ASUS P8P67 Pro/Evo/Deluxe OC discussion thread

[DStealth]

Does anyone help me? please...
I just finished my new p8p67 Pro B2 with 2600k cpu system and the pc wont boot... the cpu led is red. No any beep code from bios? what's happened?
(it's off 1 time and restart again...)
I have G-skill RJ-X 1600 4*2 memory SATA3 640*2 ANTECH TPQ1000 and GTX 580.

Thanks

Welcome to the Club...just pray for your CPU

http://www.xtremesystems.org/forums/...&postcount=969

[LeetMiniWheat]

some quirks I've found with this board...

1. voltage during loading seems inconsistent. with or without LLC, the VCore with 1 or 2 cores loaded is much lower than with 4 cores loaded. shouldn't this be the opposite since VDroop would be more noticeable with more current? for example, with 1-2 cores loaded I've seen as low as 1.336v-1.344v, but with 3-4 cores loaded I've seen generally 1.376v. but even goes as high as 1.396v-1.40v if the load is light. this to me seems backwards.

Edit: so to be clear here, my non-LLC voltage range for 4.6GHz is 1.33v-1.400v, that's quite a large range... but doesn't decrease proportional to more load, it decreases proportional to LESS load. so, since my CPU needs 1.35v for stable 4.6, it's getting both too much volts and not enough volts during different workloads.

2. LLC doesn't really seem to adjust load volts properly. enabling LLC seems to generally just give a flat all around voltage increase with not so much difference during load, compared to using no LLC and higher volts. for example, 1.396v idle compared to 1.376 idle without LLC. if I just increase non-LLC volts without to reach 1.396, sure there is a bit more load drop, but not significantly different. gigabyte seems to have done LLC better than Asus. they don't overshoot the mark, and they don't drop more than 0.010.

additionally, (related to the above quirks with LLC) it seems that setting too high of a LLC to reach the target stable load volts ends up with idle 1.6ghz power savings clocks going below the stable mark (<1.000v) which essentially makes LLC kind of worthless.

Has anyone else seen similar results?

Left CPU Z shows voltage with C states set to Auto. Right shows voltage with C states enabled in Bios. Pretty sure I've seen voltage drop below 0.900 when set to enabled. Auto does not drop below 0.992. The last bsod frigged up a Blu-ray disc while burning, which can get expensive if it continues.

I'm pretty sure anything below 1.025v-1.050v will be unstable at 1600mhz. (read that somewhere, but can't remember where) and in my testing it seems anything below 1.00v is definitely dangerous to run. (risk of BSOD for sure). you might want to loosen up the LLC a bit and increase volts like I've had to do.

[McLaren__F1]

Any tips on how to oc my ram on this board?

[danielkza]

My idle voltage drops to around 0.96V and I have experience no BSODs so far.

[ghmz]

I need some serious help dealing with a BSOD, code 0x124... I am using an offset mode with LLC off and other settings are default except fixed ram timings.

After many try outs, I seem to pass Small FFTs for 14h. However, the Blend test fails with a BSOD after about 4h. I suspect, there must be an issue with RAM voltage or something, but clueless what to change.

I also pass the LinX with AVX with 3072Mb of RAM, but also tried running with a full RAM usage and the LinX failed also with the same BSOD code.

Any help is appreciated.

[danielkza]

I need some serious help dealing with a BSOD, code 0x124... I am using an offset mode with LLC off and other settings are default except fixed ram timings.

After many try outs, I seem to pass Small FFTs for 14h. However, the Blend test fails with a BSOD after about 4h. I suspect, there must be an issue with RAM voltage or something, but clueless what to change.

I also pass the LinX with AVX with 3072Mb of RAM, but also tried running with a full RAM usage and the LinX failed also with the same BSOD code.

Any help is appreciated.

What is your RAM kit, and at what speeds and timings are you running it? It indeed seems your memory is not stable.

[ghmz]

What is your RAM kit, and at what speeds and timings are you running it? It indeed seems your memory is not stable.

I have this RAM. Running rated timings which are 9-9-9-24-2N, 1.5V.

[Jen]

Having problems with multipler staying at 41 when its set in bios . When running prime 95 i notice that it defaults back to 3500mhz (34x103 fsb) . It doesnt default if i use the 40x multipler . Trying to figure out why its doing this on this board . Cooling is adequate as its not going past 63c in this room.

