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

[Nw0rb]

And i wanna throw this computer... This F1 error when i have pll over voltage enabled is getting annoying but im doing some test and it seems to be cause by the ram atleast thats what my test show i have a cl9 gskill kit and if i run it at 9-10-9-27 it will cause the F1 error but if i let it auto set timings it doesnt do it atleast it hasnt in an hour or so.

[Ket]

Look in the "Boot" menu. Its labelled "PCI ROM Priority", or simply scan the page for "Legacy ROM". The option wasn't always in the UEFI so if you aren't running the latest UEFI it probably won't be there.

[Nw0rb]

Ok so after som more test at 4.6 with pll enabled everything boots fine no errors ... so would this mean my chip max is 4.6 ? idk tho cause i can get 5.0 stable and bench n such with no problem theres gotta be somthing causin this problem.. and ket no i dont have latest bios because they didnt fix the full reset when spread spectrum is enabled

[Ket]

You don't say what you are running your memory at, but you can try bumping up VTT. You can't go in "expecting" your hardware to hit certain levels, thats just not how it works. Take incremental steps, you will make your life a lot easier when trying to narrow problems down.

[Nw0rb]

Weird seems like updating to 1.60 has fixed the f1 problem.

[Ket]

Thats not odd, new firmware = bug fixes & improvements. Quite normal

[Nw0rb]

Weird thing is my heaven scores went down 20 points must the board sli drivers ??

[Ket]

You're worried about 20 points in Heaven? That difference can be put down to any number of things. In reality that 20 points is probably actually what? Equivillent to about 0.2FPS? Thats well within any margin of error all benchmarks have.

[Nw0rb]

No but really its like 2-7 fps depending on where heaven is on the loop. but anyway ket got a question how is it that somone at 4.3 on a 2500k can out score mine at 4.6 ? http://3dmark.com/3dmv/2928652 and its doing more then beating it i get 54,400 at 4.6 and the 4.3 for example is at 60,000 and its not just that one browse i5-2500k sli 460s and all cpu scores under 4.6 smash me by 4 - 6000 and the ones at 5 are getting 64,000 im really confused am i overclocking it wrong ?

[radier]

Your CPU is throttling down because you didn't set Turbo Boost and Core Current Limits manually to for e.g. 255/255/300

[Nw0rb]

Your CPU is throttling down because you didn't set Turbo Boost and Core Current Limits manually to for e.g. 255/255/300

Nope iv had those set from the begining. Does this mean i have a poor performing chip ?? because i dont understand it iv tried turbo use manual multi nothing seems to get me to score what a bounch of 4.3 4.4 2500ks get around 56-60,000 and 5.0 is getting 62,000 64,000...

[radier]

But CPU score with nVidia graphics is related to GPU performance not CPU (because of PhysX).

[Ket]

No but really its like 2-7 fps depending on where heaven is on the loop. but anyway ket got a question how is it that somone at 4.3 on a 2500k can out score mine at 4.6 ? http://3dmark.com/3dmv/2928652 and its doing more then beating it i get 54,400 at 4.6 and the 4.3 for example is at 60,000 and its not just that one browse i5-2500k sli 460s and all cpu scores under 4.6 smash me by 4 - 6000 and the ones at 5 are getting 64,000 im really confused am i overclocking it wrong ?

20pts in Heaven is certainly not equivillent to 2-7FPS. I can reduce the OC on my GTX460 to simulate that point difference and its about 0.5FPS tops running @ 1680*1050, tessellation on extreme, maximum AA & AF. For your difference to be so high you must be benching at a very low resolution. As for the other thing, could be any number of things, background applications, your CPU throttling, stripped down OS, unstable OC, etc.

[Nw0rb]

20pts in Heaven is certainly not equivillent to 2-7FPS. I can reduce the OC on my GTX460 to simulate that point difference and its about 0.5FPS tops running @ 1680*1050, tessellation on extreme, maximum AA & AF. For your difference to be so high you must be benching at a very low resolution. As for the other thing, could be any number of things, background applications, your CPU throttling, stripped down OS, unstable OC, etc.

So basicly its a poor performing chip because iv tried everything except getting new memory but that should not make a 4000-6000 in cpu score. and at 1680x1050 which is what i run it at ocn heaven 2.5 settings. and its the same around 2-3 fps... like i said its what ever and to be honest i think its a driver issue with SLi cause there have been alot of complaints about the 270 serires drivers everyone says use older ones and wait for 280 to be released

[radier]

Read with understanding what I have wrote. 3Dmark Vantage/11 CPU Score when using nVidia PhysX capable graphics is depended on GPU not CPU! IS THAT CLEAR? Perhaps your opponents have higher clocked ROP/Shaders on their cards than you.
Capisci?

