***Official*** ASUS P67/H67 BIOS Updates

[Mathemagician]

I had this issue as well but then I installed the patch from the quote given before by SimpleTECH. I was able to get back the CPU ratio control from this. I only had to install it once, and I did so after I installed the new AI Suite II. Quote added below with the link. Hope it works for you guys!

I've uninstalled AI Suite, installed both ME and INF, run the patcher twice, reinstalled AI Suite, and and still cannot adjust the multiplier in TurboV. Any further ideas would be greatly appreciated.



Edit: Fixed! Enter BIOS and change the Turbo Ratio setting to "By Per Core". Boot and run TurboV. Click More Settings, Advanced, and the CPU Ratio tab is there. Click the ON button to enable ratio adjustment and allow the system to reboot. TurboV changes the Turbo Ratio setting from "By Per Core" back to "By All Cores" and subsequently enables the CPU Ratio adjustment.

Thanks to all for the assistance and suggestions!

[poke349]

Does anyone has at least 4.5ghz with 4 sticks?

Yes. Check my last post for the settings. I've got 4 x 4GB @ 1333 MHz.
I can probably get to 5 GHz if I upped the vcore more, but I'm getting scared of putting it past 1.45v.

[simonmaltby]

I've uninstalled AI Suite, installed both ME and INF, run the patcher twice, reinstalled AI Suite, and and still cannot adjust the multiplier in TurboV. Any further ideas would be greatly appreciated.



Edit: Fixed! Enter BIOS and change the Turbo Ratio setting to "By Per Core". Boot and run TurboV. Click More Settings, Advanced, and the CPU Ratio tab is there. Click the ON button to enable ratio adjustment and allow the system to reboot. TurboV changes the Turbo Ratio setting from "By Per Core" back to "By All Cores" and subsequently enables the CPU Ratio adjustment.

Thanks to all for the assistance and suggestions!

I just had the same issue with the Multi adjustment following bios to 1253 and updating the AI Suite II.

I discovered that if you set the Turbo to AUTO in the bios rather than setting it manually or setting allow settings controlled by OS... then booted and ran TurboV EVO I got the enable CPU ratio. Clicked on that and let it re-boot. Then when windows re-booted I had the CPU Ratio tab with 38 set and adjustment available. Next I rebooted and set my Multi in the Bios - in my case 48 and restarted windows. Still have the CPU Ratio tab and adjustment.

Bit of a fad, but seems to work

[orion23]

I went back to 1253 this morning since it allows the PC to boot more easily after a crash, no double boot or stock bios vs 1053...

My 24/7 will be
46X100
DDR3 1600
Everything left AUTO or STOCK, except for the CPU voltage:
1.35 Vcore

BTW, Linx can't be a test tool for Sandry Bridge....
Priming away 1H + right now....

I can't get 4.8ghz to be prime stable even with 1.395V...

[wampir]

Yes. Check my last post for the settings. I've got 4 x 4GB @ 1333 MHz.
I can probably get to 5 GHz if I upped the vcore more, but I'm getting scared of putting it past 1.45v.

Could You paste your settings here ?

[poke349]

I'm not sure why you're surprised by this. The point of using higher LLC settings is to suppress the Vdroop specifically at load. Your CPU spikes to >1.4V because your it will briefly remain at your max turbo multiplier in a zero load situation and you are using a very high offset.

The fix:
CPU Voltage = Offset Mode = 0.00V
Load Line Calibration = Ultra High

Those settings will probably get you around 1.35V under load with your overclock. You can drop the offset in .05V increments and it will also drop the load Vcore by a proportional amount. If you lose stability at LOWER clocks due to undervolting, then drop your LLC from Ultra High to High (and raise your offset slightly again).

General consensus seems to be the following...
The majority of these processors can actually run at 4.8GHz on 1.37-1.39V using appropriate phase control options
4.5-4.6GHz can be achieved with 1.33-1.35V
4.9GHz isn't a wall, but it is a very steep ramp. My particular sample wants at least 1.41-1.42V at 4.9GHz vs 1.36-1.37V for 4.8GHz.

