Seeking a second opinion overclocking some dated hardware

[Blitz KriegeR]

Ok so I've had this rig (see sig) just over a year now. When I put it together originally, I intended to OC, but the stock performance was so great I didn't bother. Times change, and I'm now feeling the need to squeeze a bit more out of her.

Right now I'm on a stable OC of 300x9 (1.25v) with memory DDR2-900 @ 5-5-5-15. The memory is somewhat unrefined, as I'm still hoping to get the CPU higher.
CPU load temps at this setting are 50-50-48-46.

Originally when I got the rig I was thinking 400x8 running memory 1:1 @ 4-4-4-12, Or simply getting it to a penryn level of 333x9 would make me happy.

Anyway, more to the point, here are my current settings for 300x9; (ASUS x48 RAMPAGE 0802 BIOS)
CPU VCORE: 1.2500v
CPU PLL Voltage: 1.50v
FSB TERM: 1.200v
MHC: 1.25v
IHC: 1.05v
IHC 1.5 voltage: 1.50v
Load line Calibration: Disabled
PCIe Frequency: 100
FSB Strap to MHC: 266

Now, when I try to raise my FSB, I can't even get into windows! I just get constant bluescreens (everything from UNRESOLVABLE HARDWARE ERROR to simply STOP error, to IRQ_NOT_LESS_OR_EQUAL).

I've incrementally increased all voltages to a point where I can at least get into windows, but the system is not stable, failing any stress test within 5min.

Trying to get the OC up from 300 to 333 I've made the following changes:
CPU VCORE: 1.3625v
CPU PLL Voltage: 1.6v
FSB TERM: 1.36v
MHC: 1.35v
ICH: 1.075v
IHC 1.5 voltage: 1.52v
DRAM: 2.05v
*Memory ratio is set 1:1 when attempting to push OC up
*FSB Strap to MHC has been tested at 266 and 333

Now, my main question is: are these kind of voltage bumps normal? It seems to me that I've had to pump an excessive amount of voltage into my rig just to get it to boot, and it's not even stable. I've seen people get 3Ghz+ out of a G0 on less than 1.28v. I know all chips are different but damn...

Just fishing for some second thoughts here. Thanks in advance for any comments/input.

 

[Idontcare]

Generally when troubleshooting OC's you need to know:

1. memory overclock profile (DDR2-XXX vs. Vdimm needed to hit that speed and be stable)

2. FSB overclock profile (FSB MHz vs. NB/MCH/FSB voltage needed to hit that speed and be stable)

3. CPU overclock profile (CPU GHz vs. Vcore needed to hit that speed and be stable)

As it stands now it doesn't look like you are sure which of the three above are causing you to be held back from getting #3 up to where you want it.

The procedure I recommend has been posted before so I'll just quote it here for your review, it may seem lengthy but the idea is to systematically iterate yourself into a region where you know your rig's specific limitations so you know what it can do and what it can't...then hopefully no more frustration and wondering on your part.

Originally posted by: Idontcare

Originally posted by: Dadofamunky

Originally posted by: Idontcare

What is your Vdimm? And what is your ram clocked to? Large FFT errors can be MCH (memory controller hub) located in your NB, can be FSB (located in your NB) or can be ram (located in the dimms).

VDIMM is set to 2.1V, which is the recommended for this particular RAM at DDR2-1000 speeds. At DDR2-800, G.Skill states the sticks can run at 1.8V. RAM is now DDR2-920.

OK, thanks for the added info.

Gillbot's suggestion is worth considering. Maybe even for just the duration of doing some tests you could "ghetto" style a case fan with some zip-ties or wire-ties to direct some moving air towards your NB heatsink and ram.

So here's my recommendation: we need to isolate your error's as being due to FSB (NB), ram, or CPU.

We do this piecemeal-wise, and we do it in this order: Ram first, NB/FSB second, CPU last.

Here's what you do, first you find your ram's max clockspeed and Vdimm that it seems to be able to operate at. We do this while keeping both the CPU and the FSB/NB at speeds we know aren't going to cause errors. I.e. put that CPU and FSB back to stock in all aspect.

Then start changing your ram multiplier (1:1, 5:6, etc) only while using memtest86+ and large FFT prime 95 to find the max stable overclock for the heat and Vdimm you are willing to push them with.

