• Welcome to Overclockers Forums! Join us to reply in threads, receive reduced ads, and to customize your site experience!

HT @ 4.2Ghz w/ i7 940

Overclockers is supported by our readers. When you click a link to make a purchase, we may earn a commission. Learn More.
TaHoward

TaHoward

Registered
​
Joined
Aug 4, 2010
Location
Tempe AZ
Having a little trouble getting 4.2Ghz stable w/ HT enabled on my i7 940 in a new system build.

4.2Ghz was a breeze @ 1.35 Vcore without HT. Once enabled, it takes an aditional .07V to stop BSOD 101 errors during LinX or P95.

Just want to know if this kind of voltage difference is normal with HT enabled and if it is; is it worth keeping HT?

HT enabled currently maxes my temps out at 72C. Where as without 59-60C is the max temp.

Is this a reasonable max load temp during LinX/P95 stress loads?

*****

STABLE!!! Finally, but not at 4.2Ghz. It was becoming obvious my processor just needed more voltage than I wanted to dish out. I dropped down to 4.0Ghz 200BLKx20 with 1.38Vcore. Ran P95 blend for over eight hours without a hiccup. Now it's just finding the lowest voltage for all the other settings. Temps are much lower.

I have one more question however. I've heard even multipliers are bad voodoo on the i7 series. Is there a good chance raising the BLK a little and dropping to x19 will allow a lower stable Vcore?
 
Last edited:
I think a max of 72°C is pretty good for stress on HT. Its really up to you and how you use the PC to determine if keeping HT is worth the extra heat. It seems you have very good cooling so im guessing if you are gaming your temps are around 45-55°.

Personally i have mine with HT enabled and Turbo off for my use. I only use turbo for 4.4GHz or 4.6GHz.
 
Ok, so it sounds like 72C on stress is alright. I might reseat the block to see if it can go down a smige still.

As for a 24/7 Vcore. Should 1.45V be ok if it turns out the chip needs more?
 
Thats a bit much for 4.2GHz. Above that frequency you can use that... 1.38v to 1.4v should be sufficent. If you still need more vcore, you can up it to 1.44v (only for testing then reduce vcore) and i dont see why it would need anything higher. For now lets start with 1.38v and try goin a little lower...like 1.37v. On my set up i run 1.34v...just so you know but the chips are different so you may need more or less on yours.

With your cooling i think the temps should be lower even with HT on. Double check your bios fan control and change it to PWM to see if that may help reduce temps.
 
Last edited:
4.2Ghz seems too much for my processor to handle. I backed off to 4.0Ghz and 1.381 Vcore 20 pass LinX stable using ALL mem setting. I'll run 24hrs prime Small then Large FTT later.

I was pushing 1.4318 in BIOS for 4.2 and still getting BSOD thirty minutes in to LinX. Ran a couple passes of memtest86+ to verify QPI and memory settings are fine. Don't feel like pushing any more voltage for a 24/7 OC.

Oh well, 4Ghz is still pretty damn good.

By the time I'm ready to upgrade I'll hope there is a nice 1366 Xeon a few years down the road to grab.
 
Last edited:
Are you sure you need >1.43 Vcore for a 4.2 GHz clock? Post all of the voltages, timings, strap, etc. when you were running at that frequency. You can use this BIOS template for the P6X58D-E to post all of the settings...

Code: 
Target CPU Frequency: 3192MHz
Target DRAM Frequency: 1066MHz
Ai Overclock Tuner [Auto]
CPU Ratio Setting [Auto]
Intel(R) SpeedStep(TM) Tech [Enabled]
Intel(R) TurboMode Tech [Enabled]
Xtreme Phase Full Power Mode [Auto]
DRAM Frequency [Auto]
UCLK Frequency [Auto]
QPI Link Data Rate [Auto]
ASUS/3rd Party UI Priority [ASUS Utility] 

CPU Voltage Control [Manual]
CPU Voltage [Auto]
CPU PLL Voltage [Auto]
QPI/DRAM Core Voltage [Auto]
IOH Voltage [Auto]
IOH PCIE Voltage [Auto]
ICH Voltage [Auto]
ICH PCIE Voltage [Auto]
DRAM Bus Voltage [Auto]
DRAM DATA REF Voltage on CHA [Auto]
DRAM CTRL REF Voltage on CHA [Auto]
DRAM DATA REF Voltage on CHB [Auto]
DRAM CTRL REF Voltage on CHB [Auto]
DRAM DATA REF Voltage on CHC [Auto]
DRAM CTRL REF Voltage on CHC [Auto]