My settings for a 24/7 OC (1.016V idle, 1.368V load, 1.376V load-drop):
AI Overclock Tuner = Manual
BCLK/PEG Frequency = 100
Turbo Ratio = By All Cores = 48
Internal PLL Overvoltage = Enabled
Memory Frequency = DDR3-1866
OC Tuner = Disabled
EPU Power Saving Mode = Disabled
DRAM Timing Control = 8-9-8-24-1T
Load Line Calibration = Ultra High
Phase Control = Extreme
Duty Control = Extreme
CPU Voltage = Offset Mode = .020V
DRAM Voltage = 1.6V

I was surprised because the spikes only started appearing when I upgraded to 1253. The vdroop I noticed with the old bios was very minimal. If I ignore the random spikes with the new bios, the vdroop is still about the same. (~0.02 or so)
That's why I'm tempted to think that the spikes had nothing to do with LLC.

I'm using offset mode so that it will lower voltages when it's idle. If I forced it to a particular voltage, it stays that high even at idle.

I'll try lowering the offset and turning on LLC to see what happens.

Could You paste your settings here ?

I'll repost here:

BIOS: 1253
vcore offset mode: +0.085
LLC: forced off (standard)
PLL overvoltage: on
Everything else left on defaults...

Idle vcore: 1.09v.
Load vcore: 1.32v - 1.34v
Random vcore spikes: 1.38v - 1.41v


Results with these settings:

4.2 GHz (24/7)
4.5 GHz (completely prime stable)
4.6 GHz (4 hours prime stable)

Those are my 24/7 settings. 4 hours prime-stable @ 4.6 GHz. Completely prime-stable @ 4.5 GHz. But I'll be doing software development on it - so I'm clocking it at 4.2 GHz to be safe.

For my temporary suicide run, I put the offset to +1.25 or so (I don't remember the exact number.)
With this I was able to get 4.9 GHz at least 10 min. prime-stable. 5.0 GHz was screenshot-able but not bootable. Load voltages were over 1.43v so I didn't want to keep it there long enough to get them stable.

For this suicide run, I left my window open to let the sub-zero ambient outside air into the room and the case. (ambient outside was -10F, inside the case, it was about 40 - 50F)
So temperatures were never a problem. I'm "sorta" on water. (see my siggy for exact specs)

[kidsafe]

Don't try turning on LLC...do it now and leave it on. You are supposed to be using at least a LLC of High at your clock speed, and possibly Ultra High. Ask anyone else for their opinion, they'll say the same thing. There is no reason for you to be running a .085V offset just to run 4.6 GHz. It's forcing your idle load to be almost 1.1V when it should be at or below 1.0V. And as you point out, you can barely even hit the 4.8-4.9 GHz range on an even more ridiculous offset headed into the +1V territory. You are doing it wrong.

[poke349]

Don't try turning on LLC...do it now and leave it on. You are supposed to be using at least a LLC of High at your clock speed, and possibly Ultra High. Ask anyone else for their opinion, they'll say the same thing. There is no reason for you to be running a .085V offset just to run 4.6 GHz. It's forcing your idle load to be almost 1.1V when it should be at or below 1.0V. And as you point out, you can barely even hit the 4.8-4.9 GHz range on an even more ridiculous offset headed into the +1V territory. You are doing it wrong.

Been playing with it a bit and it's working. Thanks!

I tuned the LLC and the offset so that the load vcore is the same as it was before and it seems to be just as stable. The only difference now is that the spikes are gone and my idle vcore is low (since the offset is lower).

I didn't zero the offset. I turned it down by a lot though.

I didn't realize the LLC increased the base voltage as the load is increased. During the original i7 days, LLC only eliminated the vdroop (with very little overshoot). I turned off LLC because of the unwanted voltage spikes and the rumors that it was killing CPUs.

Now, LLC seems to actually increase the vcore - so I don't have to (as much). I guess SB is very different. (or at least the mobo...)


Right now I'm testing 4.8 GHz:
LLC = Very High and offset = 0.025.

If this is stable for a few hours, then 4.5 GHz will be my new 24/7. (I do need the safety margin since I'll be doing software dev work - so I need some degree of reliability.)
I'll be retesting everything once I get Win7 SP1 with AVX LinX... I know that's gonna put a lot more extra stress on the chip.

[BR_Impulse]

I've uninstalled AI Suite, installed both ME and INF, run the patcher twice, reinstalled AI Suite, and and still cannot adjust the multiplier in TurboV. Any further ideas would be greatly appreciated.