Hopefully you find them to be stable at 2.1V and DDR2-1000, meaning you have margin when you are clocking them at 920MHz. At this point you know when your ram will be holding you back so you don't waste time going there with FSB overclocks.

Second we find the FSB/NB max clocks and voltage. Now we don't want the CPU getting overclocked and throwing instability into this testing, so leave everything about the CPU on AUTO except the multiplier, intentionally set it to the lowest value you can in the BIOS (should be 6x) and set your ram multiplier to 1:1 (also called sync for some BIOS's) to ensure the ram is not going to be unstable, also set your Vdimm to your stable value for its highest clocks you found in the stage one testing above.

Now overclock your FSB clockspeed, increasing it in sizable chunks (50MHz or so) every iteration and then test system stability with Large FFT. Don't waste your time with small FFT or memtest here. Once you find the FSB clockspeed and voltage you are comfortable operating while knowing the rig is stable then you are done with the second stage.

Hopefully you find your FSB is stable up to 480-500MHz given that you want to operate at 460MHz. If FSB isn't stable at 460 then you've isolated your problems way above.

Third we find the CPU max clocks and voltage. Again we don't want the ram or the NB to be the cause of any instabilities we uncover, so set ram multiplier to 1:1 and Vdimm to the value you needed to reach peak ram overclocks as well as set all your NB/FSB settings (except actual FSB clockspeed) to the values you determined you needed in stage 2 above to be stable while hitting those peak FSB clocks.

Now set your CPU multiplier manually to its max value and start testing at elevated FSB clocks and checking stability with small FFT only. Presumably memtest and large FFT are passing at these lower FSB clockspeed based on them passing in stages one and two above.

As you increase FSB you will naturally need to manually increase CPU Vcc. Make sure you got the Voltage Damper enabled (not AUTO, not disabled). Step thru the same FSB increase increments you did in stage 2 above.

At some point in these three stages you are going to find something in your rig goes unstable below its current overclocked clockspeeds. It is a laborious process, but the orthogonal testing nature allows you to isolate the weakest link component that is holding back your OC.

When you do the testing there is no need to put in extensive lengthy prime95 runs, 15-20min of stable is good enough for quickly filling out the matrix of good vs. bad speed/voltage combinations. Then when you are getting close to an ideal combination you can test those few specific settings for lengthier times to figure out how much of a bump-up in voltages you'll need.

 

[Blitz KriegeR]

Well that's the opposite of what I've read in some guides, but it sounds like a solid method. I'll give it a try.

Although, shouldn't I effectively be bypassing your "stage 1" by having the memory at 1:1 @ 2.05v? That should give me 100% stability up to 400Mhz FSB (Ram is rated 800Mhz 4-4-4-12 2.0v). And I've got the timings all loose (5-5-5-15) to release pressure on the MHC. Thus, I'd anticipate having to add voltage to CPU and slightly more to MHC to compensate for increased input. That's what I got from some OC guides, is this logic flawed?

Also, I understand how to use Vcore, FSB Term, MHC, IHC, and DRAM voltage, but what the hell is CPU PLL voltage?

Anyway, off to try some testing with Idontcare's methodology.

Thanks for the input!

 

[Idontcare]

Originally posted by: Blitz KriegeR
Well that's the opposite of what I've read in some guides...

I don't claim it to be the best or most efficient method, if other guides get you to the answer you need with less work then definitely go with them. I found this method I document to have worked pretty well for me, but that doesn't mean there isn't better ways to do it.

Originally posted by: Blitz KriegeR
Although, shouldn't I effectively be bypassing your "stage 1" by having the memory at 1:1 @ 2.05v? That should give me 100% stability up to 400Mhz FSB (Ram is rated 800Mhz 4-4-4-12 2.0v). And I've got the timings all loose (5-5-5-15) to release pressure on the MHC.

Yeah you can definitely "skip" stages if you like, the idea isn't to do some exhaustive time-consuming matrix analysis. The idea is to eliminate probable suspects that can limit your OC.

I would not simply assume that your ram is stable 1:1 @ 2.05V with 400MHz FSB...I would test it with memtest and large FFT prime95 to confirm it is stable while underclocking my CPU (lower the multi) to reduce the liklihood that my CPU is causing errors during the memory testing.

Also the idea isn't to sample the clockspeed space every 5MHz or something silly like that. Make substantial stepsizes in the regions where you are pretty much assured stability.

