Core i7/X58 Overclocking Thread

[dejanh]

My new 24/7 I think for i7 940. Going for the best OC/voltage ratio.

Currently 3.8GHz CPU, 1603MHz 7-8-7-20-1T DRAM, with 1.20000V CPU, 1.23125V QPI/DRAM, 1.55081V DRAM (all BIOS settings)...C-states are enabled hence the post-run clocks are lower.

Screenshot below...not bad for a low-voltage OC I think

[tot31]

Got my i7 running at 3.4 Ghz at the moment.

BIOS settings:
vCore - 1.265
QPI/DRAM - AUTO
vDimm - 1.65 (1360 Mhz)
QPI multi - 18x (3060 Mhz)

I'm having some trouble with my QPI voltage. I've tried setting it withing 1.25-1.27v, and I can't get it stable, and Prime crashes almost instantly at those voltages. What am I doing wrong? I can't see why it should be higher

My i7 920 runs at the same speed of yours but with a very low voltage i.e below 1.1V


6hrs prime without errors.


Now it's running at 1.024V idle and 1.03V on load.

You can try to increase your Vcore to make it more stable.

[Classic Satch]

My i7 920 runs at the same speed of yours but with a very low voltage i.e below 1.1V


You can try to increase your Vcore to make it more stable.

Man, I've found it to be far better to run two instances of Intel Burntest for stability purposes. It stresses your chip waaaaaay more and you'll discover an error far faster.

http://www.xtremesystems.org/forums/...d.php?t=197835

[YukonTrooper]

What kinda of cooling are you on, Tot?

[tot31]

What kinda of cooling are you on, Tot?

I'm on air using Deepcool Ice Blade Pro with 2 x 120mm fans (push/pull) setup. My room temperature is normally between 30C - 32C.

[Hot'n'Chilly]

I'm at 3.8ghz using 1.31v CPU. Why are you guys going for such low voltages as 1.1v? Are you using a micro case or something where heat is a big deal?

[tot31]

I'm at 3.8ghz using 1.31v CPU. Why are you guys going for such low voltages as 1.1v? Are you using a micro case or something where heat is a big deal?

I'm living in a very hot climate country where all around the year is always summer. Like I posted before that my room temp is around 30C - 32C, sometimes can go up to 34C, so heat is really a big problem to me...

[Gogar]

Made a few updates to my i7 OC Optimizer tool :

- more multipliers for ram and QPI. (from P6T bios)
- ability to enter the QPI speed in terms of MHz or GT/s.
- Turbo mode options.



See sig for link, pm me if you find any errors.

[YukonTrooper]

What are your load temperatures, Tot? I bet at 4.1GHz they are

[The-Fox]

no youre safe. it has to be within .5 to be safe and you are there. But i need something like that to tell me what QPI voltage i am running!!!!!!

1. You can set your voltage on BIOS instead of AUTO, Just not sure if the + is from 1.15 or 1.375, Probably 1.15v.
2. You can use MSI OverClocking Center, get it here
3. You can use MSI GreenPowerII, get it here.

[nemesisrobot]

I've now got my QPI at 1.35v, and that allowed me to lower vCore to about 1.22-1.23 in the BIOS. Any lower seems to BSOD on Prime within an hour. How can you guys be running at sub 1.2v at higher clocks?!

[tot31]

What are your load temperatures, Tot? I bet at 4.1GHz they are

1 got max temp 72C on 100% load running at 3.5Ghz with 1.06V. I think if I run at 4.0Ghz the temp will be 90C or above.....

[roller11]

based on my own experiences, and from reading this thread, seems like there is a high correlation between week number and OC ability on the i7 920. The earlier the week, the better OC. For example, I've tested three different batch 3841xxxx 920s (week 41) and they were all duds. The chips that are getting phenominal OC's in the 45-4600Mhz range are all earlier chips, like 3834xxxx - 3836xxxx batch. Has anyone gotten a really good OC out of, say, a week 39 or later?