Edit: Fixed! Enter BIOS and change the Turbo Ratio setting to "By Per Core". Boot and run TurboV. Click More Settings, Advanced, and the CPU Ratio tab is there. Click the ON button to enable ratio adjustment and allow the system to reboot. TurboV changes the Turbo Ratio setting from "By Per Core" back to "By All Cores" and subsequently enables the CPU Ratio adjustment.

Thanks to all for the assistance and suggestions!

Glad to hear you got it working! I had mine set to "By All Cores" from the start so I'm not sure what happened there, but maybe some setting needed to instigate a change in the Suite.

[kidsafe]

Been playing with it a bit and it's working. Thanks!

I tuned the LLC and the offset so that the load vcore is the same as it was before and it seems to be just as stable. The only difference now is that the spikes are gone and my idle vcore is low (since the offset is lower).

I didn't zero the offset. I turned it down by a lot though.

I didn't realize the LLC increased the base voltage as the load is increased. During the original i7 days, LLC only eliminated the vdroop (with very little overshoot). I turned off LLC because of the unwanted voltage spikes and the rumors that it was killing CPUs.

Now, LLC seems to actually increase the vcore - so I don't have to (as much). I guess SB is very different. (or at least the mobo...)

Yeah you can't set the offset to 0.00V or it will reset itself to Auto. But you can definitely set a minimal offset of 0.005V. As far as the behavior of the Vcore is concerned, I really like the Asus LLC implementation actually and the setting of Ultra High is just about perfect for running 2500K/2600K processors at 4.7-4.9 GHz. The Asus engineers really outdid the competition on this generation of boards and the amount of power delivery control options you have.

[poke349]

Glad to hear you got it working! I had mine set to "By All Cores" from the start so I'm not sure what happened there, but maybe some setting needed to instigate a change in the Suite.

Oh wow! I didn't see Mathemagician's post until now. It worked for me to! Thanks!!!

[poke349]

Yeah you can't set the offset to 0.00V or it will reset itself to Auto. But you can definitely set a minimal offset of 0.005V. As far as the behavior of the Vcore is concerned, I really like the Asus LLC implementation actually and the setting of Ultra High is just about perfect for running 2500K/2600K processors at 4.7-4.9 GHz. The Asus engineers really outdid the competition on this generation of boards and the amount of power delivery control options you have.

Although I posted after you, I didn't see your response because I immediately went afk... lol

Been playing with it a bit more. As others have hinted, there's a steep vcore hill. For mine it's 4.8 - 4.9 GHz. I finally got 4.9 GHz > 5 min. stable. I didn't want to test longer because my temps were getting high. (Normally, I don't have this problem, but right now, I don't have my rig setup properly to get sub-zero air from outside. )

The difference between 4.8 GHz and 4.9 GHz was 0.040v. Between 4.7 and 4.8 was only 0.015v.
I tried 5.0 GHz using the marginally stable 4.9 GHz settings - blue screened about a second into my multi-threaded pi benchmark.
I'll try again later - gotta catch up on homework.

I booted up with a 4.7 GHz prime-stable setting. Then upped it via AI Suite.

These are all AI Suite settings:
4.9 GHz
Vcore = 1.230v
DDR = 1.5v

Load-Line Calibration = Ultra High
Current Capability = 120%
Frequency = 350KHz
Phase Control = Extreme
Duty Control = Extreme

Load vcore came out to be about 1.42 - 1.44. Temps were 80 - 85C. Ambient was about 72F. I'll try again later with sub-zero ambients.

[kidsafe]

Oh wow! I didn't see Mathemagician's post until now. It worked for me to! Thanks!!!

I think all the change in AI Suite does is set the EFI to "By All Cores = Auto" and that's what recovers the option in TurboV.

[poke349]

I think all the change in AI Suite does is set it to "By All Cores = Auto" and that's what recovers the option in TurboV.

Hmm... I don't think I ever had it on auto. As soon as I upgraded the BIOS, I immediately set it to what I had before. (all cores - 42x)

That's probably why...

[Daddyjaxx]

I had issues with EZ Flash saying "Not an EFI BIOS" but then I just put it on a different flash drive and it works just fine.

New BIOS is nice for the Pro but my previously stable CPU clock is now unstable with even more voltage.

The file system MUST be FAT32 to use EZ Flash.

[BR_Impulse]

The file system MUST be FAT32 to use EZ Flash.

