***X58A-OC*** Club/Discussion/Information

[XII]

Did another run using the Hypers. 217X19 is max stable, mem 7-8-7-21
What the hell

[donmarkoni]

I'm having... hmm... experience with this board...

It is very good so far, seem like CPU's need less VTT and vDIMM to run Elpida Hypers at 4000/2000/7-7-7-20-1T.

But, it also killed my favorite CPU OK, it might have died anyway, but it's like borrowing a car to a friend and it breaks while he was driving. You will not hold him responsible, but it just leaves a bad taste in your mouth.

Booted at 5GHz at 1.55V. Set vCore at 1.6V, pressed Multi+ once and made CPU-Z s/s at 5.2GHz. Restarted.
Booted at 5GHz. Then set vCore at 1.645V (my 990X ran benchmarks many times at under 1.65V), then I pressed Multi+ button twice. Took CPU-Z s/s at 5.4GHz.
Rebooted. Noticed that vCore is 1.7V. Set vCore again at 1.645 and noticed that it does jump at almost 1.7V (checked with DMM). Rebooted.

CPU worked next few hours, while I was trying to make it run at 4.2/4.0/2.0@7-7-7-20-1T unsuccessfully, then it just stopped at C1. Dead.

[CryptiK]

Suspect/self changing vcore after BIOS set value is concerning.......... sorry to hear about the CPU loss. You were pushing it but that kind of set one thing changes itself to another higher value is just not good enough.

[Helios1234]

That's unfortunate. I have a 990x myself and would love to get a hold of the OC board. I'm hoping a friend can hook me up soon.

[donmarkoni]

Here it is... 990X... Preparing for the last ride...



Nothing special... Never got to test it much...

[dinos22]

it does NOT self select vcore. What happens at about 1.65 on ALL X58 gigabyte boards is vcore is set a bit higher than value selected in bios which then settles as you go up but it sets a bit higher. donmarkoni played with fire with such high vcore on normal cooling, got burnt and now says fire is bad lol, i agree with him don't play with fire!

[donmarkoni]

Yeah, fire is hot. I got reminded.

[stasio]

BIOS F5c is out.

[maxxx]

BIOS F5c is out.

Statio, any release notes for this?

Edit: Found it...

-Improve OCZ PCIe SSD compatibility and RAID ROM
-06.Sep 11

[XII]

Tested F5c. Stable but still lacking 24Gb performance. Tested 2 flavours, but no avail.
Hypers run great, and even uncore seems to hold up icm. Hypers.

If it only could be a little better than this:

[zoson]

I thought about getting one of these to replace my R3E. Then I read this thread. It's a bunch of headache. Dead CPU's, board sets vcore WRONG, low large memory size performance, difficulty getting above 4ghz... I think I'll pass. It also seems like the report of super high bclk is wrong... None of you seem to be going very high...

[XII]

I thought about getting one of these to replace my R3E. Then I read this thread. It's a bunch of headache. Dead CPU's, board sets vcore WRONG, low large memory size performance, difficulty getting above 4ghz... I think I'll pass. It also seems like the report of super high bclk is wrong... None of you seem to be going very high...

Don`t get me wrong, the board does things my R3E can`t do.
Hypers icm. 980X or my W3680`s clock way better, @ same settings and voltage.
6x Hypers also clock better, at the same, or lower voltage.

Going to 24Gb I have to give in more than I like:
Uncore: x2+1(Rex3) vs x1,5(OC)
Vqpi : 1,4v(Rex3) vs 1,51v(OC)

If you put in the same perspective, and more important, MY WISHES: UD7v1 (Almost on par with Rex3, great 24gigs), Classy(great quad-Hyper lovin`, great 12gigs Hyper), Rex3Gene(nifty little allrounder), baby Classy(stability above all) there are no clear winners.

On air, all my core`s are stable up to ~4,4Ghz. Most of them need more voltage, than I`m prepared to give. Max Tcore for me is around 80C.
Uncore should be x2+1. Memory has to be 24Gb.
In this combo Rex3 beats X58A-OC.

Saying the X58A-OC isn`t a good board, is like saying a Lamborghini Diablo sucks.
It sucks, when you`re stuck in the middle of the Sahara, in a Lamborghini Diablo.

