Asrock P67 Extreme4 UEFI & Drivers thread [Extreme6 & Pro3 owners welcome]

[BDM123]

Ket, great info on the ASRock boards. I've followed your threads for some time now. I built a second SB system not long ago and ran with the ASRock (it's your fault). Great deal and a solid performer! I've had no issues with either of my systems. I'm not running on the bleeding edge as I have no interest to try. Just rock solid 24/7 365/day performers. I'm looking forward to maybe purchasing the Extreme7 when available......That is one sexy board

Brian

[Ket]

I'll take that as a compliment

[BranFB]

My board is back from RMA.
update to 1.9bios through Instant Flash. I cannot fine the DOS bios in USB bootable flashdrive :S Will try it soon again. If anyone has any suggestions !?

Clock is the same even worse in aspect of vCore. 4.6GHz 2600K from 1.352 to 1.36 BUT Offset from +0.065 down to +0.060 - that's strange!!!

I'm waiting on gen3 mobos. Look nice. if not will go to z68!

[trees415]

My board is back from RMA.
update to 1.9bios through Instant Flash. I cannot fine the DOS bios in USB bootable flashdrive :S Will try it soon again. If anyone has any suggestions !?

Clock is the same even worse in aspect of vCore. 4.6GHz 2600K from 1.352 to 1.36 BUT Offset from +0.065 down to +0.060 - that's strange!!!

I'm waiting on gen3 mobos. Look nice. if not will go to z68!

That's not exactly out of the ordinary, particularly with the new BIOS, which I must admit has been quite confusing and disappointing. In 1.9, the setting for LLC seems to have a much greater impact on the voltage measured in UEFI. In fact, I had to re-do all my settings for 1.9, and I couldn't get find a nice sweet spot that I felt comfortable with, so I went back to Version 1.8, so I'd suggest trying that. But again, like I said before, that isn't out of the ordinary at all when Vcore is an offset voltage setting.

[Tunagoblin]

i5 2500k 4.8GHz @ 1.382~1.392v (load to idle, vdroop) on Air. 21.5hours stable. 1600 8-8-8-24 1T 1.515v.
Max 81c. Ambient 27c.

Great board for the price. UEFI 1.9 works just fine for me. No LLC issue I've noticed... Mines at lvl2.
But somehow turning down CPU PLL Voltage made it unstable... Not usually happening on other boards.

Took me a while but got it stable.
It gets hot so this is as much as this chip can go I think.
I have to get my RAM back to 1866 next.
Other voltage settings are on the screen shots as well.
I've done IBT and LinX many times. Nice 125~130GFlops stable runs in different size of tests.
Long hours on prime95 blend test were the most tough one to clear...

[Ket]

Its recommended only to update the UEFI if you have problems, or the general rule I stick to, update the UEFI if issues that are fixed could affect you in some way, otherwise skip the version.

[Tunagoblin]

Its recommended only to update the UEFI if you have problems, or the general rule I stick to, update the UEFI if issues that are fixed could affect you in some way, otherwise skip the version.

I'm always happy to be the test subject for new UEFI

[Ket]

What happen to your RAM? Weren't you running it 1866 8-8-8 at one point?

[Tunagoblin]

What happen to your RAM? Weren't you running it 1866 8-8-8 at one point?

Yes. (Actually it was 1866 9-9-8-24) I had to make it to stock in order to get to 4.8GHz.
(Look at my previous post with screenies.)
Now I've made it to 4.8 stable. So I have to make the RAM back to 1866.
That is my next goal. I'm pretty sure I can get there.
Anyway, I love this board. None works better for the price!

[Ket]

Ahh ok thought something might of been up with the UEFI and you couldn't run 1866 anymore. If anybody has any suggestions on UEFI features that would make the P67 Extreme4 better then by all means suggest them and I'll put them forward to R&D

[BDM123]

Ket question. I noticed with most options at default in the UEFI. My 2500K will not not go to 3.7/GHz in turbo mode (loaded). It hits about 3.4 at best. Seems I have to set the multiplier myself. Probably something I'm not aware of or normal?

