******Gigabyte GA-P67A-UD4 Discussion Thread******

[chew*]

Thanks m8, much appreciated.
Brought mine in today, shop owner called me this afternoon he already got a new one.
I'm picking it up tomorrow, let's hope this one works correctly.

Which bios came with yours now?
Did you flash yours directly to the most recent one?

chew* :

did you try using the Normal + DVID setting, instead of setting VCore manually?

Should that matter? Not that i didn't try but it still shouldn't matter........

There shouldn't be a vcore issue if its not a chip limit.

I already established that the issue surfaced after a few days 1.4+ was fine now its a problem to set over 1.325.

came with a really early bios, I ended up flashing both the backup and main bios due to being plagued by flashbacks and rebbot loops kept triggering flashback.

Problem manifested itself however after that and like i said earlier it happened after messing with ez tune in windows. I was running prime and dropping a notch every hour, when i found the crash point this problem showed up.

[Falkentyne]

Did you clear the cmos with the power cord unplugged?

[chew*]

Did you clear the cmos with the power cord unplugged?

Me? If unplugging the board putting it back in box then putting it back on bench a little later means removing power cord then yes.

I'm fairly versed at debugging hardware, what i'm reporting has been tested over a 3 day period.

Yes battery was removed, been testing without it to be honest.

[LeetMiniWheat]

my i5 2500K is topping out at 4.9GHz on my UD4 with 1.476V shown in CPU-Z. gone up to 1.516v and I can't seem to get the extra 100MHz out of it. tried 102x49 and 101x49 (4.95GHz). 102 BSOD's weird 03B and 0A's, and 101 seems to 124. Ideally I'd like 100x50 though. f7b and f7h are same results.

Vcore: Normal (1.375V) +0.125v offset w/ LLC
QPI/VTT: 1.140
VCSA: 0.920
CPU PLL: 1.7200

I'm starting to wonder if the Asus or MSI would get me that extra 100MHz easier and with less Vcore.

[chew*]

my i5 2500K is topping out at 4.9GHz on my UD4 with 1.476V shown in CPU-Z. gone up to 1.516v and I can't seem to get the extra 100MHz out of it. tried 102x49 and 101x49 (4.95GHz). 102 BSOD's weird 03B and 0A's, and 101 seems to 124. Ideally I'd like 100x50 though. f7b and f7h are same results.

Vcore: Normal (1.380V) +0.135v offset w/ LLC
QPI/VTT: 1.140
VCSA: 0.920
CPU PLL: 1.7200

I'm starting to wonder if the Asus or MSI would get me that extra 100MHz easier and with less Vcore.

Lower temps likely all you need.

I tested my chip on high end water versus air, i can prime at 5 gig on the high end water with 1.5v.......doubt chip would last long like that 24/7 though.

Really 4.5-4.7 with 1.3v-1.35 is the sweet spot safety wise it seems, if you chip can do more at that voltage even better.

[LeetMiniWheat]

Lower temps likely all you need.

I tested my chip on high end water versus air, i can prime at 5 gig on the high end water with 1.5v.......doubt chip would last long like that 24/7 though.

Really 4.5-4.7 with 1.3v-1.35 is the sweet spot safety wise it seems, if you chip can do more at that voltage even better.

are they really that temp sensitive? I keep it around 68c-72c with cold air from outside and H50

this is my results sofar on this board

Batch#: L040B354 i5 2500K on P67A-UD4
1.3?-1.35 = 4.5Ghz
1.37-1.38 = 4.6GHz
1.45 = 4.8GHz
1.476 = 4.9GHz

looks like i'm 100mhz short of the 5GHz club unless there's some magic BIOS that gets me 5.0.

[chew*]

are they really that temp sensitive? I keep it around 68c-72c with cold air from outside and H50

this is my results sofar on this board

Batch#: L040B354 i5 2500K on P67A-UD4
1.3?-1.35 = 4.5Ghz
1.37-1.38 = 4.6GHz
1.45 = 4.8GHz
1.476 = 4.9GHz

looks like i'm 100mhz short of the 5GHz club

For 5 gig prime I had 50C cpu temps, obviously at those temps and 1.5v I was using cool water and a very high end water setup

Same chip on normal air fails around 4.7 with 1.4v more so the temps are 70C plus though.

Your chip sounds similar to mine clock per voltae scaling.

[LeetMiniWheat]

I may be on to something here... lowered the VCSA from 0.920 to 0.875 and 4.9 seems stable sofar @ 1.452 instead of 1.476 but as soon as I post this, i bet it will BSOD lol.

anyone have similar results from lowering system agent? anyone find a good range for it?

EDIT: nevermind, BSOD 124 after 10 minutes. damn this vcore hungry chip

[chew*]

I may be on to something here... lowered the VCSA from 0.920 to 0.875 and 4.9 seems stable sofar @ 1.452 instead of 1.476 but as soon as I post this, i bet it will BSOD lol.

anyone have similar results from lowering system agent? anyone find a good range for it?

EDIT: nevermind, BSOD 124 after 10 minutes. damn this vcore hungry chip

IF you got more stable with lower volts i'd be inclined to say damn this "hot" chip.

Likely heat went down with less volts allowing you to be stable longer.

Hang the system out a window bet it stabilizes

[LeetMiniWheat]

IF you got more stable with lower volts i'd be inclined to say damn this "hot" chip.

Likely heat went down with less volts allowing you to be stable longer.