[Splave]

I have a week 37 and its a dud with HT on, HT off im 4.3ghz stable 1.4375 vcore

[burningrave101]

Something I think people need to quit confusing in their overclocks is that a 920 running at 20x200 = 4Ghz is NOT the same thing as 21x190 = 4Ghz with Turbo mode enabled. The higher multiplier from Turbo mode does not function in the same manner as actually selecting a higher multiplier. Turbo mode only engages when there are idle cores. When 1 or more cores are running idle it increases the speed of the cores that are under load. If however you have all 4 cores running under full load such as when you are running Prime95 then Turbo mode is not increasing the speed of the cores to 4ghz in the example above. You are at 3.8Ghz. And that is why if you disable Turbo mode and try running 20x200 instead your system most likely isn't Prime stable unless you increase the vCore more.

[ReverendMaynard]

Something I think people need to quit confusing in their overclocks is that a 920 running at 20x200 = 4Ghz is NOT the same thing as 21x190 = 4Ghz with Turbo mode enabled. The higher multiplier from Turbo mode does not function in the same manner as actually selecting a higher multiplier. Turbo mode only engages when there are idle cores. When 1 or more cores are running idle it increases the speed of the cores that are under load. If however you have all 4 cores running under full load such as when you are running Prime95 then Turbo mode is not increasing the speed of the cores to 4ghz in the example above. You are at 3.8Ghz. And that is why if you disable Turbo mode and try running 20x200 instead your system most likely isn't Prime stable unless you increase the vCore more.

Are you generalizing all X58's with that statement? If the person is using a Gigabyte board that 21x multi does not drop to 20. It's permanent

[Truckchase!]

Are you generalizing all X58's with that statement? If the person is using a Gigabyte board that 21x multi does not drop to 20. It's permanent

Asus sticks @ 21 with all cores loaded as well provided your temps are in the clear.

[ReverendMaynard]

Asus sticks @ 21 with all cores loaded as well provided your temps are in the clear.

gotchya, so it has a TDP limit. The Gig's don't which is why I bought it

[burningrave101]

Are you generalizing all X58's with that statement? If the person is using a Gigabyte board that 21x multi does not drop to 20. It's permanent

Well I'm just generalizing it based on how the technology works according to the info I've read about Intel Turbo mode.

Turbo mode and overlocking

The Turbo mode of the Core i7 is not an architectural characteristic but a functionality that Intel has already implemented on certain versions of the Core 2 Mobile under the name IDA (Intel Dynamic Acceleration). This mechanism consists of accelerating in a dynamic and temporary way the clock speed of one or several of the cores when others are not called for. The concept is based on the fact that many applications consist of one or two threads and therefore do not use all of the multi-thread processing potential of a multi-core processor. When the case arises, the Turbo mode comes into play (under control of the operating system) and increases the multiplying coefficient of the one or several cores in question.

Thus, Nehalem is an evolutionary step in the control of internal voltage and clock frequency. Up until now this control was handled by the operating system and the processeur offered the possibility of external control of the multiplying coefficient (FID: Frequency Identifier) and voltage (VID: Voltage Identifier) which forms the foundation for EIST (Enhanced Intel SpeedStep Technology). With the Core i7, these parameters are no longer external and it alone can change them in order to have control over its thermal dissipation. Indeed, the processor is capable of estimating the power it consumes at each moment (voltage x intensity of the current consumed) and of course to control it with the help of frequency and voltage parameters. Therefore, a Core i7 model will no longer be characterized by its maximum FID and VID but by its maximum TDP. Software (the BIOS and operating system) no longer controls the FID/VID combo as it did with previous generations but rather « Power-States » (or P-State). These power levels are defined based on the overall TDP of the processor at the maximum frequency outside of Turbo mode.

As an example, the Core i7 has an announced TDP of 130 Watts at 2.93 GHz, or 22 x 133. For each intermediate multiplying coefficient, the estimated TDP equals:

TDP[coeff] = (coeff / max_coeff)3 x TDP_core + TDP_uncore

The “uncore” part of the processor (IMC and L3 cache) is not subject to P-states and thus 20 of the 130 Watts is constant. We therefore obtain for example at 14 x 133 = 1.86 GHz:

TDP[14] = (14/22)3 x 110 + 20 = 48 Watts.

The multiplying coefficient varies between 12x and 22x which gives us ten P-states between 37 and 130 Watts.

