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

[bluestang]

Anyone care to supply me with a screenshot of the list of devices under "System Devices" in Device Manager for this MB? I'd appreciate it.

[Bullet92]

[bluestang]

^^ Thanks Bullet!!!

[Demo]

I measured in my red dots, the voltage is slightly higher

1.452v cpuz / digital meter 1.46v

Try to measure on the bluedots and its even higher..

[Mafio]

ATI ain't that bad in 3dmark 01
118325 pt with single HD4890 @ 1050-1150 MHz and a 2600K @ 5400 MHz



FPS:
CL 1694,1 :: CH - 541,2 :: DL - 1684,3 :: DH - 687,2 :: LL - 1623 :: LH - 762,2 :: N - 1124,6

[Gaul]

DON'T try F4 :
- Cold BOOT
- Lower score CPU Score VANTAGE -700 pts

[Frost_Bite]

After doing some pretty intense testing over the last three days, I'm going to have to agree with Gaul.

I initially thought F4 was good to go, but I started noticing some oddities here and there such as random BSOD and instability problems. I finally decided to fire LinX up and run it against my 4.8Ghz settings with the F4 BIOS. Three times in a row LinX failed within 7 seconds of being started on my 2500k. I upped the vcore from 1.38v all the way to 1.43 and still couldn't get it past 7 seconds. It passed 3DMark loops just fine and had no problems with SuperPI 32M.

I switched back to F3D and my initial 4.8GHz settings...and what do you know it passed 25 loops of LinX without breaking a sweat and my hottest core reached 51C according to HWMonitor. From now on, I'll be using F3D until something better comes along and I'm going to keep pushing this chip. I thought I had reached its max, but it was probably the F4 BIOS causing the majority of my headaches.

Full specs in case anyone was interested (same used for F3D and F4):

Multi: 48X
Blck: 1005 (100.5)
Mem: 1600MHz @ 9-9-9-24 1T
Vcore: 1.38v (1.37v under load F4 / 1.36v under load F3D)
QPI/vtt: 1.12v
VCSSA: 0.945v
PCH: 1.05v
PLL: 1.7v

[Gaul]

@ ^ = ha ha ha, then u must try F5 beta, better than F3D

[Demo]

@ ^ = ha ha ha, then u must try F5 beta, better than F3D

What makes F5 beta, should I expect great improvements like lower vcore requirements?

[Frost_Bite]

I have F5b on my BIOS flash drive, just haven't tested it since I was told it simply fixes S3 sleep functions.

What sort of improvements does F5b offer over F3d? Does it have better vdroop control or enhanced stability? My vcore is already somewhat low so I think expecting it to be lower through a BIOS change is a bit unrealistic. I'm curious as to what your input is regarding F5b.

Thanks

[Gaul]

What makes F5 beta, should I expect great improvements like lower vcore requirements?

I have F5b on my BIOS flash drive, just haven't tested it since I was told it simply fixes S3 sleep functions.

What sort of improvements does F5b offer over F3d? Does it have better vdroop control or enhanced stability? My vcore is already somewhat low so I think expecting it to be lower through a BIOS change is a bit unrealistic. I'm curious as to what your input is regarding F5b.

Thanks

Well in my CASE, i got same vcore to touch 5GH, BUT..........
- lower pll @ 1.6
- lower vtt @ 1.08
- lower pch @ 1.08
- lower vSA @ 0.92 def

[Falkentyne]

And what were those values set to on the F3 betas? Did you actually test them? I highly doubt a BIOS would suddenly let you use lower values....

[Gaul]

And what were those values set to on the F3 betas? Did you actually test them? I highly doubt a BIOS would suddenly let you use lower values....

as i remember with F3
- pll @ 1.70
- vtt @ 1.12
- ch @ 1.12
- vSA @ 0.945 def

[Falkentyne]

as i remember with F3
- pll @ 1.70
- vtt @ 1.12
- ch @ 1.12
- vSA @ 0.945 def

Why would you need those values? what happened if you used auto?

[Mafio]

did you measure those voltages with DMM or they are just what you set in the bios?
i'm asking that because with UD5 i see that with bios F3 real VDIMM was 0.12 volt higher of bios setting, with bios F4 or F5 it was 0.15 volt higher.
maybe that's because you need lower voltages.

[Falkentyne]

You SURE About that???

0.12v HIGHER means, 1.50v is 1.62v, and 0.15v higher is 1.65v.

Do you mean 0.012? Decimal places matter !

[Gaul]

Why would you need those values? what happened if you used auto?

good question, i'll never use auto

[Mafio]

You SURE About that???

0.12v HIGHER means, 1.50v is 1.62v, and 0.15v higher is 1.65v.

Do you mean 0.012? Decimal places matter !

yes, i'm sure and i know that 0.015 is different from 0.15
with bios F5 VDIMM 1.8 volt in bios results in 1.95 idle and near 2 volt during 32M
probably that's why i see alot of people doing crazy frequency with kinda low voltage, the board overvolt like hell.....

[Falkentyne]

Did you test this on F3G (f3h)?

Because that's the first I ever heard anyone say the vdimm was that high. Sin tested it and it was almost spot on (that may have been UD7, though).

[Demo]

@Gaul

I've played around those voltages and my CPU doesnt seem to respond to any aside from just vcore.. Im actually concerned if lowering those values aside from PLL voltage would cause instabilities.. They dont seem to reduce temps or anything.

[Gaul]

@ ^ = Which BIOS ?

[Demo]

F3b

EDIT:

Does upping the vcore increase power provided to the IMC or make it stable at higher multipliers or thats plainly controlled by the vDIMM ?

[Hondacity]

its independent...

[Falkentyne]

NOT exactly.
I found out the difference between BSOD 101 and 124.
BSOD 101 only happens if one of the processors crashes at high load due to not enough vcore or too high temps. I've NEVER seen a BSOD 101 at idle or light load--ever---its always 124.

Likewise, I've never seen a BSOD 124 at max load...(like prime 8 threads), always seems to be 101.

BSOD 124 happens if the processor is at medium, light load or idle or unable to function at that voltage at light or medium load. Since 0124 can also occur if using 4 dimms at low VTT, I think it happens if the CPU crashes while communcating with the IMC itself, while 0101 is purely an instruction/cache error.

[Demo]

NOT exactly.
I found out the difference between BSOD 101 and 124.
BSOD 101 only happens if one of the processors crashes at high load due to not enough vcore or too high temps. I've NEVER seen a BSOD 101 at idle or light load--ever---its always 124.

Likewise, I've never seen a BSOD 124 at max load...(like prime 8 threads), always seems to be 101.

BSOD 124 happens if the processor is at medium, light load or idle or unable to function at that voltage at light or medium load. Since 0124 can also occur if using 4 dimms at low VTT, I think it happens if the CPU crashes while communcating with the IMC itself, while 0101 is purely an instruction/cache error.

I've actually experienced 101 and 124 at high loads BUT its almost always 124 on my end, so probably on this scenario both Vcore and VTT lacks!.. too high temps would be exactly how much under primeblend.. 75c and above?

I've only experienced 124 at Idle.