[bobekk]

If you're on 5 GHz all core overclock, then you're clock stretching, you've set voltage too low and losing over 1k points.

setting 5Ghz all core resulted in some horrid voltage and temps spikes that resulted in throttling (using arctic freezer ii 360) so after 2 days of fighting i gave up and stopped with :
1C: 50x, 2C: 50x, 3C: 50x, 4C: 49x, 5C: 49x, 6C: 49x, 7C: 49x, 8C: 49x
is that also a reason for HWinfo to rep IA: Max Turbo Limit - YES since it's not all core ?

 

[BroadPwns]

setting 5Ghz all core resulted in some horrid voltage and temps spikes that resulted in throttling (using arctic freezer ii 360) so after 2 days of fighting i gave up and stopped with :
1C: 50x, 2C: 50x, 3C: 50x, 4C: 49x, 5C: 49x, 6C: 49x, 7C: 49x, 8C: 49x
is that also a reason for HWinfo to rep IA: Max Turbo Limit - YES since it's not all core ?

Get a contact frame to fix Intels terribleness, or sand it down. Max Turbo limit is just a flag, that the CPU runs at maximum allowable turbo ratio.

 

[bobekk]

Get a contact frame to fix Intels terribleness, or sand it down. Max Turbo limit is just a flag, that the CPU runs at maximum allowable turbo ratio.

well i didnt see any problems with my mounting tbh paste was spreaded evenly(no odd singe cores temps) ..checked again 2 days ago when artic sended me fix kit after month of waiting

EDIT: Ordered frame anyway...

 

[Laserco]

Is anyone seeing slightly higher CPU temps lately? I have noticed idle temps for package CPU around 34 (Hwinfo) after several Windows 11 Dev updates. Used to be around 30c. Or could 9 month old MSI Mag Corelquid 360 cooler be failing? Wouldnt increase in fan speed offset higher temp? Other temps such as PCH, DRIVES, ram, have not changed. Tried using stock bios settings with only XMP oc. It helped a little maybe lower by 3c. Would it need repaste after only 9 months? Thanks for any input.

 

[KyKo.]

Is anyone seeing slightly higher CPU temps lately? I have noticed idle temps for package CPU around 34 (Hwinfo) after several Windows 11 Dev updates. Used to be around 30c. Or could 9 month old MSI Mag Corelquid 360 cooler be failing? Wouldnt increase in fan speed offset higher temp? Other temps such as PCH, DRIVES, ram, have not changed. Tried using stock bios settings with only XMP oc. It helped a little maybe lower by 3c. Would it need repaste after only 9 months? Thanks for any input.

MSI MAG z690 tomahawk wifi ddr4 with noctua NH-U12A .WINDOWS 11 with latest updates. 4x8 GB PATRIOT VIPER STEEL 4400CL19 OC 4000 CL15. CPU core voltage mode adaptive 1.390v LLC2

 

[BroadPwns]

Is anyone seeing slightly higher CPU temps lately? I have noticed idle temps for package CPU around 34 (Hwinfo) after several Windows 11 Dev updates. Used to be around 30c. Or could 9 month old MSI Mag Corelquid 360 cooler be failing? Wouldnt increase in fan speed offset higher temp? Other temps such as PCH, DRIVES, ram, have not changed. Tried using stock bios settings with only XMP oc. It helped a little maybe lower by 3c. Would it need repaste after only 9 months? Thanks for any input.

Not that CoreLiquid is even remotely a good unit, tons of earlier batches also had significant pump failure rate. 4 degree difference also sound like paste degradation, if the measurements were taken after same passage of time from cold boot. My best guess is to check vcore readings and power draw (always keep track of power draw when doing temperature testings), microcode updates sometime can create vcore output differences.

 

[Laserco]

Not that CoreLiquid is even remotely a good unit, tons of earlier batches also had significant pump failure rate. 4 degree difference also sound like paste degradation, if the measurements were taken after same passage of time from cold boot. My best guess is to check vcore readings and power draw (always keep track of power draw when doing temperature testings), microcode updates sometime can create vcore output differences.

Appreciate all your thoughts. Have read about cooler issues. vcore and power draw are low. However I have not methodically checked with temp as you describe. Your input (and my gut) tells me re-paste at least or new cooler. Have to wait to work on system either way for a few weeks due to shoulder surgery.

 

[digitalfrost]

I have a 12700K, I kept it at 52x4 and 51x8 OCTVB for the last couple of months the system was stable. PC is running 24/7 so plenty of idle time at night and it was always rock solid.
Vcore under gaming loads is around 1.35v, temps are in check (below 80°). Now suddenly the random idle reboots are back. I already increased Vcore some but it happened again. Very strange phenomenon.

I can run Geekbench all day long without any issue. But then during light load scenarios it just dies after being stable for months.

 

[F7LTHY]

I have a 12700K, I kept it at 52x4 and 51x8 OCTVB for the last couple of months the system was stable. PC is running 24/7 so plenty of idle time at night and it was always rock solid.
Vcore under gaming loads is around 1.35v, temps are in check (below 80°). Now suddenly the random idle reboots are back. I already increased Vcore some but it happened again. Very strange phenomenon.

I can run Geekbench all day long without any issue. But then during light load scenarios it just dies after being stable for months.

What BIOS are You on? Idle instabillity indeed seems like low Vcore but seems to not be it.. Maybe some program got updated so it is spiking in usage and LLC cant handle it..