The Turbo mode thus operates within the framework of this internal control of the processor’s overall TDP. The absence of activity of one or several cores results in a lowering of the overall TDP and thereby offers the Turbo mode the opportunity to accelerate cores that are called for.

With this new protection mechanism via control of the TDP, there is the obvious question of overclocking as the processor’s maximum TDP is quickly surpassed. A priori, Intel will finally not set any limitations here and it will be possible to go beyond this value, but of course the Turbo mode will not be in effect.

Only variations of Core i7 “Extreme Edition” models will enable modifying the TDP ceiling. However, note that these modifiable parameters on the Core i7 XE will only concern the Turbo mode. Should the need arise, it will thus be possible to modify the maximum multiplying coefficient as well as the TDP but this will not mean that the processor will function the entire time under these parameters.

http://www.behardware.com/articles/7...hitecture.html

So how are you guys finding out that the 21x multiplier sticks and is enabled at all times under full load of all cores? Is this something that someone found out during testing? How did they test?

Because for my Gigabyte EX58-Extreme this is what is says on the box:

Intel Turbo Boost Technology
Dynamically reroutes power to improve performance
For situations where all cores are not needed for a particular workload, Intel Turbo Boost Technology powers down the idle cores and dynamically reroutes the power to the active cores, boosting their performance without wasting power. This allows the active cores to achieve significant overclocks.

All I know is that it's easier to get 20x190 w/ Turbo stable than it is 20x200 w/ Turbo disabled. And I doubt that it's simply due to a 10Mhz increase in the BLCK which would raise the QPI, Uncore, and Memory a little but still well within stable spec.

[ReverendMaynard]

Don't lean on articles from before the cpu was actually released.

[burningrave101]

Don't lean on articles from before the cpu was actually released.

Well like I said, link me to where you guys are reading that the 21x multi sticks just like a normal multiplier on the Gigabyte or Asus boards.

[dejanh]

Well I'm just generalizing it based on how the technology works according to the info I've read about Intel Turbo mode.



http://www.behardware.com/articles/7...hitecture.html

So how are you guys finding out that the 21x multiplier sticks and is enabled at all times under full load of all cores? Is this something that someone found out during testing? How did they test?

Because for my Gigabyte EX58-Extreme this is what is says on the box:



All I know is that it's easier to get 20x190 w/ Turbo stable than it is 20x200 w/ Turbo disabled. And I doubt that it's simply due to a 10Mhz increase in the BLCK which would raise the QPI, Uncore, and Memory a little but still well within stable spec.

Yeah, that was the case initially, but then Gigabyte decided to permanently scrap sticking to the specs for TurboBoost and just have it on all the time, no matter TDP limts or anything. Hence, Giga boards have this on all the time.

Well like I said, link me to where you guys are reading that the 21x multi sticks just like a normal multiplier on the Gigabyte or Asus boards.

This is not "written" anywhere. People own these boards, and they know, they have tested it time and time again. I have an Asus board and mine shuts off Turbo as soon as temps get into the 80s. Asus - TDP limit enabled and cannot be disabled, Gigabyte - TDP limit permanently disabled.

[JWMc]

Someone corrected me earlier on this. The way turbo boost works is it gives you stock multi +1 for -all- cores if the thermal limits aren't reached, and the processor is loaded enough. If 3 of the 4 cores are doing nothing and switched off, the processor multi will go to stock multi + 2 for all cores but only one of them is active so effectively it's one.

The reason it's easier to get stable at 4.0ghz @ 20x190 with turbo instead of 4.0 ghz @ 20x200 without turbo is that the memory system and Uncore aren't being run at as high of a speed due to the reduced BCLK.

See this document:

http://www.intel.com/support/process.../CS-029908.htm

[The-Fox]

burningrave101, here is a TON of real world results provided by people who own the P6T Deluxe/RIIE and i920:

http://www.xtremesystems.org/forums/...d.php?t=212389

The Multi is depended on Temp in ASUS, they implemented it differently than other vendors like Gigabyte.

[ReverendMaynard]

TDP is determined by the board, not the cpu so it is up to the manufacturers discretion.