 

[error8]

The extreme retarded method I'm using: (I'm not responsible for any hardware failure )

Set the FSB to 400
multiplier 9X
Vcore 1.46V
MCH 1.45V
SB whatever
VTT 1.4 V
don't oc the ram

Try to pass 10 "standard tests" in Linpack. If it fails, then up the vcore a bit, or if the temps you're having go beyond 75 C, back the oc and voltage a bit .

 

[Blitz KriegeR]

Just wanted to chime in and say thanks for the input. I've adapted a hybrid method between what I use and what you suggest and found something that seems to be working for me.

I started by bottoming out all voltages except CPU Vcore, which I kept at 1.25v. Then anticipated some needs; boosted NB to 1.35v to keep it out of the equation (at a safe voltage), dropped RAM to 1:1 divider @ 5-5-5-15 tRD 12 ("performance level" also known as "MCH Read Delay") running 2.02v, increased FSB Term by one step to 1.22v, dropped my CPU multiplier to minimum (6), and started to raise FSB frequency.

I got it stable at 2.4Ghz (400x6) {Stock frequency, but 50% bus OC} without changing any other settings from what I mentioned above.

I then increased the multiplier to 7 (2.8Ghz) and started testing for stability. Had to bump CPU Vcore to 1.3v (1.2875v might work, didn't test for long). Stable at 2.8Ghz (400x7) 1.30v.

Finally, increased the multiplier to 8 (3.2Ghz) and started to test for stability. I got bluescreens on boot all the way to 1.425v, which I am now testing for stability. So far I haven't had to increase any other settings except CPU Vcore, and have been testing for stability with a quick pass of small FTT and large FTT in-between each run.

So yeah, so far so good. All I need to do now is stabilize this frequency, then start tuning in the memory. Going to keep it either 800 @ 4-4-4-12 or 1066 @ 5-5-5-15, and bring the tRD down as far as I can, hoping for 6.

Thanks again for the well advised input!

EDIT: My Board seems to be under-volting my CPU by .05v even at idle. 1.425v in BIOS only yields 1.375 as read by even the boards own BIOS hardware monitor. Vdroop is even worse. It is also overvolting *everything* else by .025v to .085v. And this is supposedly a rock solid board? :S

 

[imported_Shaq]

Originally posted by: Blitz KriegeR
Just wanted to chime in and say thanks for the input. I've adapted a hybrid method between what I use and what you suggest and found something that seems to be working for me.

I started by bottoming out all voltages except CPU Vcore, which I kept at 1.25v. Then anticipated some needs; boosted NB to 1.35v to keep it out of the equation (at a safe voltage), dropped RAM to 1:1 divider @ 5-5-5-15 tRD 12 ("performance level" also known as "MCH Read Delay") running 2.02v, increased FSB Term by one step to 1.22v, dropped my CPU multiplier to minimum (6), and started to raise FSB frequency.

I got it stable at 2.4Ghz (400x6) {Stock frequency, but 50% bus OC} without changing any other settings from what I mentioned above.

I then increased the multiplier to 7 (2.8Ghz) and started testing for stability. Had to bump CPU Vcore to 1.3v (1.2875v might work, didn't test for long). Stable at 2.8Ghz (400x7) 1.30v.

Finally, increased the multiplier to 8 (3.2Ghz) and started to test for stability. I got bluescreens on boot all the way to 1.425v, which I am now testing for stability. So far I haven't had to increase any other settings except CPU Vcore, and have been testing for stability with a quick pass of small FTT and large FTT in-between each run.

So yeah, so far so good. All I need to do now is stabilize this frequency, then start tuning in the memory. Going to keep it either 800 @ 4-4-4-12 or 1066 @ 5-5-5-15, and bring the tRD down as far as I can, hoping for 6.

Thanks again for the well advised input!

EDIT: My Board seems to be under-volting my CPU by .05v even at idle. 1.425v in BIOS only yields 1.375 as read by even the boards own BIOS hardware monitor. Vdroop is even worse. It is also overvolting *everything* else by .025v to .085v. And this is supposedly a rock solid board? :S

I was going to add to this but it looks like you got it figured out. The CPU voltage was way too low. My Q6600 G0 needed 1.37 volts to be stable at 3.0. I had to set 1.42v in the BIOS like you described (voltage drop). 3.3 was the max for mine and it needed a true 1.54v to be stable, which was 1.6v in BIOS. lol

 

[Jabbernyx]

C'mon! Dated?! You're trying to make my Rev 1.0 965P-DS3 + E6600 @ 3.6GHz feel bad aren't you
*EDIT* WTF timewarp!