 

[Ichirou]

I have a 12700K, I kept it at 52x4 and 51x8 OCTVB for the last couple of months the system was stable. PC is running 24/7 so plenty of idle time at night and it was always rock solid.
Vcore under gaming loads is around 1.35v, temps are in check (below 80°). Now suddenly the random idle reboots are back. I already increased Vcore some but it happened again. Very strange phenomenon.

I can run Geekbench all day long without any issue. But then during light load scenarios it just dies after being stable for months.

Any BSODs? If not, likely to be a PSU issue.

 

[markusz2n]

Just did some overclocking on my 12700k after running 50/39 all core with 1.25 Vcore (fixed, LLC3) and got Vcore down to 1.206 V under load and 1.244 V at idle with LLC4 and fixed Voltage at 1.24 in BIOS. Stable in Prime95 with small FFT with no AVX for 7 hours and -3 AVX offset stable aswell.

186 Watt down from 205 Watts with old Vcore. Maxed out at 70°C with Contact frame and big custom loop with fans @ 600 RPM

 

[BroadPwns]

Just did some overclocking on my 12700k after running 50/39 all core with 1.25 Vcore (fixed, LLC3) and got Vcore down to 1.206 V under load and 1.244 V at idle with LLC4 and fixed Voltage at 1.24 in BIOS. Stable in Prime95 with small FFT with no AVX for 7 hours and -3 AVX offset stable aswell.

186 Watt down from 205 Watts with old Vcore. Maxed out at 70°C with Contact frame and big custom loop with fans @ 600 RPM

AVX offset doesn't count at all. Remember to up cache ratio as well, I've mine set to 4.4 GHz and for some reason, it doesn't drop under e core load, which should not be happening lol.

 

[markusz2n]

AVX offset doesn't count at all. Remember to up cache ratio as well, I've mine set to 4.4 GHz and for some reason, it doesn't drop under e core load, which should not be happening lol.

I got cache ratio at 4.2 GHz (havent tried any higher). And what exactly do you mean with "AVX offset doesn't count at all" ? The Prime95 7 hours stable is without AVX, so cores at 50/39 and then I tested some AVX Prime to see how much offset I need.

 

[Ichirou]

I got cache ratio at 4.2 GHz (havent tried any higher). And what exactly do you mean with "AVX offset doesn't count at all" ? The Prime95 7 hours stable is without AVX, so cores at 50/39 and then I tested some AVX Prime to see how much offset I need.

Testing with an offset is a relic of the past like 10 years ago.
Now, most programs/games even use AVX.

 

[BroadPwns]

Testing with an offset is a relic of the past like 10 years ago.
Now, most programs/games even use AVX.

This and any kind of AVX load triggers offset for entiery of cores, so even a light, single threaded AVX load will lower core ratio for entire CPU.

 

[markusz2n]

Testing with an offset is a relic of the past like 10 years ago.
Now, most programs/games even use AVX.

Oh, ok. I haven't ever seen my clocks drop from 5 GHz to the theoretical 4.7 GHz, so I never thought about it... good to know, I'll try to get it running with 0 AVX offset then...

Haven't overclocked to the limit of stability (with low voltages for power saving) in quite some time, so for me AVX offsets were never an issue for the games I played in the past. Good to know that has changed.

 

[Dash8Q4]

Just picked up this board. What’s the most up to date stable bios for it? I’ll be putting in two 16gb TeamForce 3600C14 sticks so any help in getting the best out of these will be greatly appreciated.
cheers. Cpu will be 12900K

 

[digitalfrost]

Just picked up this board. What’s the most up to date stable bios for it? I’ll be putting in two 16gb TeamForce 3600C14 sticks so any help in getting the best out of these will be greatly appreciated.
cheers. Cpu will be 12900K

I had good results with 1.70 so far. Stay away from 1.80. Latest is 1.81 which I haven't tested yet. Also a lot of people here a very fond the old 1.24U2.

Any BSODs? If not, likely to be a PSU issue.

No BSOD, the screen goes black, then it hard reboots and I am greeted with the OC has failed BIOS message. PSU is rather new Corsair 850W.

What BIOS are You on? Idle instabillity indeed seems like low Vcore but seems to not be it.. Maybe some program got updated so it is spiking in usage and LLC cant handle it..

1.70. I might try 1.81...


e: Found something interesting in the system log. Before this happens, I always have this in the log:

The speed of processor 16 in group 0 is being limited by system firmware. The processor has been in this reduced performance state for 71 seconds since the last report.
The speed of processor 17 in group 0 is being limited by system firmware. The processor has been in this reduced performance state for 71 seconds since the last report.
The speed of processor 18 in group 0 is being limited by system firmware. The processor has been in this reduced performance state for 71 seconds since the last report.
The speed of processor 19 in group 0 is being limited by system firmware. The processor has been in this reduced performance state for 71 seconds since the last report.

These are the E-cores no? I will reduce E-Core clock by 100Mhz and see if that helps...

 

[Laserco]

I had good results with 1.70 so far. Stay away from 1.80. Latest is 1.81 which I haven't tested yet. Also a lot of people here a very fond the old 1.24U2.

Actually 1.80 is latest stable bios and 1.81 was a prior beta version. numbering is counterintuitive. 181 date was early august. 18 stable version Sept 23. Good luck. I have tried 1.24u2 &1.81. not 18 yet. 1.81 less user friendly oc cores labels than older bioses--for me anyway.

 

[digitalfrost]

1.81 is missing a fan speed sensor in HWiNFO for me. I already wrote about the 1.80 VF curve problems in this thread. I go back to 1.70...