Brian

[Ket]

Load UEFI defaults and see if turbo mode works correctly. Without knowing what settings you changed thats all I can suggest ATM.

[BDM123]

Ok Ket, I'll try that and let you know the results. Thanks.

Brian

[BDM123]

Ket, here is a shot with the UEFI default loaded. One other thing I've noticed. If the BCLK is set at 100, it actually operates at 99.8. If I set it to 99.9 or 100.1, it operates exactly at those settings. Nothing I'm concerned about, just strange.

P.S. Turning off spread spectrum brought the BUS to a set 100.0. But it had no effect on the turbo boost.

[Ket]

Can you flash to UEFI 1.8 and see if turbo boost works correctly then? Perhaps try putting a real world load on the system instead of a synthetic one as well, try converting a 800MB AVI file and monitor the CPU speed to see if it increases to 3.7GHz.

[swampy]

Has anyone been successful reading an IDE drive with this board, I'm just trying to save peoples data. I've tried 4 different USB/IDE controllers, no go. Could it be a UEFI setting I'm not seeing? I haven't tried a PCI-E/IDE card, I'm spending too much money, I do it for free.

[papatsonis]

are you trying to connect it to a usb3 port? if yes, try again in a usb2, usb3 may still have some issues. Also consider finding a ide-sata adapter

[swampy]

I've tried all you said, including ide-sata adapter, but thanks. I could try wasting money on a PCI-E/IDE card.

[Tunagoblin]

Hey Ket I have a question about RAM OC.
I'm still a noob about OC RAM. I want you to take a look at my settings.
Thanks a lot in advance!!


When I had 4.6Ghz (1.315v), I could make my RAM to go 1866 9-9-8-24 1T 1.545v, VTT@1.051
Now I'm at 4.8GHz (1.384v), I needed my RAM to be 1866 9-10-9-28 1T 1.650v, VTT@ 1.077
Also this MB and RAM don't like lowering CPU PLL Voltage at high clock.... so I'm at default 1.832...


Both stable with HCI memtest and LinX Max run. But does this look right to you?
I'm surprised that I needed a big jump in terms of DRAM voltage.... (anything less than 1.65 will give me unstability, errors in memtest)
Also I've been messing around with VTT(VCCIO) and I found out that 1.077 is the sweep spot.
Anything lower will make my RAM unstable, anything higher will give me a bit lower GFlops in LinX and also more heat.
I get healthy GFlops gain in LinX (130+GFlops) but going from 4.6 to 4.8 makes this much difference in terms of the settings goes...
I know I have a 2x4G so it is making the IMC stress harder than 2x2G sets. Even I'm getting healthy GFlops and stable, Is the settings look OK to you?
I might can make the timings a tad better but as far as voltage goes, I got it stable with as less voltage as I can make it I think... Unless I'm doing something wrong...

[BDM123]

Ket I'm working right now so I can't try anything at the moment. Few things I can tell you as I've noticed that since I built it three weeks ago. I received my board with version 1.6. Without any bios changes, I noticed it wouldn't turbo up to 3.7. I have tested it while playing BC2 and using LinX also. So I flashed it to 1.8 soon after it was released. I again noted the same issue. Now I can set the multiplier to whatever I want with no other change and it will go to where I set it. So during lunch I called a friend with the same board. Only difference is he has a 2600K. He reset his bios 1.8v back to default. He tried LinX and BC2 also. It would only turbo up to 3.5. Notice both chips only turbo up about 100/MHz from their stock speed. It's not a big deal as I can obviously OC my system without issue and at low vcore to boot. But it's something I've wondered. Compared to my P8P67 (with 2600K) board, it auto turbos to 3.8 in stock form under any load I give it.

[Ket]

Hey Ket I have a question about RAM OC.
I'm still a noob about OC RAM. I want you to take a look at my settings.
Thanks a lot in advance!!


When I had 4.6Ghz (1.315v), I could make my RAM to go 1866 9-9-8-24 1T 1.545v, VTT@1.051
Now I'm at 4.8GHz (1.384v), I needed my RAM to be 1866 9-10-9-28 1T 1.650v, VTT@ 1.077
Also this MB and RAM don't like lowering CPU PLL Voltage at high clock.... so I'm at default 1.832...