 

[Blitz KriegeR]

Just finished my last round of testing, and I actually needed 1.4875v to keep it stable at 3.2Ghz (400x8). 1.445v idle, 1.40v load; a lot more voltage than I expected!

How the hell do people reach 3.5-3.8 on air with this thing? lol! It is running at 68'C under a ''T.R.U.E.'' with a 110CFM fan in a push configuration, plus a pair of 133CFM fans on either end (one blowing onto the RAM, at a right angle to the CPU, and the other as the primary rear exhaust.)

So in the end I made it to my 15month old goal of matching the P4 "C" Northwood this baby replaced in pure frequency. God it needs almost as much voltage though (my P4 ran at 1.52v idle); albeit it results in much greater performance!

So, now I'm thrilled, and I get to spend the rest of my weekend off tuning DRAM frequency and tRD! ^.^

EDIT: Thanks anyway Shaq! God, eyeing that i7 of yours at 4Ghz makes me jealous, and prone to unnecessary impulsive expenditure! X.X

 

[Assimilator1]

Lol well I guess my Q6600 is a good chip then , I'm currently running mine at 3.34 GHz, bios vcore 1.40625v with a loaded vcore of 1.344v running DPAD (all cores nearly 100%).

Mbrd is an Asus P5Q Pro (P45), see my thread here for more info on cpu & mbrd voltages.
Although it's now running 417 MHz FSB, 1046 MHz RAM, NBv 1.18v, vDIMM 2.14v, CPU PLL 1.52v, & vFSB Term. 1.22v.

It's been a while since I set it up but IIRC temps were my limiting factor, I've since bought but not yet fitted a better cooler which should hopefully let me o/c higher .

I'd suggest for your rig you could drop the CPU PLLv to 1 above the base manual setting, that'll cut down heat output, retest for stabilty. You might then be able to drop vcore or maybe increase clock speed.

Btw does your mbrd show CPU PLL, vFSB TERM & NBv in the H/W monitor?

 

[Blitz KriegeR]

Or my Q6600 just *isn't* a very good sample. I see a lot of folks in the 3.0-3.6Ghz range that need as little as 1.38v - 1.45v for stability. I passed several hours of OCCT small & large with these settings:

3.2Ghz (400x8), DRAM 1:1 @ DDR2-800 4-4-4-15 tRD 10
CPU Vcore: 1.4875v
CPU PLL: 1.5v
FSB Term: 1.2v
MCH: 1.45v (probably could get a lot lower, just put it there from anand's original review of my board, going to try to push the tRD down to 6 or 7. I'm in the middle of that today)
ICH: 1.05v
ICH 1.5v: 1.5v
DRAM: 2.02v (board over-volts by 20-80 mV so its more like 2.05v - 2.1v)

As soon as I tried some Linpack early this morning it fails in under 10min, had to put the Vcore up to 1.50v in BIOS. Now reading 1.456V Idle and 1.39-1.40v load.

My limiting factors are now both temp & voltage. I'm not willing to anywhere above 1.5, and I'm running up to 68'C after hours of OCCT load, and up to 76'C after an hour of Linpack. Using a "T.R.U.E" with a 110CFM fan, with two 133CFM fans on either side. (one as primary exhaust and the other blowing onto the RAM, perpendicular to the CPU/HS)

Yes my motherboard shows all seven of the above mentioned readings in its hardware monitor, as well as the output on my PSU rails. I don't trust anything except a multimeter for PSU readings though, heh.

I'm a bit confused about what you said about the PPL voltage? How can I "drop to 1 above the base manual setting"? It's 1.5v by default, up to 3.0v in 20 mV increments. Are you suggesting I increase it one step to 1.52v? Not sure how that will help cut heat output; won't that just do the opposite?

Anyway, off to drop my tRD down to 8 and rerun all my tests. It's doing wonders for memory read speeds and access latency.

