Gigabyte EX58-UD5/Extreme Discussion Thread

[M.Fixit]

@Mr.Fixit, you will need to run your mem multi @ 6x if you want to use anything lower than 16x on Uncore, as the rule is that Uncore should be twice that of the RAM multi, you can set it higher than double, but if you trying going lower it results in BSOD.

I have found that with my UD5's, I cannot use the 12x Uncore multi when using a 6x mem multi, that refuses to POST for me altogether, when using 6x on mem, I have to use 13x uncore for my system to POST.

As for WC, I am going to be purchasing parts to setup a custom loop for my rig, I will be using the HK 3.0 as well after seeing a few builds where people are using them, as the temps compared to air are fantastic.

I seen someone on another forum with temps of 61c full load @ 4GHz, that is amazing compared to air imo.

Thanks, I am already running with the x6 mem multiplier. I too tried x12 for Uncore with no Post, hence why I set it at 16. But I will try 13 as you suggest, although I am not sure whether this will stop the alarm going off.

I look forward to seeing your before and after w/cooling temps. I assume you will need to use a triple rad in the loop.

EDIT:

OK, so I have just tried the 13x Uncore and it posts OK, but again after about 45 secs of LinX I get an alarm souding. So, what can this be due to? It's not overheating or throttling (around 74 deg C). Any more ideas to try anyone? BTW - LinX will keep running OK, so it is just the alarm problem I can't live with - it's way too loud!

[orig3n]

@Mr.Fixit, you will need to run your mem multi @ 6x if you want to use anything lower than 16x on Uncore, as the rule is that Uncore should be twice that of the RAM multi, you can set it higher than double, but if you trying going lower it results in BSOD.

I have found that with my UD5's, I cannot use the 12x Uncore multi when using a 6x mem multi, that refuses to POST for me altogether, when using 6x on mem, I have to use 13x uncore for my system to POST.

As for WC, I am going to be purchasing parts to setup a custom loop for my rig, I will be using the HK 3.0 as well after seeing a few builds where people are using them, as the temps compared to air are fantastic.

I seen someone on another forum with temps of 61c full load @ 4GHz, that is amazing compared to air imo.

depends on many factors, mine still hits 80c with the HK v3 and a TFC 360 radiator when im pushing 1.35-1.4v @ 4400mhz or higher, though it might have helped if i could mount the HK the right way across the cores but the caps next to the socket on this board stop that...

[jalyst]

I won't be going to Tiawan, I meant my Contact at Gigabyte will be back in the office after the 31st. So I would ask him then, he is pretty fast at getting back to me and the last he told me (in June) was that they would be released at the same time as P55 so they could release all the new Smart 6 features at once.

Of course things may have changed since then as we know, but I will find out from him once he is back next week.

Sorry, I don't know anyone at Asus though, Just Gigabyte

Okay very keen to know...
So please let us all know once you hear from him.

Thanks!

[123456iambelial]

I made it. All I did was pump the vcore a little and fixed the problem.
I was fooled by the BSoD that I was getting that was why I thought I did not need to add more vcore.

[123456iambelial]

WTH!!! I can pass Linx now, but I could not pass Prime95 blend.
The BSoD I got was 0x0000003B. I think that error is for display driver's bug or something.
I am going to reinstall the display driver and re-run P95 blend.

Do you guys have any input about this?

[Jonny Manc]

WTH!!! I can pass Linx now, but I could not pass Prime95 blend.
The BSoD I got was 0x0000003B. I think that error is for display driver's bug or something.
I am going to reinstall the display driver and re-run P95 blend.

Do you guys have any input about this?

I had BSOD's with the new WHQL 190.6x drivers from nv.....did my head in!!
Went back to my old faithful 190.15 and BSOD's gone

[cliffy]

Stupid question maybe but can anyone tell me if it is possible to select the 21 multiplier???

[ReverendMaynard]

Stupid question maybe but can anyone tell me if it is possible to select the 21 multiplier???

Nope, but you can get the 21x multi by disabling some of the power saving options (I disabled all) and enabling Intel Turbo Mode.

[Jonny Manc]

Apparently, it is possible to FORCE the x21 multi....I'll see if I can find the post over on G3D

[ReverendMaynard]

Apparently, it is possible to FORCE the x21 multi....I'll see if I can find the post over on G3D

With this board ( I owned the Extreme...same ), you can only select 20x multi then add the extra with Turbo Mode.

[Supershanks]

yep you can do it on any x58 as long as it's running reasonably cool.
see Turbo mode

[cliffy]

Great Avatar Supershanks, Disappointing start to the season and an even more disappointing transfer window so far. Y.N.W.A

[Supershanks]

Great Avatar Supershanks, Disappointing start to the season and an even more disappointing transfer window so far

I know don't even go there Alonso is a great loss.

[DavyBoy]

Got my replacement chip today, and I am up and running again @ 4GHz with normal temps!!

Idle temps are 36c, and full load hits about 79c.. although my system still isn't inside the case yet, so I assume these temps will rise a little bit once it is all back inside the case.

I am just happy that I can get 4GHz stable again!

I was supposed to get my 2nd chip today that I ordered for my Fiancees build, but OcUK sent it out with 2 day delivery for some reason... even though it says on their site that I was supposed to get next day delivery..

It will be here tomorrow though, hopefully it is a goodun as well

::edit::

Temps will rise a little when the rig is build up inside the case right?

Or will they go down because of my 5 x 120mm case fans blowing air across everything?

[orig3n]

Got my replacement chip today, and I am up and running again @ 4GHz with normal temps!!

Idle temps are 36c, and full load hits about 79c.. although my system still isn't inside the case yet, so I assume these temps will rise a little bit once it is all back inside the case.

