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

[banwell]

I've got a 2600K with the Extreme4 board and I seem to be having some issues that are similar to what I'm hearing other people say here.

Having read all of this thread I cannot see what 'these' issues are. Many people (myself included) are happily running decent overclocks without issue - call us lucky winners in the CPU lottery if you like. If you don't believe me (or others here) - then go read Anandtech's review of the Ex4. They got along together just fine. And I am sure they did a much better job of putting the board and a CPU through its paces. But even they stopped at 4.8 (and I am sure they have more technical know how than most).

Yes - some people may need to RMA a board. It happens.

Not being able to get stable at some pre-defined idea of 'decent' overclock at whatever cost (including trashing your gear) is 'by the by'.

No CPU has a god given right to hit 4.6 / 4.8 or 5.0Mhz+. Its becoming plain to me that 4.5 and beyond is where most chips start to hit issues. Hell if they could all do 4.0Mhz even - don't you think Intel would market them as such?

All chips are different - and as has been proved many, many times over it's not the boards that are holding you back its the specific CPU. That why people will BIN them and charge extra to play.

[mean6]

I asked one of my contacts about your predicament with canadacomputers to see if theres anything that can be done to help you out. I'll let you know

Thanks alot Ket!

[mean6]

*Update* I ended up just going and buying a new board and cpu. I installed the board first, and used the processor that I got yesterday. And guess what? Same thing. I installed the new processor, and guess what happened? Same thing. I looked at the batch number of the cpu I got yesterday, and the one today, and they are the exact same. My fault for not checking before hand, but who would have guessed. So, my conclusion is... the 2500K batch# 3049A355 SUCKS!!! You have been warned. Now I have to return all this hardware, and probably lose the 5% restocking fee. I will see if I can swap my cpu, for another batch. Fingers crossed.

[piquadrat]

Your memory is common denominator. Underclock it.

[pythonse]

Your memory is common denominator. Underclock it.

thank you!

[mean6]

Strange thing, I have been screwing with this thing all day, and right now it's running at 4.8 I remember I did try scaling the memory back, but it was only to 1866, I don't think I tried 1600. A little confused because with my other process I was able to run the stock settings of 2133. This IMC must not like the memory. Any adjustments I could do to try and get the memory to running at rated speeds? So, right now to have the system running at 4.8 I have adjusted the following

Muliplier: 48 (obviously)
PLL: Enabled
Turbo Limits set at: 255,255,255
Memory: Auto @1600
CPU Core Voltage: Offset +90
Load Line Calibration: Level 3

Everything else is left at default.

It is idling at 27 degrees, and reaches 78 degrees using Prime. (according to Coretemp)
How do these settings look?

[piquadrat]

>mean6
VTT up but below 1.2V
Verify that autodetected mem timings meeting G.Skill specification.

[VonZippa]

Having read all of this thread I cannot see what 'these' issues are. Many people (myself included) are happily running decent overclocks without issue - call us lucky winners in the CPU lottery if you like. If you don't believe me (or others here) - then go read Anandtech's review of the Ex4. They got along together just fine. And I am sure they did a much better job of putting the board and a CPU through its paces. But even they stopped at 4.8 (and I am sure they have more technical know how than most).

Yes - some people may need to RMA a board. It happens.

Not being able to get stable at some pre-defined idea of 'decent' overclock at whatever cost (including trashing your gear) is 'by the by'.

No CPU has a god given right to hit 4.6 / 4.8 or 5.0Mhz+. Its becoming plain to me that 4.5 and beyond is where most chips start to hit issues. Hell if they could all do 4.0Mhz even - don't you think Intel would market them as such?

All chips are different - and as has been proved many, many times over it's not the boards that are holding you back its the specific CPU. That why people will BIN them and charge extra to play.

When I hear someone mention that they can run a particular overclock just fine at one moment but then not be able to cold boot Windows at those same settings at some other moment, as mean6 did a couple pages back, then I'd say that sounds pretty reminiscent of the problem I described with my rig.

I could more easily accept that I've just reached the natural limit of my particular chip if I had just hit a point where the chip never booted or worked stable past. But when I've got a particular overclock running 8 hours Prime95 stable one moment and the rig won't even boot Windows at those same exact settings the next moment, it makes me wonder. If the CPU could run at that speed just fine before, then what's keeping it from booting now? Maybe it is the CPU, but if so that seems like awfully quirky behavior to me. Surely running a stress test overnight is a more difficult feat than booting Windows.

The only thing I can imagine that might explain this is if the chip was running at the bleeding edge of its overclocking capability at 4.6 and 8 hours of stressing it somehow degraded the chip slightly in such a way that it would no longer even boot Windows at that same speed. But that’s just pure unsubstantiated speculation as I’ve not heard anyone else mention anything like that with their CPUs.

[BranFB]

If you're using offset voltage - it might be trying to boot with less voltage that requires !

[VonZippa]

If you're using offset voltage - it might be trying to boot with less voltage that requires !

True, but I tried it with fixed voltage, additional turbo voltage, auto voltage and many different combinations of those options and still no go.

[Falkentyne]

*Update* I ended up just going and buying a new board and cpu. I installed the board first, and used the processor that I got yesterday. And guess what? Same thing. I installed the new processor, and guess what happened? Same thing. I looked at the batch number of the cpu I got yesterday, and the one today, and they are the exact same. My fault for not checking before hand, but who would have guessed. So, my conclusion is... the 2500K batch# 3049A355 SUCKS!!! You have been warned. Now I have to return all this hardware, and probably lose the 5% restocking fee. I will see if I can swap my cpu, for another batch. Fingers crossed.