 

[error8]

10 standard tests of Linpack are more then enough to test your stability. On guru3d it's written "should use at least 5 at minimum, no more than 20". Linpack it's not one of those Prime95 that can run for days, it's a very dangerous program that doesn't need to run that long. I can say that 10 standard test are enough for testing your stability. If you pass those, your overclock is ready for every day use. And 76 C is a good temperature for that voltage, under Linpack.

Just make a 10 tests pass and see if it's stable. I have the feeling that you can stabilize it at 1.5 V at more then 3.2 ghz. Maybe at 3.4-3.5 ghz.

Linpack guide

 

[VirtualLarry]

Wow, those later date Q6600s just don't clock like they used to. Mine is an older one, and I got to 3.6Ghz at 1.325v (IIRC). LLC enabled, on a DFI X48 board. Problem is temps, not voltage, for me. Temps get to like 85C under OCCT: Linpack.

 

[error8]

Originally posted by: VirtualLarry
Temps get to like 85C under OCCT: Linpack.

Tjunction set at 90C?

 

[Blitz KriegeR]

God, so many different programs floating around. Every time I think I've found a set I'm happy with someone shoots me another one.

Just tried that link of yours error8, failed after 3m23s even at 1.5v.
I've been OCCT stable for over 8hrs. I'm testing both small FFT, and then large using 1GB RAM. Then running a full pass of memtest using all available RAM (6-8GB). No errors in that.

In terms of memory testing, my tRD is down to 8 so far; shaved 17.6ns off my original memory access latency, and increased read speed by over 850MB/s (from same settings at tRD12 which I used for guaranteed stability during CPU testing). Going to try putting it at 7 or even 6. Still using 1.45v MCH, 1.5v Vcore, 2.0-2.1v DRAM.

Unrelated: I don't think my Q6600 is *that* new. I bought it in early Mar08. I can tell you it's a SLACR G0 with a packing date of 12/19/07. My TJ max setting is showing 100'C.

 

[error8]

Originally posted by: Blitz KriegeR
My TJ max setting is showing 100'C.

Your TJ max should be 90C. Intel has corrected that value in November last year I believe

 

[Blitz KriegeR]

I had a read through that specsheet. Most of it is over my head, but it did say 65nm Q6000 series TJMAX was 90'C on G0. Then again I'm not sure what to say coz Coretemp and Realtemp both report my TJmax 100'C.

 

[error8]

Originally posted by: Blitz KriegeR
I had a read through that specsheet. Most of it is over my head, but it did say 65nm Q6000 series TJMAX was 90'C on G0. Then again I'm not sure what to say coz Coretemp and Realtemp both report my TJmax 100'C.

You have to set 90 C as Tj max, manually in CoreTemp or RealTemp and there you'll have the real temperature. Those programs are using the older Tj max for your chip.

 

[imported_Shaq]

Originally posted by: VirtualLarry
Wow, those later date Q6600s just don't clock like they used to. Mine is an older one, and I got to 3.6Ghz at 1.325v (IIRC). LLC enabled, on a DFI X48 board. Problem is temps, not voltage, for me. Temps get to like 85C under OCCT: Linpack.

It depends on your VID. Mine was 1.325. It is the highest or second highest value. You should be able to go .2 volts over your VID when you OC. You probably had a 1.2 VID chip, or less, which is quite good. Your chip would burn up before you got to 1.5 volts. lol

 

[error8]

Originally posted by: Shaq

Originally posted by: VirtualLarry
Wow, those later date Q6600s just don't clock like they used to. Mine is an older one, and I got to 3.6Ghz at 1.325v (IIRC). LLC enabled, on a DFI X48 board. Problem is temps, not voltage, for me. Temps get to like 85C under OCCT: Linpack.

It depends on your VID. Mine was 1.325. It is the highest or second highest value. You should be able to go .2 volts over your VID when you OC. You probably had a 1.2 VID chip, or less, which is quite good. Your chip would burn up before you got to 1.5 volts. lol

Who says you can only go 0.2 V over your VID? You can go as far as 1.55V, as Intel states for a 65 nm chip, as long as you can keep it cool enough.

 

[BonzaiDuck]

Blitz KriegeR said:

Or my Q6600 just *isn't* a very good sample. I see a lot of folks in the 3.0-3.6Ghz range that need as little as 1.38v - 1.45v for stability.

Assimilator1 said:

Lol well I guess my Q6600 is a good chip then , I'm currently running mine at 3.34 GHz, bios vcore 1.40625v with a loaded vcore of 1.344v running DPAD (all cores nearly 100%).