Load-Line Calibration [Auto]
CPU Differential Amplitude [Auto]
CPU Clock Skew [Auto]
CPU Spread Spectrum [Auto]
IOH Clock Skew [Auto]
PCIE Spread Spectrum [Auto]
 
Last edited:
Target CPU Frequency: 4200MHz
Target DRAM Frequency: 1600MHz
Ai Overclock Tuner [Manual]
CPU Ratio Setting [21]
Intel(R) SpeedStep(TM) Tech [Disabled]
Intel(R) TurboMode Tech [Disabled]
Xtreme Phase Full Power Mode [Auto]
DRAM Frequency [1600]
UCLK Frequency [3200]
QPI Link Data Rate [7.2]
ASUS/3rd Party UI Priority [ASUS Utility]

DRAM Timing Control

CPU Voltage Control [Manual]
CPU Voltage [1.4318]
CPU PLL Voltage [1.8]
QPI/DRAM Core Voltage [1.225]
IOH Voltage [1.2]
IOH PCIE Voltage [Auto]
ICH Voltage [1.2]
ICH PCIE Voltage [Auto]
DRAM Bus Voltage [1.66]
DRAM DATA REF Voltage on CHA [Auto]
DRAM CTRL REF Voltage on CHA [Auto]
DRAM DATA REF Voltage on CHB [Auto]
DRAM CTRL REF Voltage on CHB [Auto]
DRAM DATA REF Voltage on CHC [Auto]
DRAM CTRL REF Voltage on CHC [Auto]

*************************************************

Load-Line Calibration [Auto]
CPU Differential Amplitude [Auto]
CPU Clock Skew [Auto]
CPU Spread Spectrum [Auto]
IOH Clock Skew [Auto]
PCIE Spread Spectrum [Auto]


This setting lasts about thirty minutes in to P95 small FTT before BSOD 101 or 124 error. I also ran 2 passes of memtest86+ to see if the QPI was too low.

Edit: Tried your settings. Boots in to windows fine, BSODs 2 seconds in to LinX with 124 error. P95 will just lock up. I'll bump the vcore a .025 and see what happens.
 
Last edited:
Give these settings a try for 4.2 GHz w/ HT enabled...

Code: 
Ai Overclock Tuner.....................[Manual]
CPU Ratio Setting......................[21.0]
Intel(r) SpeedStep(tm) Tech............[Disabled]
Xtreme Phase Full Power Mode...........[Enabled]
BCLK Frequency.........................[200]
PCIE Frequency.........................[100]
DRAM Frequency.........................[DDR3-1603MHz]
UCLK Frequency.........................[3208MHz]
QPI Link Data Rate.....................[Auto]

CPU Voltage Control....................[Manual]
CPU Voltage............................[1.35]
CPU PLL Voltage........................[1.80]
QPI/DRAM Core Voltage..................[1.335]
IOH Voltage............................[1.12]
IOH PCIE Voltage.......................[1.50]
ICH Voltage............................[1.20]
ICH PCIE Voltage.......................[1.50]
DRAM Bus Voltage.......................[1.64]
DRAM DATA REF Voltage on CHA...........[Auto]
DRAM CTRL REF Voltage on CHA...........[Auto]
DRAM DATA REF Voltage on CHB...........[Auto]
DRAM CTRL REF Voltage on CHB...........[Auto]
DRAM DATA REF Voltage on CHC...........[Auto]
DRAM CTRL REF Voltage on CHC...........[Auto]

Load-Line Calibration..................[Enabled]
CPU Differential Amplitude.............[800mV]
CPU Clock Skew.........................[Delay 300ps]
CPU Spread Spectrum....................[Disabled]
IOH Clock Skew.........................[Auto]
PCIE Spread Spectrum...................[Disabled]

C1E Support............................[Disabled]
Hardware Prefetcher....................[Enabled]
Adjacent Cache Line Prefetch...........[Enabled]
Intel(r) Virtualization Tech...........[Disabled]
CPU TM Function........................[Enabled]
Execute Disable Bit....................[Enabled]
Intel(r) HT Technology.................[Enabled]
Active Processor Cores.................[All]
A20M...................................[Disabled]
Intel(r) SpeedStep(tm) Tech............[Disabled]
Intel(r) C-STATE Tech..................[Disabled]
 
TaHoward said:
Edit: Tried your settings. Boots in to windows fine, BSODs 2 seconds in to LinX with 124 error. P95 will just lock up. I'll bump the vcore a .025 and see what happens.
Click to expand...