Both stable with HCI memtest and LinX Max run. But does this look right to you?
I'm surprised that I needed a big jump in terms of DRAM voltage.... (anything less than 1.65 will give me unstability, errors in memtest)
Also I've been messing around with VTT(VCCIO) and I found out that 1.077 is the sweep spot.
Anything lower will make my RAM unstable, anything higher will give me a bit lower GFlops in LinX and also more heat.
I get healthy GFlops gain in LinX (130+GFlops) but going from 4.6 to 4.8 makes this much difference in terms of the settings goes...
I know I have a 2x4G so it is making the IMC stress harder than 2x2G sets. Even I'm getting healthy GFlops and stable, Is the settings look OK to you?
I might can make the timings a tad better but as far as voltage goes, I got it stable with as less voltage as I can make it I think... Unless I'm doing something wrong...

The higher your CPU clock the more VTT / vDIMM you may need as the IMC (Integrated Memory Controller) is on-die the CPU so as you increase clocks you are also putting additional demands on other areas of the CPU architecture. For 4.8GHz I'd say you are doing very well for such low voltages.

Ket I'm working right now so I can't try anything at the moment. Few things I can tell you as I've noticed that since I built it three weeks ago. I received my board with version 1.6. Without any bios changes, I noticed it wouldn't turbo up to 3.7. I have tested it while playing BC2 and using LinX also. So I flashed it to 1.8 soon after it was released. I again noted the same issue. Now I can set the multiplier to whatever I want with no other change and it will go to where I set it. So during lunch I called a friend with the same board. Only difference is he has a 2600K. He reset his bios 1.8v back to default. He tried LinX and BC2 also. It would only turbo up to 3.5. Notice both chips only turbo up about 100/MHz from their stock speed. It's not a big deal as I can obviously OC my system without issue and at low vcore to boot. But it's something I've wondered. Compared to my P8P67 (with 2600K) board, it auto turbos to 3.8 in stock form under any load I give it.

I never noticed turbo not working correctly, before I bring the issue up with my contacts can you try disabling the temperature throttling options and increase the power draw durations to make sure your CPU isn't throttling itself.

[Tunagoblin]

The higher your CPU clock the more VTT / vDIMM you may need as the IMC (Integrated Memory Controller) is on-die the CPU so as you increase clocks you are also putting additional demands on other areas of the CPU architecture. For 4.8GHz I'd say you are doing very well for such low voltages.

Thanks! I wasn't sure 0.2GHz makes this much difference to the other parts.
I guess it reached its limit, though...

[BDM123]

Ket, loaded defaults. Disabled thermal protection, but left everything else stock. Ran Prime, LinX (half hour each), played a couple of games of BC2 and it never clocked up above 3.4/GHz. I noticed the multiplier never climbs above x34. So I called my bud and he noticed the same thing. Though his multiplier never climbs above 35 (remember, same board but he's using a 2600K and UEFI 1.80). Has anyone else noticed this? I know a third guy who has an Extreme4 with a 2500K. But I haven't asked him to test his system yet. If you'd like screen shots I can take them. Or anything else you may require. Again I'm not really bugged by it. But I'm sure that's not how it's intended to operate when at UEFI default settings. If you're not looking, I'd bet you'd never notice. The SB system runs so well as it is

Brian

[Ket]

Can you increase short and long power duration limits to 150 and increase "Long Duration Maintained" to 5? If I'm remembering correctly how the new intel turbo boost works it fluctuates up and down all the time based on load and the maximum turbo speed (in your case, 3.7GHz) will never be maintained longer than 1 second with default power duration and long duration maintained settings.

[Tunagoblin]

I knew I couldn't stop OC...
5Ghz on first test. Prime blend 1 hr. (I know it's short but it's my first test.)
1.470v (1.464 load) It is way too HOT right now I know. But I found a way to make it 85c max late last night.
So I'll try that test tonight when I get home. I'll get there without reaching 90c....