VirtualLarry said:

Wow, those later date Q6600s just don't clock like they used to. Mine is an older one, and I got to 3.6Ghz at 1.325v (IIRC). LLC enabled, on a DFI X48 board. Problem is temps, not voltage, for me. Temps get to like 85C under OCCT: Linpack.

I'm not setting myself up as the great expert or the last word on this. We followed the Kentsfield from B3 to G0 stepping through 3rd quarter 2007. Some people -- on water-cooling -- were able to get the B3 close to 3.6 Ghz with voltages exceeding 1.47V. There had even been a review that I read boasting that you could get it to 3.5 Ghz with a 1.5+V BIOS "setting" value.

But a lot of posters through end of '07 and into 2008 seemed to converge on the idea that reasonable expectations for air-cooling with the G0 stepping tended to average out at around 3.4 Ghz.

Let's face it: the Kentsfield -- even the G0 stepping -- was spec'd at 1066 FSB, and getting it to host-frequency 400 Mhz (and 1600 FSB) is a 50% over-clocking above the stock 2.4 Ghz -- if you plan to use the stock multiplier of 9.

Look at VirtualLarry's temperatures. It doesn't so much make sense to me, except that he's got LLC enabled, how his temperatures could be so high.

My B3 stepping of the Q6600 just won't go past 3.2+ Ghz without pushing the voltage setting above 1.45-something, with the CPU-Z reading idle at around 1.38V.

With my motherboard, I can't just jump over to a Yorkfield -- I'm stuck with Wolfdale dual-cores. I've thought about getting a Q6700 G0 stepping (I think that's all there was for the 2.66 Ghz core). You can grab OEM-wrap CPUs of that model for about $173 right now.

And you have to ask yourself: Between DDR2 and DDR3 dual-channel motherboards and chipsets, and DDR3-tri-channel . . . between Yorkfields and I7 cores . . . the price of new motherboards -- all of it . . . . Is getting a faster Kentsfield G0 rated at 1066 FSB -- REE-ally worth it -- even for that price?

 

[Blitz KriegeR]

Originally posted by: error8

Originally posted by: Blitz KriegeR
I had a read through that specsheet. Most of it is over my head, but it did say 65nm Q6000 series TJMAX was 90'C on G0. Then again I'm not sure what to say coz Coretemp and Realtemp both report my TJmax 100'C.

You have to set 90 C as Tj max, manually in CoreTemp or RealTemp and there you'll have the real temperature. Those programs are using the older Tj max for your chip.

Somehow I just can't believe that. If my temps are really 10'C off, that would mean that when I'm on stock voltage/frequency, my idle temp is 17-22'C (because it's 32-30-28-27). You're going to tell me that my fourth CPU core idles below my room temp with my TRUE? That's a physical impossibility. Air coolers can't cool below the temp of the air they are using, it's against the laws of physics.

Although, to your credit, Everest and Speedfan do show my temps 10'C lower, so it seems I now know why. Still, as much as I'd like to believe it the above makes me see it's impossible.

Originally posted by: Shaq
It depends on your VID. Mine was 1.325. It is the highest or second highest value. You should be able to go .2 volts over your VID when you OC. You probably had a 1.2 VID chip, or less, which is quite good. Your chip would burn up before you got to 1.5 volts. lol

Wait, isn't VID something that is specific to an entire line, or at least batch of CPUs? It's not different from individual CPU to CPU...? Coretemp reads my G0's VID as 1.325v, the Intel label on the box says 1.35v. Either way, sounds fishy.

Originally posted by: Bonzaiduck
My B3 stepping of the Q6600 just won't go past 3.2+ Ghz without pushing the voltage setting above 1.45-something, with the CPU-Z reading idle at around 1.38V.

About the same as me. At 3.2ghz I need 1.425v to post, 1.4875v to pass 4hrs OCCT, and 1.5v to pass Linpack.


My latest OC thought: God trying to stabilize 8GB of RAM sucks.

 

[imported_Shaq]

Originally posted by: Blitz KriegeR

Originally posted by: error8

Originally posted by: Blitz KriegeR
I had a read through that specsheet. Most of it is over my head, but it did say 65nm Q6000 series TJMAX was 90'C on G0. Then again I'm not sure what to say coz Coretemp and Realtemp both report my TJmax 100'C.