Start bumping the Vcore up in increments starting at 1.35V, and leave all the other settings I posted as is.
 
This is starting to look some what good. I'm starting to get an error catch in LinX instead of BSODs at 1.375. Bumping up to 1.4 as I type.

System used to just give me a BSOD @ 4.2 w/ HT no matter what instead of error catches. :)

What's the trick?

Edit: Stopped on error 300 seconds in to LinX with 768Mb mem usage. Gonna bump to 1.4125
 
Last edited:
Normally you don't need that much Vcore for a 4.2 GHz clock on a 940, of course all chips aren't created equal. You might want to try bumping VTT to 1.35V and IOH to 1.18V at a Vcore of ~1.40V.
 
1.4125 stable after a short LinX run 5m 30s. I'll give ALL mem setting a run now. i7 runs hot! 73C Think I should remount my block or is this expected?
I got a lot to learn still.

Edit: Failed about 16 minutes in to LinX full mem usage. Going to run Prime small FFT while I sleep. Hopefully I don't need anymore vcore and just IOH/QPI adjustment.
 
Last edited:
redduc900 said:
Normally you don't need that much Vcore for a 4.2 GHz clock on a 940, of course all chips aren't created equal. You might want to try bumping VTT to 1.35V and IOH to 1.18V at a Vcore of ~1.40V.
Click to expand...

The 940 chips are known to be a pain to OC far more than the 920. Im not sure what the deal is with those chips.
 
TaHoward said:
1.4125 stable after a short LinX run 5m 30s. I'll give ALL mem setting a run now. i7 runs hot! 73C Think I should remount my block or is this expected?
I got a lot to learn still.

Edit: Failed about 16 minutes in to LinX full mem usage. Going to run Prime small FFT while I sleep. Hopefully I don't need anymore vcore and just IOH/QPI adjustment.
Click to expand...

Thats a lotta vcore for 4.2GHz... the 940's can be a pain to OC for some reason. Anyway, whats your uncore frequency running @? Run your CPU multiplier @ 20X. Change the uncore multiplier to 19 or 18 and turn Turbo off for now - you want that uncore running less than 4GHz for now. Then if its stable reduce the vcore. You should not need more than 1.37v to run that speed.
 
Its probably an old chip like mine. I need 1.4 just to run 4ghz.

So all you guys running 4.2+ with 1.35 have it made in the shade.
 
Ya, the 940's never had a D0 revision. I got this chip for nothing on the dollar from an Intel program; I might just sell it and grab a 930/920 D0.

My last settings posted BSODed at some point during the night on P95 small FFT. I didn't see the error code, but I'll bump IOH to 1.18 and re-run the test before trying VTT increase. Otherwise I'll back off to 4.0Ghz and see what happens.

It's strange LinX will pass or give errors. P95 seems more BSOD happy. Pretty sure Vcore is good now, unless i7 chips are just BSOD happy. I'll run another memtest if this P95 run BSODS on me again even though I'm running small FFT.

Edit:

Got another BSOD Error code 124 about 30 minutes in to P95. Bumping VTT...
 
Last edited:
Re applied my thermal paste after cleaning the CPU. With 1.418 my load temps dropped 3C. Staying below 70C now :D

I originally used the dot in the middle. This time I spread the compound evenly across the top.

Bumping VTT to 1.35 caused a BSOD only 16 minutes in to the test. Seems I found my bug. I'll try VTT between 1.32 and 1.35.

Made it past the 30 minute mark in P95 small FFT with 1.300 VTT/QPI. Hope there is a trend.