Dude!!!!
It's NOT THE PROCESSOR!!!!!!
Don't you even know that batch numbers are virtually meaningless? There is some minor variation, and SOME batches do show gravitation towards one end, but a lot of it is random. And Plant assembly codes (Costa Rica/Malaysia) mean nothing, as the CPU's aren't actually fabbed at those locations (They come from elsewhere).

We already explained what the problem may be.
It may be a bad flash where you didn't clear the RTC correctly.
It may be a problem with the board not applying PLL Overvoltage on a cold boot.

It's NOT THE CPU.

If you find another CPU where the problem happens at X48 multiplier instead of X44, then you know it's a bug with something that went wrong in the flash, with some setting.

It could also theoretically be a memory related issue, or even a USB issue. Not likely to be, but it is possible.

Uh...seems from your above post it could be the RAM.

I know that someone was having cold boot problems on a UD5 with a certain brand...he switched to a different RAM brand and the problem became 100% solved....

[pkdick]

Hello,

First I would to tell you all thank you for this very interesting thread about the Asrock P67 boards: I own myself an Asrock P67 Extreme 6 and it helped me a lot to configure it.

I have however a problem now: I bought a Highpoint Rockethybrid 1220 extension card and I want to update its bios to the latest version. Using Windows 7 64 bit it is not possible to use directly the updating software provided by Hihgpoint (due to an incompatibility between 64 bit windows version and the mvf_mag.exe program).

So I unsuccessfully tried two things:

- first I plugged my old floppy disk unit but it could not strangely work,
- then I tried to create a DOS bootable usb stick (using a 1 Gb stick): every trial I did failed as the computer did not seem to see it as bootable.

Is there anyone that has a piece of advice ?

Thank you in advance.

[banwell]

When I hear someone mention that they can run a particular overclock just fine at one moment but then not be able to cold boot Windows at those same settings at some other moment, as mean6 did a couple pages back, then I'd say that sounds pretty reminiscent of the problem I described with my rig.

I could prime 95 blend for 12 hours at 0.03v offset at x46 (no llc). It ran games and folded without issue. Then one day BSOD out of the nowhere starting a game. I went back and Primed and it BSOD after 1h30. With any stability test you require the specific combination of factors that reveal the vulnerability - that takes time and is completely random.

I went back and revisited my overclock. After much time I settled on a mild LLC and lower offset (llc by its very nature adds vcore). I knew my target vcore and solved for the two other variables.

So you can indeed Prime for hours and fail to boot windows. In fact many other people have had this issue. It's the peculiar nature of SBs. And as I said before, you need to test with IBT to check Gflop for no down clocking too.

IMO the common denominator here is simply bad overclocks.

[Ket]

Hello,

First I would to tell you all thank you for this very interesting thread about the Asrock P67 boards: I own myself an Asrock P67 Extreme 6 and it helped me a lot to configure it.

I have however a problem now: I bought a Highpoint Rockethybrid 1220 extension card and I want to update its bios to the latest version. Using Windows 7 64 bit it is not possible to use directly the updating software provided by Hihgpoint (due to an incompatibility between 64 bit windows version and the mvf_mag.exe program).

So I unsuccessfully tried two things:

- first I plugged my old floppy disk unit but it could not strangely work,
- then I tried to create a DOS bootable usb stick (using a 1 Gb stick): every trial I did failed as the computer did not seem to see it as bootable.

Is there anyone that has a piece of advice ?

Thank you in advance.

If you are sure the USB stick is bootable, just put it in a USB2 port and press F11 during boot then just select the device.

[pkdick]

If you are sure the USB stick is bootable, just put it in a USB2 port and press F11 during boot then just select the device.

Hello Ket,

Thank you for your support: I tried all USB ports I had without success (even at the back of my chassis. To be sure that my USB stick was bootable I used the memtest software USB stick version: it does not work...

With regards to my floppy disk, here again I do not understand: I have two of these and none of them work (I tried different cables)...

[Ket]

Your USB stick doesn't appear to have the necessary files to be bootable. Download the HP USB Disk Storage Format Tool and use that. You may also need to pay a visit to bootdisk.com to get the necessary files to use with the HP tool.

[swampy]

Is your usb fat32?

[pkdick]

Is your usb fat32?

Yes it is, and I already tried the HP tool without success...

Have you tried to boot on a USB stick ?

[Ket]

Yes, I can boot with any of my USB sticks no problems.

[pythonse]

anyone have the 1.7? i need to dump the 1.61 as its buggy

[BranFB]

yeah. It's stable for now... but my VID got higher :O
and it takes higher voltage to be stable!

[pythonse]

wow...so what is the best bios for overclocking? 1.64?

[gurneyhck]

yeah. It's stable for now... but my VID got higher :O
and it takes higher voltage to be stable!

On what frequency your testing it?

[pythonse]

so i passed the intel burn test at 5ghz 67 gigaflops on the first and 69 on every other one after..but when i start a game it bsods to memory dump...back it off to 4.8 and it works fine

[Ket]

You wern't testing with AVX enabled thats why you thought you were stable @ 5GHz