Hang the system out a window bet it stabilizes

nah it's stable between 1.464v and 1.476v @ 4.9, but temps are 68-76 depending on room temp. (i.e. if I'm sitting here in jacket/gloves with windows open)

I'm still hoping a new BIOS will somehow get me that extra mile for 5.0

[Falkentyne]

Me? If unplugging the board putting it back in box then putting it back on bench a little later means removing power cord then yes.

I'm fairly versed at debugging hardware, what i'm reporting has been tested over a 3 day period.

Yes battery was removed, been testing without it to be honest.

Yes but did you actually press the CMOS clear button?
I don't know if the UD4 has a clear cmos button, or if it uses a jumper...

Because clearing the CMOS with the button fixed many strange issues for people with the Asus board. I don't know if running without a battery without clearing the CMOS with the button counts or not.

If you still find that the vcore is still capped, then just RMA...

[chew*]

Yes but did you actually press the CMOS clear button?
I don't know if the UD4 has a clear cmos button, or if it uses a jumper...

Because clearing the CMOS with the button fixed many strange issues for people with the Asus board. I don't know if running without a battery without clearing the CMOS with the button counts or not.

If you still find that the vcore is still capped, then just RMA...

Yah i stuck a jumper on it and let it sit.

Board is a review board, not gonna RMA unless giga advises me to.

EDIT

I should note that i believe the issue was caused by myself. I installed ET6 and had limited functionality but proceeded to use it without determining the cause ( which was lack of intel management drivers ).

I believe using ET6 without these drivers can be problematic, using it without them is not reccomended and it is required to work with them as well as other manufacturers software.

I suggest you learn from my personal mistake and install the correct software to use windoows based OC utilities.

[Falkentyne]

Ouch...do you think you damaged the board this way?

The CPU still runs at higher vcore on the UD7 (up to 1.6v) right?

[chew*]

Ouch...do you think you damaged the board this way?

The CPU still runs at higher vcore on the UD7 (up to 1.6v) right?

Yep no issues, UD7 is rocking fine, then again i was wise to the Intel management driver needing to be installed by the time i got to that board.

[Gaul]

[aytacaksel]

i want a bios allow me oc to 5 ghz at 1.4v :'(

[wekktor]

easy tune 6 doesnt work to change the multiplier Is it possible in any way to change it in windows? Bclk and voltages works fine to adjust

Thx

[Andrea deluxe]

easy tune 6 doesnt work to change the multiplier Is it possible in any way to change it in windows? Bclk and voltages works fine to adjust

Thx

Do you enabled in bios real time multiplier change?

[Andrea deluxe]

i want a bios allow me oc to 5 ghz at 1.4v :'(

the problem is

the 80% of cpu wall at 4600
the 15% at 4700
the 5% wall at 4800
the rest around 4900/5000/5200

[LeetMiniWheat]

the problem is

the 80% of cpu wall at 4600
the 15% at 4700
the 5% wall at 4800
the rest around 4900/5000/5200

I've been hearing that some BIOS allow lower VCore though. like F2 or F3. I haven't personally tested.

your list seems fairly accurate, mine walls at 4600 1.38v, and needs 1.44v for 4.8GHz and 1.476v for 4.9GHz. Any idea how safe 1.44v is for 24/7?

also, fine tuning of VCore on this board seems to not work at all... it only makes large jumps. like 1.416 1.426 (1.43?) 1.440 1.452 1.464 1.476 1.488

Edit: only tested with DVID, not sure about just strictly manual voltage and not the offset
Edit2: fixed typos in voltages
Edit3: these volts are what CPU-Z reports. it doesn't change unless I make at least +0.010 or +0.015 to DVID.
Edit4: Also, VDIMM is not precise at all either... why only 1.64v or 1.66v? detects 1.48V with 1.64 and 1.68V with 1.66 --> are Asus, MSI, and Biostar boards more precise with the Vcore and VDIMM?

[wekktor]

Do you enabled in bios real time multiplier change?

yes I have enabled it, but I can only lower the multi, no raise ?

[GAR]

This board has been a weird one for me.....every once in a while i will get the non stop reboot, the computer turns on and off when trying to power on, even tho when it was on, it was 100% stable and properly shut down, passes linx, prime95 and every single game thrown at it, im kind of glad the SATA issue has risen, will give me a reason to return this board, when its on, and booted into windows, its rock solid, but everytime i turn it on i have to cross my fingers for it to boot, and wtf is wrong with this board, the bios will not reset with the jumper, you need to power down, take the power cable out, wait 5 mins then try again hoping it turns on, this has to be one of the worst motherboards i have ever used, i thought it was good, boy was i wrong, i should have gone with asus, even tho i prefer gigabyte.

[LeetMiniWheat]

clear cmos works fine for me. I just stick a screwdriver in there.

and I know what you mean about the reboot loops, but I haven't seen it randomly like you're saying. only when I change something in the bios it will sometimes do it even tho it should be working fine. sorry if this is incoherent, it's 2am here...

[LeetMiniWheat]

so my system was stable with 1T command rate 2133MHz DDR3 for about a week but now suddenly it doesn't want to do 1T anymore. (granted, the ram is rated for only 2T) stable as rock with 2T but with 1T it BSOD's the memory error stuff... I wonder if all this high cpu voltage has somehow weakened my sandy bridge? or maybe something with the board? no amount of QPI/VTT seems to make a difference anymore. It used to want 1.140v QPI/VTT for 1T. weird, huh

[X.T.R.E.M.E_ICE]

2 Questions.

How do I change the LLC level?
how do I change the dynamic vcore setting?
dynamic vcore worked fine on MSI but not on Gigabyte-UD4.