**Gigabyte GA-X58A-UD7 OC-Tips-Bios-Discussion Thread**

[Sam_oslo]

^^ The VVT would dependent on your CPU-batch, but generally, you need more VVT for x20 Uncore which is needed for x10 RAM. Try 1.32v.

Which reminds med about the locked x20 Uncore on my VVT-demanding X5650 that can push the VVT to "unsafe", very fast. Many 45nm i7 has survived higher VVT without damage, but we don't know much about these new 32nm CPU yet.

Anybody have a DMM-reading of VVT on this MB? Is the real-VVT higher or lower than the BIOS-set?

[Lsdmeasap]

Anybody have a DMM-reading of VVT on this MB? Is the real-VVT higher or lower than the BIOS-set?

Looks to be a little less according to
http://www.hardwarecanucks.com/forum...review-19.html

[Sam_oslo]

Looks to be a little less according to
http://www.hardwarecanucks.com/forum...review-19.html

Thx for the link

That's almost 0.02v drop in real-VVT on load, at least around my actual voltages ~1.33v.
Intel recommends Max 1.4v for new 32nm CPUs, but I've been using VVT=1.32v for keeping it on the safe side, just in case. At least until we get to know these new CPUs better. These drops makes me braver for pushing the VVT up to 1.34v.

[Lsdmeasap]

Well we all know what us overclockers do with Intel Recommended amounts, it's all up to you.

Some memory XMP settings are 1.5+, so I think you are fine. If you would like a link to where I go more in depth on Vtt and XMP programmed amounts let me know and I will link you, but I think far above the recommended is fine. Of course you are right though, most of that would pertain to the 45nm i7's and memory made back then for those.

[Sam_oslo]

Yeah, those i7 45nm are almost indestructible. They take a lot of torture, including unsafe VVT. I hope these new 32nm are the same, and we need some time to get know these. But until then, I don't want to risk degrading or killing this CPU, at least not before I've tried it on a double socket MB.

But if I know myself right, I won't be able to hold out, and will try to push it to the limits anyways, soon or later, but rather sooner.

[Lsdmeasap]

I know what you mean, I wouldn't either especially if you paid $1000 for a CPU.

Some ES X980 and ES Xeon users should be able to tell us soon though, or shortly we will find out from the new cheaper Xeons coming out.

[Fritz_the_germ]

^^ The VVT would dependent on your CPU-batch, but generally, you need more VVT for x20 Uncore which is needed for x10 RAM. Try 1.32v.

Which reminds med about the locked x20 Uncore on my VVT-demanding X5650 that can push the VVT to "unsafe", very fast. Many 45nm i7 has survived higher VVT without damage, but we don't know much about these new 32nm CPU yet.

Anybody have a DMM-reading of VVT on this MB? Is the real-VVT higher or lower than the BIOS-set?

I checked the "hidden" voltages out with EasyTune. Do you think the voltages are o.k. for the 175*23@2100 RAm-speed settings?

[Sam_oslo]

I checked the "hidden" voltages out with EasyTune. Do you think the voltages are o.k. for the 175*23@2100 RAm-speed settings?

http://www.abload.de/img/easytuneqam2.png[/img][/url]

To be honest, I don't think VVT=1.54v is a good idea for 24/7.

i7 has been good at taking torture, but one shouldn't take chances by pushing (and rubbing) it over Intel's recommendations for 24/7, unless one can get something really extraordinary. Besides, that much VVT will put a lot of extra heat in your CPU and VRM. You can use the saved heat to get higher CPU OC with the same cooling.

That x20 Uncore for x10 RAM needs extra VVT, I would suggest rather using x16 Uncore and x8 RAM with tight timings and ~1.32v VVT. You won't notice any different between @2100Mhz CL9 and @1700 CL7 RAM anyways.

I would suggest using less on IHO Core too, something around 1.2v. Otherwise, it looks good to me, but every system is different and you have to try it yourself.

[Fritz_the_germ]

Thank you for your recommandation. I reduced the QPI/VTT to 1,34V.
If I don´t have enough I will get freeze (prime95 in-large FFTs) ,right?
I am kind of a newbe

[Sam_oslo]

Yeah, run some memtest, prime95 large FFTs, HyperPi32M, these will test the IMC, Umcore and RAM, and would react to low VVT too.

[newhit]

To be honest, I don't think VVT=1.54v is a good idea for 24/7.

i7 has been good at taking torture, but one shouldn't take chances by pushing (and rubbing) it over Intel's recommendations for 24/7, unless one can get something really extraordinary. Besides, that much VVT will put a lot of extra heat in your CPU and VRM. You can use the saved heat to get higher CPU OC with the same cooling.

That x20 Uncore for x10 RAM needs extra VVT, I would suggest rather using x16 Uncore and x8 RAM with tight timings and ~1.32v VVT. You won't notice any different between @2100Mhz CL9 and @1700 CL7 RAM anyways.

I would suggest using less on IHO Core too, something around 1.2v. Otherwise, it looks good to me, but every system is different and you have to try it yourself.

Excellent advice Sam. I've got G.Skill 2000 ram and had to jack up Vtt to 1.49v to make it run as advertised. Maybe I should rethink

[Fritz_the_germ]

O.K.
Can I reduce QPI/VTT if set 160*25 for a goal of 4 GHz?
Even a VTT of 1,54v couldn´t prevent a BSOD running prime95 in-place-large FFTs.
I really want to use my Xeon a little longer...