[zoson]

I didn't say it sucks or is a bad board... I said this thread makes it seem like a bunch of headache, and it seems like the initial report of high bclk is wrong.
For 24/7 I'm running 4.4GHz with 18GB Mushkin with Elpida BBSE @ 1834MHz with x2 uncore(3668). With what you said above, you've confirmed that this would not be 'better' for me. Maybe I'll get a few more mhz on core, but my uncore would be even lower...

[XII]

In that case, you`re probably right.

[Silverlight]

I need some help clocking with this board, seem to have hit the 222 bclk wall

http://valid.canardpc.com/show_oc.php?id=2145159

Tried;

VTT - up to 1.38v
IOH - up to 1.36v
QPI is in slow mode
PCIE - 115

Any idea what the issue could be? Going from 110 to 115 PCIE allowed me to boot at 220 bclk (110 PCIE - 215 bclk boot but it still froze at exactly 222, increasing bclk with the onboard buttons) so more PCIE will give more bclk?

Can only test higher PCIE later this week, don't want to damage my 6850.

[Hondacity]

some of those gainestowns can't clock high bclk. we had the same limit like you. we were even at -60c

the board works @ 250bclk with a good 920, some cpus just don't have the ooomph

[sin0822]

yup 250 but with slow mode under LN2 http://valid.canardpc.com/show_oc.php?id=1809457 with a 930

Try more qpi/vtt to 1.5v and some QPI PLL to 1.3v, but yu don't want to run 24/7 at those volts. you can even try higher qpi/vtt

more PCI-E can give more BLCK, but that seems like enough already

[Silverlight]

That sucks, thought these Xeons would clock higher than the consumer parts. Swapped the chip with another E5540 and that hit a wall even earlier, 218 bclk :/

Tried 1.5v QPI and 1.3 QPI PLL on the 222 chip and it was stable at 223.1 bclk for a few seconds then froze.

Guess I'll have to look for a Bloomfield. Glad it's not the board, Googled "222 bclk wall" and the results that came up scared me.

[sin0822]

well you had slow mode on right? My 930 there had a wall at 222 but under LN2 is went free. BLCk just isn't something you are going to get at 225+ stable working 24/7 the voltages required are usually enough to hurt. I just haven't seen it done much i guess. You are just looking for max OC? Put the system under the AC lol. Good luck mate, i am sure you will get it at some point. BTW you are using a lot of volts, but for suicide youc an try 1.55v qpi/vtt, some even have used 1.6v. i stay under 1.6v and that is only under LN2. Cold is what you need mate.

[Silverlight]

Yep, it was on slow mode. Tried swapping the RAM for 1x stick of Hypers and that didn't help (Tried x6 and x8 mem multi)

I will put it under DICE at the end of the week and report back with regards to how it goes, maybe I can prove Honda wrong and get it going at -60 hehe .

I'm not going for a 24/7 OC on this (Have another setup I use). Got the X58A-OC to clock a few 1366 chips, starting with the Xeons.

Thanks for the help, much appreciated.

[acebmxer]

After stugaling to get 225 bclk stable here we go. Here are some pics,

LinX stable



Voltage settings
1.525 vcore set in bios via offset. 1.25+0.275

[Hondacity]

Yep, it was on slow mode. Tried swapping the RAM for 1x stick of Hypers and that didn't help (Tried x6 and x8 mem multi)

I will put it under DICE at the end of the week and report back with regards to how it goes, maybe I can prove Honda wrong and get it going at -60 hehe .

I'm not going for a 24/7 OC on this (Have another setup I use). Got the X58A-OC to clock a few 1366 chips, starting with the Xeons.

Thanks for the help, much appreciated.

there is a reason why they disabled 2 cores. sh|ty chips

[Silverlight]

Didn't know they were leftovers .

Will post subzero OC results tomorrow.

How do the 32nm Xeons compare to the E5500s, Bloomfields , 970s? Like the X5650 and E5620?

[Silverlight]

http://valid.canardpc.com/show_oc.php?id=2148403



I ran out of DICE (Wasted on a stupid P5Q Pro).

Is it normal for the board to reboot multiple times when you change something in BIOS? Whenever I changed something, at "Verifying DMI" screen it would display another line saying something to the effect of "Detecting DRAM size" and then switch off, this will happen 2-4 times then the detecting message would disappear and it would boot normally.

[zoson]

It's the memory recheck feature, to make sure all dimm's are detected correctly.