You have to set 90 C as Tj max, manually in CoreTemp or RealTemp and there you'll have the real temperature. Those programs are using the older Tj max for your chip.

Somehow I just can't believe that. If my temps are really 10'C off, that would mean that when I'm on stock voltage/frequency, my idle temp is 17-22'C (because it's 32-30-28-27). You're going to tell me that my fourth CPU core idles below my room temp with my TRUE? That's a physical impossibility. Air coolers can't cool below the temp of the air they are using, it's against the laws of physics.

Although, to your credit, Everest and Speedfan do show my temps 10'C lower, so it seems I now know why. Still, as much as I'd like to believe it the above makes me see it's impossible.

Originally posted by: Shaq
It depends on your VID. Mine was 1.325. It is the highest or second highest value. You should be able to go .2 volts over your VID when you OC. You probably had a 1.2 VID chip, or less, which is quite good. Your chip would burn up before you got to 1.5 volts. lol

Wait, isn't VID something that is specific to an entire line, or at least batch of CPUs? It's not different from individual CPU to CPU...? Coretemp reads my G0's VID as 1.325v, the Intel label on the box says 1.35v. Either way, sounds fishy.

Originally posted by: Bonzaiduck
My B3 stepping of the Q6600 just won't go past 3.2+ Ghz without pushing the voltage setting above 1.45-something, with the CPU-Z reading idle at around 1.38V.

About the same as me. At 3.2ghz I need 1.425v to post, 1.4875v to pass 4hrs OCCT, and 1.5v to pass Linpack.


My latest OC thought: God trying to stabilize 8GB of RAM sucks.

VID is the voltage specified by the CPU to run at stock settings. It tells the motherboard BIOS what voltage to set. Some CPU's are more efficient and run at lower voltages. The voltage on the box is the range of VID's. 1.175-1.35 I believe. 1.65v is max spec according to Intel.

Originally posted by: Shaq

Originally posted by: VirtualLarry
Wow, those later date Q6600s just don't clock like they used to. Mine is an older one, and I got to 3.6Ghz at 1.325v (IIRC). LLC enabled, on a DFI X48 board. Problem is temps, not voltage, for me. Temps get to like 85C under OCCT: Linpack.


It depends on your VID. Mine was 1.325. It is the highest or second highest value. You should be able to go .2 volts over your VID when you OC. You probably had a 1.2 VID chip, or less, which is quite good. Your chip would burn up before you got to 1.5 volts. lol

Who says you can only go 0.2 V over your VID? You can go as far as 1.55V, as Intel states for a 65 nm chip, as long as you can keep it cool enough.

I meant that as a rule of thumb. If you have a 1.2 VID CPU you won't be able to go over 1.5v on air without it being too hot. .2v is just a rough estimate for overclocking on air.

 

[firewolfsm]

You should stick to a more modest OC and you'll be able to lower your voltages dramatically.

 

[error8]

Originally posted by: Blitz KriegeR

Originally posted by: error8

Originally posted by: Blitz KriegeR
I had a read through that specsheet. Most of it is over my head, but it did say 65nm Q6000 series TJMAX was 90'C on G0. Then again I'm not sure what to say coz Coretemp and Realtemp both report my TJmax 100'C.

You have to set 90 C as Tj max, manually in CoreTemp or RealTemp and there you'll have the real temperature. Those programs are using the older Tj max for your chip.

Somehow I just can't believe that. If my temps are really 10'C off, that would mean that when I'm on stock voltage/frequency, my idle temp is 17-22'C (because it's 32-30-28-27). You're going to tell me that my fourth CPU core idles below my room temp with my TRUE? That's a physical impossibility. Air coolers can't cool below the temp of the air they are using, it's against the laws of physics.

Although, to your credit, Everest and Speedfan do show my temps 10'C lower, so it seems I now know why. Still, as much as I'd like to believe it the above makes me see it's impossible.

Just set the Tjunction manually and see what happens. Idle won't go bellow ambient for god sake and even if it does, idle is not something that you want to care about. You are only interested in full load temps. My idle temps now are at 33 C with Tjunction set at 90C. You haven't even tried it and your so scared about it. Tjunction set at 90C will give you the real temperature, 10 C lower then what you're having now, with it at 100C. If intel states that the Tjuncion value is 90C, why would you use another wrong Tjunction value????