2600k
crucial memory
Asus P8P67 M

Ai Overclock Tuner - manual
Bclk/peg freq - 103.0
By All Cores (Can Adjust In OS) , currently set at 41
Internal PLL Overvoltage - AUTO
Memory Freq (1090mhz)
Offset Mode sign (-)
CPU Offset Voltage 0.020
DRAM Voltage (1.50)
VCCIO Voltage 1.00000
PCH Voltage (AUTO)
Load Line Calibration (AUTO)
-----------------------------------

Ai Tweaker / CPU Management
CPU Ratio (AUTO)
Enhanced Intel Speedstep Technology (ENABLED)
Turbo (ENABLED)
---------------------------------
Turbo Mode Parameters
Long Duration Power Limit (AUTO)
Long Duration Maintained (AUTO)
Short Duration Power Limit (AUTO)
Additional Turbo Voltage (AUTO)
Primary Plane Current Limit (AUTO)
________________________
Advanced\CPU configuration
CPU ratio = Auto
Intel adaptive thermal monitor - disabled
Hyper threading - enabled
Active Processors (ALL)
limit CPuid maxium disabled
Execute disable bit - enabled
Intel virtualization tech = disabled
Enhanced Intel Speedstep Technology (ENABLED)
Turbo Mode -enabled
CPU C1E, (AUTO)
CPU C3 Report (AUTO)
CPU C6 report (AUTO)

[Darkstormz]

I´ve got some serious issues going on my P8P67 WS Revolution.

When I run 3DMark Vantage / 3DMark 11, I get a message "Nvidia display driver stopped responding and has recovered"

Tried everything, GTX480, GTX580 and 9600GT card. Still same problem.
Tried to switch PSU from Corsair HX620 -> Corsair AX850. Still same problem.
Re-installed Win7 Ultimate SP1 and still got the same problem.

Sometimes I can run 3DMark for 30sec, sometimes a couple of minutes.
The most strange thing is: Can run Unigine Heaven benchmark for 30min without any trouble.

Everything is at stock 2600K, 16GB Vengeance 1866MHz...

Anyone?

[McLaren__F1]

I´ve got some serious issues going on my P8P67 WS Revolution.

When I run 3DMark Vantage / 3DMark 11, I get a message "Nvidia display driver stopped responding and has recovered"

Tried everything, GTX480, GTX580 and 9600GT card. Still same problem.
Tried to switch PSU from Corsair HX620 -> Corsair AX850. Still same problem.
Re-installed Win7 Ultimate SP1 and still got the same problem.

Sometimes I can run 3DMark for 30sec, sometimes a couple of minutes.
The most strange thing is: Can run Unigine Heaven benchmark for 30min without any trouble.

Everything is at stock 2600K, 16GB Vengeance 1866MHz...

Anyone?

You have the latest drivers?

[Sauron_Vie]

I need some serious help dealing with a BSOD, code 0x124... I am using an offset mode with LLC off and other settings are default except fixed ram timings.

After many try outs, I seem to pass Small FFTs for 14h. However, the Blend test fails with a BSOD after about 4h. I suspect, there must be an issue with RAM voltage or something, but clueless what to change.

I also pass the LinX with AVX with 3072Mb of RAM, but also tried running with a full RAM usage and the LinX failed also with the same BSOD code.

Any help is appreciated.

raise ur vcore a little

[Darkstormz]

You have the latest drivers?

Yes I do, still the same trouble...

[Pacha]

you using offset or manual voltages ? upping the vcore a tiny bit more ?

manual mode. I don't know what to expect as default voltage when using offset mode as you can't set default vcore but only offset value.

some quirks I've found with this board...

1. voltage during loading seems inconsistent. with or without LLC, the VCore with 1 or 2 cores loaded is much lower than with 4 cores loaded. shouldn't this be the opposite since VDroop would be more noticeable with more current? for example, with 1-2 cores loaded I've seen as low as 1.336v-1.344v, but with 3-4 cores loaded I've seen generally 1.376v. but even goes as high as 1.396v-1.40v if the load is light. this to me seems backwards.

Edit: so to be clear here, my non-LLC voltage range for 4.6GHz is 1.33v-1.400v, that's quite a large range... but doesn't decrease proportional to more load, it decreases proportional to LESS load. so, since my CPU needs 1.35v for stable 4.6, it's getting both too much volts and not enough volts during different workloads.

2. LLC doesn't really seem to adjust load volts properly. enabling LLC seems to generally just give a flat all around voltage increase with not so much difference during load, compared to using no LLC and higher volts. for example, 1.396v idle compared to 1.376 idle without LLC. if I just increase non-LLC volts without to reach 1.396, sure there is a bit more load drop, but not significantly different. gigabyte seems to have done LLC better than Asus. they don't overshoot the mark, and they don't drop more than 0.010.

additionally, (related to the above quirks with LLC) it seems that setting too high of a LLC to reach the target stable load volts ends up with idle 1.6ghz power savings clocks going below the stable mark (<1.000v) which essentially makes LLC kind of worthless.

Has anyone else seen similar results?