Edit: Dropped QPI to 1.281 after a BSOD 120 minutes in to the last P95 test. Ran P95 on this for 90 minutes so far. Going to sleep on this and hope for the best in the morning.

Edit Again: BSOD Error 101 some point in the night. Read 101 is usually Vcore related. Considering the trend I'm going to try dropping QPI to 1.275

STABLE!!! Finally, but not at 4.2Ghz. It was becoming obvious my processor just needed more voltage than I wanted to dish out. I dropped down to 4.0Ghz 200BLKx20 with 1.38Vcore. Ran P95 blend for over eight hours without a hiccup. Now it's just finding the lowest voltage for all the other settings. :D

I have one more question however. I've heard even multipliers are bad voodoo on the i7 series. Is there a good chance raising the BLK a little and dropping to x19 will allow a lower stable Vcore?
 
Last edited:
TaHoward said:
Re applied my thermal paste after cleaning the CPU. With 1.418 my load temps dropped 3C. Staying below 70C now :D

I originally used the dot in the middle. This time I spread the compound evenly across the top.

Bumping VTT to 1.35 caused a BSOD only 16 minutes in to the test. Seems I found my bug. I'll try VTT between 1.32 and 1.35.

Made it past the 30 minute mark in P95 small FFT with 1.300 VTT/QPI. Hope there is a trend.

Edit: Dropped QPI to 1.281 after a BSOD 120 minutes in to the last P95 test. Ran P95 on this for 90 minutes so far. Going to sleep on this and hope for the best in the morning.

Edit Again: BSOD Error 101 some point in the night. Read 101 is usually Vcore related. Considering the trend I'm going to try dropping QPI to 1.275

STABLE!!! Finally, but not at 4.2Ghz. It was becoming obvious my processor just needed more voltage than I wanted to dish out. I dropped down to 4.0Ghz 200BLKx20 with 1.38Vcore. Ran P95 blend for over eight hours without a hiccup. Now it's just finding the lowest voltage for all the other settings. :D

I have one more question however. I've heard even multipliers are bad voodoo on the i7 series. Is there a good chance raising the BLK a little and dropping to x19 will allow a lower stable Vcore?
Click to expand...

Lowering the multiplier does help if you are trying to find your chipsets highest base clock rate. To be honest, 19X mulitplier isn't that far from 20X so you will probably run into a BSOD unless its lower than 16X. The lowest you should hit is 12X to really see what the max base clock is...but its still not worth it because you need to add almost the same amount of voltage (if not more) to keep the processor stable.

All i can suggest at this point is to try lowering your Vcore to 1.375v or 1.366v at the very least to hold her steady @ 4GHz. You can also run turbo for an effective 4.2GHz if its stable with those 2 voltages. Give it a shot :) If they don't work, bump it back to 1.38v. If it does work @ 1.366v i suggest you leave it here. I don't think the 940 will accept anything lower...but again, it doesn't hurt to try if you don't mind risking damagin your OS. Don't touch the VTT or QPI voltages for now, just leave them on auto until its stable. Then when you find the stable voltage you can play with those.
 
My 930 is overclocked at 4.038 with HT and turbo enabled. My system passed 50 passes of linx, 72 hours of prime 95 and 72 hours of World community grid. I personally prefer world community grid because I feel it stresses the computer components better.

I can overclock higher with a bit more vcore but I will take a 7/24 hour stable overclock over an unstable on. With more vcore temps tend to rise. Although my temps rarely break past low to mid 70s under stress. I decided I can't be to greedy and anything stable at or above 4ghz is a decent overclock.
 
You must log in or register to reply here.

Similar threads

Alvin777
Further Help in Overclocking the Q6600 for Two Disabled People
Replies
5
Views
921
Aleslammer
Aleslammer
Al_bundy
Alucryl - 3DFX - CNC - Retro pc
Replies
2
Views
398
EarthDog
EarthDog
M
10900k real world temperatures, what is too high/expected? msi z490 Settings
2
Replies
37
Views
11K
EarthDog
EarthDog
S
Higher temps with vcore offset than manual, am i missing something?
Replies
5
Views
928
ATMINSIDE
ATMINSIDE
B
First Successfull 10700k 5GHZ Overclock Opinion
Replies
5
Views
10K
ShrimpBrime
ShrimpBrime
Back