Is this for external devices only (USB flash drives and the like)? I update in EZ Flash just by navigating to my windows desktop folder where I downloaded my BIOS to from within EZ Flash itself with the SATA drive being in NTFS and I was able to update my BIOS both times successfully.

[zoob]

Try this please, shutdown the system. Remove power from the power supply for a minute or two, clear CMOS, then reattach power. Startup, enter UEFI and input your required settings, for your memory please manually change the memory speed, timings and dram voltage based on your kit specifications. Reboot, enter Windows, then shutdown again. Now restart the system. This will clear the ME and memory/CPU tables. It could be that after flashing the BIOS up from a lower version (pre-1053) that the tables were never properly cleared. This is especially true if the MEI driver is not loaded. Let me know if this helps, if not we will go to step two.

In testing I found that the voltages are slightly lower overall but stability was still maintained. We also changed how the auto settings work on C1E, C3, and C6 and are dependent now on EIST being enabled or not. In addition LLC was fine tuned so previous settings might not work as before. If you continue to have problems, please PM me and we will work it out offline.

Thanks, I actually did the CMOS reset after flashing up to 1053 and that cured my double post problem. I meant that this new 1253 BIOS lets me stress test longer at 49x100 with the same settings compared to 1053.

I still am getting power downs after about an hour of testing... the CPU is getting just about as much voltage as I'd like to push through it at the moment, so I'll just keep 48x100 as my OC.

Regarding the slightly lower voltages, I found I had to bump up my vcore from 1.330 to 1.335 to maintain stability w/Extreme LLC. However, after reading about the voltage spikes I'm now switched over to +0.05V offset on Ultra High LLC. It keeps my CPU at about 1.344V under full load which is the same load voltage.

[Utnorris]

Ok, so I went from the Rampage 3 Formula to the Deluxe and tried installing the AI suite. I had the one for the Rampage already on there thinking it would just upgrade it, but it did not. I tried to unistall it, but I get the error message that my 2500K is not supported. Is there a way to uninstall without doing a fresh build so I can reinstall the one for the Deluxe?

[some_one]

I've tried EZFLASH; Bupdater from DOS and Windows update utility. All will not allow you to fallback to previous bios.
If there is a way, i would appreciate someone posting the info here. This is for the P8P67 series boards.

It seems that with 1253 the hardware version has been increased for at least P8P67 evo, pro, saber and standard while deluxe remains unchanged. I'm guessing this means those with updated HW versions can not simply downgrade as the deluxe can.

Not sure why there was no warning of this unless it's a bug in which case if the next official release is with the original hardware version then you may not be able to upgrade easily either.

Unless there is an override switch, options would appear to be to try and find a 3rd party flasher or remove the chip and program externally. IMHO best though if Asus can offer some feedback as to why they did this and hopefully offer some sort of solution if required.


Sorry about the pic of bupdater.exe, it's what was available at the time and I was to lazy too find/make something better.

[Sam_oslo]

I'm considering to get a ASUS P8P67-M PRO for a compact build.

The big border seams to be doing good, but I'm wondering if there is a big difference between BIOSes (or other important factors impacting 24/7 OC) for ASUS P8P67-M PRO vs ASUS P8P67 PRO?.

[One_Hertz]

Just in case some people are not aware, the 1253 bios does not like the PC being put to sleep. Mine refused to wake up and then refused to post until I removed the battery from the mobo for a few minutes to wipe CMOS... That sure was annoying.

[Chri$ch]

ASUS Maximus IV Extreme BIOS 0950

Changelog:
- Fixed Memory ratio sometimes not changing bug
- Fixed XMP + OC Load profile bug
- Adds Memory Bandwidth Turbo for performance improvement
- Fixed Wrong Target CPU Speed shown after Clear CMOS

MaximusIV-Extreme-ASUS-0950.ROM

[Forceman]

Just in case some people are not aware, the 1253 bios does not like the PC being put to sleep. Mine refused to wake up and then refused to post until I removed the battery from the mobo for a few minutes to wipe CMOS... That sure was annoying.

If you are using the PLL Overvoltage feature, it is a known bug. Disabling PLL Overvoltage normally lets sleep start working again, depending on overclock level.

[Hooti3]

So can we expect to see a new bios for the WS Revolution any time soon?

[splmann]

So can we expect to see a new bios for the WS Revolution any time soon?

I also Need this BIOS !!