I'm pretty sure anything below 1.025v-1.050v will be unstable at 1600mhz. (read that somewhere, but can't remember where) and in my testing it seems anything below 1.00v is definitely dangerous to run. (risk of BSOD for sure). you might want to loosen up the LLC a bit and increase volts like I've had to do.

I can see we're more and more to find the same non sense in how this board is operating and seeing how many guys here have LynX stability but light load BSOD, I regret I've not bought a Gigabyte... P8P67 or Commando, years apart but hands together in the crap..

I need some serious help dealing with a BSOD, code 0x124... I am using an offset mode with LLC off and other settings are default except fixed ram timings.

After many try outs, I seem to pass Small FFTs for 14h. However, the Blend test fails with a BSOD after about 4h. I suspect, there must be an issue with RAM voltage or something, but clueless what to change.

I also pass the LinX with AVX with 3072Mb of RAM, but also tried running with a full RAM usage and the LinX failed also with the same BSOD code.

Any help is appreciated.

Same weirdness and dumb behaviour of the board, confirmed. So many guys here with "rock stability" but BSOD 0124 when browsing or so

raise ur vcore a little

As suggested by LeetMiniWheat I would say also that LLC at auto (which may correspond to low value) and higher vcore do the trick, at least it's better. Push the vcore too high for LynX stability, and even if it's too high for load and idle, it's high enough to prevent BSOD with light load of 1-2 cores.

[Falkentyne]

some quirks I've found with this board...

1. voltage during loading seems inconsistent. with or without LLC, the VCore with 1 or 2 cores loaded is much lower than with 4 cores loaded. shouldn't this be the opposite since VDroop would be more noticeable with more current? for example, with 1-2 cores loaded I've seen as low as 1.336v-1.344v, but with 3-4 cores loaded I've seen generally 1.376v. but even goes as high as 1.396v-1.40v if the load is light. this to me seems backwards.

Edit: so to be clear here, my non-LLC voltage range for 4.6GHz is 1.33v-1.400v, that's quite a large range... but doesn't decrease proportional to more load, it decreases proportional to LESS load. so, since my CPU needs 1.35v for stable 4.6, it's getting both too much volts and not enough volts during different workloads.

2. LLC doesn't really seem to adjust load volts properly. enabling LLC seems to generally just give a flat all around voltage increase with not so much difference during load, compared to using no LLC and higher volts. for example, 1.396v idle compared to 1.376 idle without LLC. if I just increase non-LLC volts without to reach 1.396, sure there is a bit more load drop, but not significantly different. gigabyte seems to have done LLC better than Asus. they don't overshoot the mark, and they don't drop more than 0.010.

additionally, (related to the above quirks with LLC) it seems that setting too high of a LLC to reach the target stable load volts ends up with idle 1.6ghz power savings clocks going below the stable mark (<1.000v) which essentially makes LLC kind of worthless.

Has anyone else seen similar results?




I'm pretty sure anything below 1.025v-1.050v will be unstable at 1600mhz. (read that somewhere, but can't remember where) and in my testing it seems anything below 1.00v is definitely dangerous to run. (risk of BSOD for sure). you might want to loosen up the LLC a bit and increase volts like I've had to do.

This is about the weirdest thing ever.
Can you try clearing CMOS with the power unplugged?
Every board that uses fixed (not auto and no offsets) voltages, and with C1e, C3, C6 disabled, should NOT be raising the voltage at load for any reason; vdroop (loadline slope) is supposed to lower the vcore at load on an even slope and it's been that way ever since the Pentium 4 was introduced and VRD 8.x, instead of the P3's always fixed voltage.

If you're getting increased vcore at load, and you're not using offsets and not using LLC, then either the BIOS or the board is defective somehow, because that's not how vdroop is supposed to work.
It's like the board is giving you LLC ultra high, even though you have it disabled...

[LeetMiniWheat]

This is about the weirdest thing ever.
Can you try clearing CMOS with the power unplugged?
Every board that uses fixed (not auto and no offsets) voltages, and with C1e, C3, C6 disabled, should NOT be raising the voltage at load for any reason; vdroop (loadline slope) is supposed to lower the vcore at load on an even slope and it's been that way ever since the Pentium 4 was introduced and VRD 8.x, instead of the P3's always fixed voltage.

