SOLVED [email protected] on 1.08vcore?

Hiya guys

I've just been delivered my replacement GB UD4-B3 board, so lost all my old BIOS settings, etc.

Set her up from memory with Dynamic vcore control on... standard vcore with a dynamic of +.20

Looking from the screenshot that it's holding steady at 1.08vcore, which is waaaay less than I was planning, but it's running LinX as I type (just finished 20# in 30m15s without errors) and is as happy as could be...

Temps are a bit high right now, but given that it's a fresh application of Zalman's paste and it's under a CNPS9900, I'm pretty happy.

Do the Times & Gflops look about right?

It's running 2x4Gb of G.Skill @ 665 presently. (It is 2000Mhz ram, but I forgot to set it to XMP). Good test for the cpu I guess.

Should I leave the vcore that low? I've just kicked her off again at the same settings, for a 100 run of LinX and we'll see how that goes.

Value a bit of help here guys, as this all seems a bit wierd to me.

What is the voltage set at in the BIOS? The Gigabyte P67 boards are notorious for reporting voltages much lower than they are actually feeding the CPU.

In BIOS... vcore is set to 'standard' and in the 'dynamic' setting... the one just under vcore... it is set to +.20.

Here is a s/s of the 100 pass of LinX... check out the max vcore this time (with no BIOS changes at all from the previous run)



I think I might set the vcore manually to 1.08 in BIOS and set her up for another 100x run overnight... see what happens.

I'm guessing it will be closer to 1.4v based on temps and +.2v

tool_462 said:

I'm guessing it will be closer to 1.4v based on temps and +.2v

Click to expand...

Agreed... I set it manually to 1.08... rebooted before even displaying the windows logo (trying to boot Win7).

Manually set it to 1.290 and cpuz & HWMon both said 1.07

Have set it up to do another 100x run, so will be interesting to see the results when I wake up.

edit... There must be a way to limit the upper vcore level using dynamic... ideas appreciated

Okay... Ramped the ram up to XMP and it's part-way through the next run... loving the temp's...

Go grab CPUZ 1.57. I know voltages werent reported too well in 1.56.1 (didnt fix till 1.56.3 for most people).

Please use HWinfo to monitor voltages on the B3 boards.
CPU-Z doesn't support B3 yet.
The 1.08v you saw is the VTT voltage.

Falkentyne said:

Please use HWinfo to monitor voltages on the B3 boards.
CPU-Z doesn't support B3 yet.
The 1.08v you saw is the VTT voltage.

Click to expand...

Interesting... the 1.08v is sitting right out from the vcore title (in HW-Mon)... or are you meaning another program again?

CPUZ works just fine on B3 boards. There was no change to anything on the B3 boards other than the SATA port fix...... And HWMonitor has always had issues with everything (not reporting correctly).

doz said:

CPUZ works just fine on B3 boards. There was no change to anything on the B3 boards other than the SATA port fix...... And HWMonitor has always had issues with everything (not reporting correctly).

Click to expand...

Cheers... good to know.

So CPU-Z in the first s/s above, showing 1.08vcore... would that have been right or wrong?

Dont know. I know 1.56.1 for me was garbage at reporting vcore. I upgraded to 1.56.3 (now up to 1.57) and it corrected the voltage readings.

If you're truly running 1.08Vcore you either have the most perfectly lithographed, flawlessly packaged, perfectly pin arrayed CPU on with the most perfect mount to the most perfectly made motherboard ever produced or there is a reporting error. I don't think it is actually possible for a 2600K to reach 4.5Ghz on that vCore. I've never heard of it on anything less than about 1.25 on a very good bin chip.

Weirder things have happened though. I have a very strange Conroe that only needs about 1.4V to hit 4Ghz from a stock of 2.3Ghz and about 1.28V under load. And that's 65nm. You're 32.

Cheers guys... appreciate the input.

I've set the BIOS at 1.29 (from memory) and Ram at XMP, and it is passing LinX as many as I like without errors at 4.4Ghz, so very happy, as is the Mission agency using it!.. encoded a 20 min vid last night just as a quick test... 4 minutes and it was done. All 8 cores in use, but only to 40-60%.

The thing I love the most is the temp... during the final LinX runs... the max temp was 65 degrees... much cooler than the original UD4, so I suspect that the application of paste wasn't as good as this time, as it's the same cooler, and this time... it's in a nice HAF case, not out on the bench.

Might throw up a pic later. Might also see how far I can push the multi seeing it is so cool.

When the Gigabyte B2 revisions first came out there was some kind of bug where CPUZ reported VCCio (vtt) instead of vcore. I believe a CPUZ update corrected the problem. I think I saw somwhere else on one of the forums that the new Gigabyte B3 boards might be doing/reporting this same type of bug again? If your chip didn't show report vcore like that with your old board and CPUZ then I would think that something here seems buggy.

yeh, Y not download the latest cpuz anyway...

I'd bet my 2600K that's the VccIO voltage showing in 1.5.1