There is no such thing like "poor performing chip".

[Ket]

So basicly its a poor performing chip because iv tried everything except getting new memory but that should not make a 4000-6000 in cpu score. and at 1680x1050 which is what i run it at ocn heaven 2.5 settings. and its the same around 2-3 fps... like i said its what ever and to be honest i think its a driver issue with SLi cause there have been alot of complaints about the 270 serires drivers everyone says use older ones and wait for 280 to be released

Heres a crash course checklist that may help

Make absolutely sure your OC is stable. What works for me is running BOINC, if your system passes 6+hrs of that you can say you are good.
Use the latest UEFI
Set memory timings and voltage manually
make sure you are using the latest drivers for all your hardware
Optimise Windows (download regcleaner, click the "startup" tab and bare witness to all the junk starting with Windows)
Run services.msc for further optimising.
For Vantage, you will get a significantly higher score if PhysX is set to run on the GPU, likewise your scores in most other benchmarks will likely go down a bit if you don't set Physics to run on the CPU

[BranFB]

My E4 just died. No warning, no nothing... this morning I pressed the start button and nothing happened! R.I.P.

[Ket]

Well, thats not giving me much to work with You sure its not the PSU that died without warning? Or even the fuse thats simply gone in the plug.

[BranFB]

I'm sure.
I have tested it with another PSU. - NO GO!
I'm waiting on a multimeter to check the voltage in cables... if I can.
Otherwise there's not a single black mark,spot or anything on the board. It just died.... WOW!

[bat4]

Hi all,

There is a new UEFI 1.80 for Asrock P67 Extreme4 on the Asrock support site:

http://www.asrock.com/mb/download.as...xtreme4&o=BIOS

Has anyone from the user community tested this UEFI?

[xdan]

I'm sure.
I have tested it with another PSU. - NO GO!
I'm waiting on a multimeter to check the voltage in cables... if I can.
Otherwise there's not a single black mark,spot or anything on the board. It just died.... WOW!

There where some extremely rare cases with SB, when the cpu just died from nothing...
Check also the cpu on another mb, than it's clear if it is the mb.

[BranFB]

If it was the CPU... there should be still some activity on the mb, right. Like drDebug indicating somehting or at least some leds light up

I'm RMA it tomorrow

[Ket]

Its worth checking the CPU, if that checks out ok would seem safe to think its the board. Deffinately very odd with no warning your system just doesn't work anymore.

[trees415]

Hey I wanted to post this just in case anyone was having difficulties achieving decent clock speeds. When I first built my computer, I was able to easily hit 5ghz with my setup. However, for the last month or so, I've been completely unable to get my system to boot at anything higher than 4.2 ghz (multi=42 x FSB=100). When I would raise my vcore voltage to 1.4+, I could get it to boot at 44 MAX, but my system was extremely unstable, and would fail almost every stability test.

After toying with latency here and there, I tried making some adjustments to my memory timings. I had set them originally to the manufacturer recommended settings of 9-9-9-24-1N, which was also the recommended settings stated in the UEFI screen XMP profile. However, CPU-Z suggested at 800mhz (I have 2x4gb PC12800 ram), that my timings should be 9-9-9-25-2N.
Out of curiosity, I loosened my ram timings slightly, and I immediately noticed a reasonable increase in performance. So rebooted to the UEFI, set my Multi to 46 and it booted fine, and ran Lynx for 20 passes, set to a problem size that used nearly all my ram. To my amazement, the test was able to complete without any stability issues, and more importantly with incredible results. Every pass took nearly the exact same amount of time, and calculated 120 Gflops with each pass without more than a .7 deviation. All this and thevoltage never went above 1.40v (Excellent voltage for my CPU). I re-ran it at a 48 multi, and achieved similar results. At 50 (5.0 Ghz overclock), it wasn't as precise, but no issues stability wise. While it required roughly 1.48v for the cpu, temps never went about 80C (thank god I upgraded my cpu cooler).

Long story short, if you can't seem to get your multiplier to budge without boot or stability issues, try making some adjustments to your ram timings, or even drop your FSB down a hair (.01-.02), then see where your at.


Honestly, the performance benefits I've seen as a result totally outweighs the potential loss of performance from looser ram timings. Manufacturer recommended ram timings are not always the ideal settings for your memory, as two memory packs of the exact same brand and model are usually quite different performance-wise, hence the reason why upgrading ram typically requires you buy a whole new ram set.

I'm slightly new to the OC scene so this is probably common knowledge, but I figured I'd share it just in case it might help anyone. Thanks for the thread Ket, it's been super useful for me the last few months since I got my P67 E4 mobo.

[BranFB]

Will do.
My seller said that this is not the first board acting like that