[Oblique82]

Guy I am having a lot of reboot issues with this board. eg at bclk 190 and above (and maybe less havent tested) when rebooting or cold booting about 1 time in 20 it will fail.

you will hear a click several times during the initial power on and then it will boot fine usually, but sometimes it will "click" AFTER the post beep, and then fans will power off and it will give the red overclocking failed warning.

this is on settings that are 24/7 prime and linx stable

have latest bios and have tried all differernt voltage under and over combinations.

I need a board that ALWAYS boots at its current settings.

has anyone encountered this? Im thinking of RMA

[dinos22]

what PSU? they can be a culprit here too

whats the rest of system spec

where does the click come from?

[rezax58]

Anyone else getting sick of the electric whine they get when they scroll down and up on a page? It stopped for a few days and it's back on.

[Oblique82]

the click is from the motherboard "c1" code- when the LED's turn on.

corsair hx 1000w
920 d0
6gb gskill 2000mhz rated

have tried all differnt safe combinations of voltages.

basically the problem occurs simply because the bclk is high- even if I turn down the cpu multiplier and ram divider to underclocking, but keep bclock moderately high, and have generous safe voltages for everything it still occurs.

the lower the bclock the fewer "fail" clicks from the mobo. but yeh - the main issue is sometimes it will turn on for about 5 seconds then off(all fans pwoered down) then boot from backup bios with warning/error

I note that the green overvolt LED for the ?northbridge (the one nearest the PCI slot) is always lit even in "failsafe values" mode, and the machine will not post from cold boot if IOH >1.18

its giving me a great deal of stress because just when I think its happy it fails to boot, despite being prime stable for 24 hrs +

[Intratech]

Anyone else getting sick of the electric whine they get when they scroll down and up on a page? It stopped for a few days and it's back on.

Haven't heard any.

[Musho]

Anyone else getting sick of the electric whine they get when they scroll down and up on a page? It stopped for a few days and it's back on.

Disable C1E, Intel Speedstep/EIST and the whine should go away. Let us know if it works out for you!

[rezax58]

Disable C1E, Intel Speedstep/EIST and the whine should go away. Let us know if it works out for you!

WOW thanks alot dude. It's goneeeeeeeeeeee!!


One more thing, the led on the top left side of the ram is on amber and green, also one green led is lit right by the water cool block, This only happens when I manually set the timing, speed and voltage of my memory according to OEM spec; when it's on fail safe mode in bios the leds are all off.

[M@rcB]

I am looking for the latest BIOS and found the Gigabyte GA-X58A-UD7 F6a -08.Mar 10 bios version. However, I am confused now. Does someone know if the F6a is the latest or is the official F6 the latest BIOS version?

Another question, I would like to upgrade the Marvell 9128 to the latest firmware version on the same board. Any idea how to?

[tipes]

Hi all,

Feeling a bit peeved here - let me explain. Just recently bought cosair dominator GT 2000 triple kit thinking it would be great to run a straight 200x20 for 4Ghz with turbo disabled with a mem multi of 10 and uncore of 20/21 but no dice. Allkinds of blue screens etc have tried quite a few things to get it to work but no dice. It will run on 8 multi - no problem. Why is it seemingly impossible to run a 10 memory multi or am I missing something. Seems a waste to run the memory at 1600 or less.

Any suggestions from you knowleadgeable peeps would be appreciated. Anybody out there managed to do this???

Mark

[PalmeAsus]

Hi all,

Feeling a bit peeved here - let me explain. Just recently bought cosair dominator GT 2000 triple kit thinking it would be great to run a straight 200x20 for 4Ghz with turbo disabled with a mem multi of 10 and uncore of 20/21 but no dice. Allkinds of blue screens etc have tried quite a few things to get it to work but no dice. It will run on 8 multi - no problem. Why is it seemingly impossible to run a 10 memory multi or am I missing something. Seems a waste to run the memory at 1600 or less.

Any suggestions from you knowleadgeable peeps would be appreciated. Anybody out there managed to do this???

Mark

Almost all cpus need a uncore / qpi voltage at ~1.5v-1.6v to run ram at 2000mhz. So question is have you tried to rise qpi/vtt voltage ?

[tipes]

Almost all cpus need a uncore / qpi voltage at ~1.5v-1.6v to run ram at 2000mhz. So question is have you tried to rise qpi/vtt voltage ?

I have not dared to raise to the figures you are suggesting as I would not think it would be good to run at such high volts 24x7. See Sams post above. I think I will try up to 140 ish and if no dice - well i have wasted my money on the ram - unless any other suggestions????

thanks for your reply

Mark

[PalmeAsus]

I have not dared to raise to the figures you are suggesting as I would not think it would be good to run at such high volts 24x7. See Sams post above. I think I will try up to 140 ish and if no dice - well i have wasted my money on the ram - unless any other suggestions????

thanks for your reply

Mark

Well , for the 45 nm i7s i can't speak for everyone but i have had my vtt on 1.53v to be able to set uncore as high as i have it, and i haven't had any degradations or problems so far after ~7 months and i know for sure alot other users use 1.5-1.6v vtt to get their ram stable at 2000mhz 24/7 without problems but ofc it's all up to you =).

[tipes]

Well , for the 45 nm i7s i can't speak for everyone but i have had my vtt on 1.53v to be able to set uncore as high as i have it, and i haven't had any degradations or problems so far after ~7 months and i know for sure alot other users use 1.5-1.6v vtt to get their ram stable at 2000mhz 24/7 without problems but ofc it's all up to you =).

Many thanks for your further input - suppose every chip is different so maybe I may try going to these high vtt and see what happens. Do you find the gains of ram at 2000 worth it????

Pity the 920 DO does not have unlocked multi otherwise you could probably do baseclock of 250 x 16 multi with mem multi at 8 giving the 2000 ram - mind you don't know whether you could attain a bclk of 250???