If you're getting increased vcore at load, and you're not using offsets and not using LLC, then either the BIOS or the board is defective somehow, because that's not how vdroop is supposed to work.
It's like the board is giving you LLC ultra high, even though you have it disabled...

this baffles me too because this isn't how vdroop is supposed to work. but this happens only with offset voltages (and I don't believe C-states are related). manual voltage is alright but nowhere near as tight as gigabyte. (but gigabyte's have their own slew of problems...) I can stay between 1.352v and 1.364v with manual mode, but offset ranges between 1.336 and 1.40v during loading and about 1.072 idle. (read from CPUZ VID)

I'm starting to think it's not even worth it to keep offset on anymore and just let it stay at 1.35-1.36 all the time. though I am rock stable for almost a week now, but my chip is getting more voltage than it needs. Asus needs to fix their offset to work better like gigabyte. (and the LLC) though I don't regret trading in my P67A-UD4 for P8P67 Deluxe. Asus definitely made a superior product this time around. but every P67 board has it's own set of issues.

<Edited multiple times>

[Peter949]

First time Asus mobo user here also and probably might use EVGA next if they get around to releasing their boards. Asus just has too many problems.

[LeetMiniWheat]

First time Asus mobo user here also and probably might use EVGA next if they get around to releasing their boards. Asus just has too many problems.

I wouldn't count on that. Rumor has it, EVGA's mobo team up and left. (multiple sources from this) and then mysteriously another brand started popping up with almost identical designs to EVGA's X58 boards. so apparently they jumped ship. (I forget what brand it was though) I was a loyal EVGA board customer for awhile myself until I realized this. and the P67 delay only proves this further. if you look at their photos of it, there's no innovation compared to the last P55 model. it's just a P55 remake with a P67 chip slapped in, nowonder they can't get it to work right in order to release a product.

[Sauron_Vie]

i am a little bit confused to read that some members have mich trouble with the asus boards., so i can only spreak for myself, and i have really not a trace of problems with the pro versions. meanwhile i changed to the b3 rev.. also no problems at all.
maybe it depends also at my combo CPU - RAM - MOBO that mine is running without any probs. some friends of mine bought at first the giga´s.. meanwhile they changed to asus cause the gigas are also not exactly brilliant either, are they ??

[ghmz]

raise ur vcore a little

Didn't help. Increase from +0.080 to 0.090. Crashed during the night. I know it's getting enough voltage for a 45x.

[Sauron_Vie]

do u guys use only the offset function or r u using offset & additional turbo vid ?

[Leeghoofd]

Didn't help. Increase from +0.080 to 0.090. Crashed during the night. I know it's getting enough voltage for a 45x.

Do you call that an increase, 0.010 ? I give up... no idea why you run LLC off... just leave it on AUTO... try upping VCCIO to 1.15. The dimms are 100% okay ?

[LeetMiniWheat]

i am a little bit confused to read that some members have mich trouble with the asus boards., so i can only spreak for myself, and i have really not a trace of problems with the pro versions. meanwhile i changed to the b3 rev.. also no problems at all.
maybe it depends also at my combo CPU - RAM - MOBO that mine is running without any probs. some friends of mine bought at first the giga´s.. meanwhile they changed to asus cause the gigas are also not exactly brilliant either, are they ??

I switched from Gigabyte p67 to Asus p67 too and I agree they have problems. but none are perfect. they all have problems of their own. but Asus is still superior in features.

[Darkstormz]

I´ve got some serious issues going on my P8P67 WS Revolution.

When I run 3DMark Vantage / 3DMark 11, I get a message "Nvidia display driver stopped responding and has recovered"

Tried everything, GTX480, GTX580 and 9600GT card. Still same problem.
Tried to switch PSU from Corsair HX620 -> Corsair AX850. Still same problem.
Re-installed Win7 Ultimate SP1 and still got the same problem.

Sometimes I can run 3DMark for 30sec, sometimes a couple of minutes.
The most strange thing is: Can run Unigine Heaven benchmark for 30min without any trouble.

Everything is at stock 2600K, 16GB Vengeance 1866MHz...

Anyone?

I did a clean install without SP1, and the issue is gone. But as soon as I install SP1, boom! Display driver stopped blabla....

[crossg]

I did a clean install without SP1, and the issue is gone. But as soon as I install SP1, boom! Display driver stopped blabla....

I've seen that issue 5 times today, but it's because I'm pushing my Cpu and Gpu's to their limits and 3DMK 11 really stresses the card more than the cpu. Happens on my 980X rig as well so it's probably vid card related. Right now I'm using beta 270.51 drivers but any driver so far has done the same when pushed.

[Sauron_Vie]

I switched from Gigabyte p67 to Asus p67 too and I agree they have problems. but none are perfect. they all have problems of their own. but Asus is still superior in features.

u r right and i think asus is one of the stablest of the big 4

Do you call that an increase, 0.010 ? I give up... no idea why you run LLC off... just leave it on AUTO... try upping VCCIO to 1.15. The dimms are 100% okay ?

0,08 or 0.09 is for my taste much too much. can´t understand why some guys pump so much voltage into. i can run mine for 4,6ghz with 0,02 and a little add turbo vid