I am just happy that I can get 4GHz stable again!

I was supposed to get my 2nd chip today that I ordered for my Fiancees build, but OcUK sent it out with 2 day delivery for some reason... even though it says on their site that I was supposed to get next day delivery..

It will be here tomorrow though, hopefully it is a goodun as well

::edit::

Temps will rise a little when the rig is build up inside the case right?

Or will they go down because of my 5 x 120mm case fans blowing air across everything?

They will usually go up because your ambient temp will be higher.

[DavyBoy]

They will usually go up because your ambient temp will be higher.

That's what I thought, but I just wanted to be sure.

Thanks mate

[phsinc1]

Nice Blunt, what is the CPU batch and what voltage did you use for 4GHz?

[DavyBoy]

I haven't really tried to tinker with my voltages to see how low I can get them, they are currently at 1.2815v in the BIOS (1.264v in CPUZ), with QPI/Vtt @ 1.285v, and I have tightened up my RAM timings to 8-7-7-21-1t as well.

I've been trying to get it stable at 4.2GHz today, but it isn't happening for me, I have managed to get it to stop BSOD'ing in LinX and now it only stops because of errors, so I am getting somewhere at least, although I think I am just going to keep it at 4GHz until I get the chip under water.

I will be building a custom loop in the next few months, I am just buying up my parts slowly.

I've managed to get a 2nd hand EK GTX280 block and backplate from OcUK Members Market for £35, so I am well happy about that.

And I'm sending my TJ10 case off to ChilledPC to have them mod the roof so I can fit a 120.3 radiator in there, and they will be powder coating the internals red for me also, as well as putting some holes in the mobo tray for routing cables to.

I should hopefully have all this work done before the end of the year, I can't wait!!!

Still waiting on my second chip to be delivered, I thought it would have been here by now...

[Pyr0]

hi blunt,
just wondering... what version of dominator mem do you have?
the errors in linx might be due to memory
i had a kit of the ver2.1 dominator mem which wasn't very flexible at all. best i could get out of them was 7-8-7-21 1T

[DavyBoy]

Hey Pyr0, my Dominators are v3.1..

Like I said, I didn't really play around with it enough to try and get 4.2GHz stable, I think if I did mess around with them and tried to loosen up RAM timings I might have managed to get it stable.

When I first tried it, I kept getting 124 BSOD's, but raising QPI/Vtt stopped the BSOD's, and LinX would just stop due to errors..

This is the same thing that happened when I first tried for 4GHz, and I got that sorted out by raising QPI/Vtt a little bit more.

But tbh, I am just going to leave it at 4GHz until I get this chip under water.

[Pyr0]

yeah i just mentioned it cos i was playing with my mem last night trying to find out what voltages it needed for 1900 stable(ish)
i also stopped the 124 bsods with more qpi vtt, but then linx was just giving errors until i raised mem voltage

[DavyBoy]

You have your dominators running @ 1900Mhz?

What timings are you using for this?

And what do you have your CPU clocked to?

My Dominators struggle to do anything more than their rated frequency, I am starting to hate this RAM actually, I think I will buy another brand of memory soon to replace these ones.. When I bought them I thought that I would be able to O/C them by a couple 100MHz and for them to run without any issues, but that just isn't the case for me unfortunately..

Even at 1688MHz they give me BSOD's all over the place.. though in saying that, I haven't really tried using more VDimm than 1.65v, as I am afraid of damaging my processors IMC.

Which RAM would you say is the best out there atm for overclocking?

I have heard that the Patriot Viper 1600MHz is supposed to be good from a few people, they aren't to badly priced neither, and the color of their heatsinks would go nicely with the UD5 color scheme as well!

[Pyr0]

tbh i haven't tried any other sets, so cannot really recommend anything
i still have to check full stability with this kit, as i've only been testing by running 25 loops of LinX for now, but so far managed
1600 7-7-7-21 1.70V
1600 8-8-8-24 1.56V
1800 8-8-8-24 1.68V
1800 9-9-9-24 1.60V
1860 9-9-9-27 1.62V
1900 9-9-9-27 1.68V

at this moment i just have them running at 1600 7-7-7-21 1T (200 bclk) because it only needs a low qpi/vtt

while testing i just kept my cpu at or under 3.8Ghz cos its a crappy C0 chip, sometime i'd like to try a D0 i920

[DavyBoy]

Thanks for posting your timings for these clocks mate, I will give this a go later on today and see if I can get my modules stable, I never used anything higher than 1.65v VDimm because of the Intel specification, but I have read a lot of posts on here where people are running 1.8 to 2v VDimm without any issues.

Do you know why Intel have said that 1.65 is the limit?

Having them running at their rated speed with only 1.58v is not bad actually.

What did you have the advanced timings set to when testing these clocks, did you just leave them on auto?

I always set my advanced timings to the same values as the XMP profile sets them to, I haven't actually tried playing around with these yet.. I just have mine set to 8-7-7-21-1T atm, and it is stable with the CPU at 4GHz, I let it run 20 passes of LinX using all available RAM, and I left it running Prime95 Blend Test overnight to, and it was still going when I got up this morning, so I would say that this is stable.

[esdxx]

RAMs are giving me stability issues too in new LinX (i think it is RAMs atleast). I'v got OCZ Platinum tripple channel - http://www.ocztechnology.com/product...triple_channel

I'v tried to lose the timings to 8-8-8-24 and bupmed the VTT to 1.35V and VDIMM to 1.64V, still cannot pass LinX (usually BSOD after 2 passes). My multiplier is set to 6.0 and uncore to x16. I am at 4GHz btw (190x21).