1 ratio ...
The first thing I do to a Core2 system is take the memory settings off Auto and set the RAM clock to twice the FSB frequency. If you don't, you can break the overclock. Let's say that you are using DDR2-800 RAM with the Q6600. You turn the system on and the BIOS reads the CPU pins and sets the FSB freq to 266 MHz. And the BIOS reads the SPD chip on the RAM modules and sets the RAM clock to 800 MHz. So you OC the CPU to a not unreasonable 3.33 GHz (367 X 9). The memory clock will be increased the same proportion to (333/266)*800 or DDR2-1100 speed. And your overclock fails.

Load line calibration ...
I use whatever works better. I have 2 P45 Gigabyte boards and a P35 Gigabyte board. On two of them, LLC gives a more stable overclock. On the second P45 board, it works better with LLC off.
a b à CPUs
a b K Overclocking
July 29, 2010 4:18:40 PM

i have mine @ 1.36 vcore and 400 X 9, had to bump the RAM voltage a little bit from 1.9 to 2.0 and it is 100% stable since September @ 3.6 GHz
July 29, 2010 5:12:08 PM

jonpaul37 said:
i have mine @ 1.36 vcore and 400 X 9, had to bump the RAM voltage a little bit from 1.9 to 2.0 and it is 100% stable since September @ 3.6 GHz


Wow, if only I could reach that speed with that vCore. What is you VID? Also 1.3250? Perhaps you have a lower vid that let's you overclock better? My Ram has a recommended voltage of 1.8, what about yours? Can I safely up the voltage to 2.0 without smoke coming out of my pc case?

a b à CPUs
a b K Overclocking
July 29, 2010 5:43:22 PM

Only thing i have had to change was the following, granted in the beginning it took some time to establish this sweet spot but it is stable:

FSB = 400
Vcore = 1.365
RAM voltage to 2.0
RAM timings to 4-4-4-12
Set memory multiplier to 1:1

disabled any speedstep or C1E

there may have been some other minor adjustments i made but nothing big or that involved votlages.
July 29, 2010 6:17:18 PM

jonpaul37 said:
Only thing i have had to change was the following, granted in the beginning it took some time to establish this sweet spot but it is stable:

FSB = 400
Vcore = 1.365
RAM voltage to 2.0
RAM timings to 4-4-4-12
Set memory multiplier to 1:1

disabled any speedstep or C1E

there may have been some other minor adjustments i made but nothing big or that involved votlages.


Thanks for the info. Could you also tell me your CPU VID? You might have a lower/better one which enables you to overclock that high with such a Vcore. My CPU VID is 1.3250, I could read it using CoreTemp which I also use to monitor my core temperatures.

Any idea if I can safely up the voltage of my ram to 2.0 even though it's recommended and rated value is at 1.8?

Thank you for all the help!
a b à CPUs
a b K Overclocking
July 29, 2010 6:39:15 PM

yes, never go past 2.1 for RAM but since you are rated at 1.8 i would not go past 2.0

just remoted into my rig at home and downloaded core temp, VID says 1.2750v
July 29, 2010 7:48:57 PM

jonpaul37 said:
yes, never go past 2.1 for RAM but since you are rated at 1.8 i would not go past 2.0

just remoted into my rig at home and downloaded core temp, VID says 1.2750v


Thank you. I will give it a shot. Your VID is better than mine, I might need a higher Vcore.
August 2, 2010 9:16:50 AM

jsc said:

RAM ...
RAM timing and frequency with the Core2's just don't have much effect on overall performance. Theoretically, maximizing bus speed for a given core speed yields increasing performance. Unfortunately, performance does not scale linearly with FSB frequency. For a given core speed, say 3.0 GHz, (375 KHz X 8) yields an insignificant increase in performance compared to (333 X 9). Likewise, overclocking RAM doesn't do much.

1:1 ratio ...
The first thing I do to a Core2 system is take the memory settings off Auto and set the RAM clock to twice the FSB frequency. If you don't, you can break the overclock. Let's say that you are using DDR2-800 RAM with the Q6600. You turn the system on and the BIOS reads the CPU pins and sets the FSB freq to 266 MHz. And the BIOS reads the SPD chip on the RAM modules and sets the RAM clock to 800 MHz. So you OC the CPU to a not unreasonable 3.33 GHz (367 X 9). The memory clock will be increased the same proportion to (333/266)*800 or DDR2-1100 speed. And your overclock fails.

Load line calibration ...
I use whatever works better. I have 2 P45 Gigabyte boards and a P35 Gigabyte board. On two of them, LLC gives a more stable overclock. On the second P45 board, it works better with LLC off.


1) So 400*7 1600 Fsb won't gain much from the "1600 Fsb" part. I don't know if the gain mostly come from multiplier, though.. Nice info.

2) 1:1 Fsb ratio Hence since the bus speed there don't have any bottleneck, it ought to increasing performance from stock speeds right? (0,66:1 ratio at stock 266 Mhz)
And thanks so much for the info for setting the RAM to not-auto. I would leave it Auto if I didn't read it, possibly crashing caused by the RAM to go past 800.

But would it yield noticeable gain because of 1:1 ratio now?
Thanks alot for the in-depth explanation.
a c 172 à CPUs
a c 197 K Overclocking
August 2, 2010 4:11:12 PM

You do not get a noticeable gain in performance. What you do get is a noticeable gain in stability compared to the same system with RAM running faster that the FSB.
August 5, 2010 8:09:20 PM

So, I have been playing some more with my overclock. I have it now at 3.4 Ghz. I had 3.5 first but I decreased the vCore just to be on the safe side, specially with my horrible VID of 1.3250. I thought I had it very stable but I get an error after 2,5 hours from Prime95. Good idea to keep increasing the vCore. My RAM cannot get higher than 1.8 volt. I did Memtest86+ on my RAM just to be sure, it passed perfectly. Can this error be due to my RAM or should I just keep increasing vCore? Advice is very appreciated, thank you!



Uploaded with ImageShack.us
August 6, 2010 5:17:46 AM

jsc said:
You do not get a noticeable gain in performance. What you do get is a noticeable gain in stability compared to the same system with RAM running faster that the FSB.


No performance gain at all? Hmm, okay, I can keep up with that.
Thanks for the info.

@bactrian: Maybe it's caused by bottleneck in RAM:FSB? Maybe others can help regarding on this matter.
August 6, 2010 4:51:53 PM

andrern2000 said:
No performance gain at all? Hmm, okay, I can keep up with that.
Thanks for the info.

@bactrian: Maybe it's caused by bottleneck in RAM:FSB? Maybe others can help regarding on this matter.


Yes you might be right, I have 2x1 GB RAM from Hynix. Eventhough it passed memtest86+ I still think it's not very high quality Ram and might be the cause of my problem.
August 10, 2010 4:31:48 AM

bactrian said:
Yes you might be right, I have 2x1 GB RAM from Hynix. Eventhough it passed memtest86+ I still think it's not very high quality Ram and might be the cause of my problem.


Wait, you passed the Memtest86+ @ 1.8V, which means that the RAM is quite good.
Hmm,.. I think before judging the RAM's quality, maybe increase the voltage just one notch, to be sure that it's caused by need more voltage, or not.
!