Pro z690-A ddr4 Stops booting after failed oc

[zvinjavo]

I'm trying to find a stable OC for 12700K. Anytime the system freezes/BSOD's or fails to boot I can't get it to boot anymore. Shorting the clear CMOS jumper does absolutely nothing. The only thing that helps is reflashing with the button.

V11 Bios.
Fast boot is off
XMP is off.

Pretty unstable BIOS, hope they get a fix out soon.

 

[zerijelly]

How did the OC being set in the BIOS?

 

[citay]

Yeah, it will probably be improved in a future BIOS. But i wouldn't really OC the 12700K to begin with. These latest Intel CPUs use 99% of their capabilities by default, with all the different turbo modes. To get the last 1%, you have to increase the voltage too much. The CPUs are already highly binned (pre-selected) by Intel according to their capabilities. So you can't turn a 12700K into a 12900K, the silicon for the 12900K was binned much more by Intel, and then pushed much harder. If you attempt something like that with overclocking, the efficiency will suffer dramatically, because it needs way more voltage and power, and quickly becomes uncoolable. So just leave it as it is, you cannot do much better without big downsides in power consumption.

See here for example, https://www.techpowerup.com/review/intel-core-i7-12700k-alder-lake-12th-gen/22.html

However, enable your XMP. Running the safe JEDEC RAM profile of DDR4-2133 or so with loose timings will cost you massive amounts of performance, because it's bottlenecking the CPU.
Optimizing the RAM to run for example at DDR4-3600 with tight timings will give only benefits with no downsides, unlike CPU OC.

 

[zvinjavo]

I am not attempting any big OC's, on the contrary, I just set it to 5 GHz and then slowly bringing the voltage down (in override mode while testing) to find the sweet spot and have low temperatures (silence is very important). I have disabled XMP as the temps go wild for some reason when it is enabled, so I set the frequency, timings and voltage manually, not exceeding memory specs. I am not overclocking memory ass I run four 16Gb sticks.

What puzzles me is that I can not clear cmos at all. When it goes down the only thing I can do is reflash. I also noticed that sometimes when I reboot it won't boot immediately before I hit the restart button again.

 

[zvinjavo]

Another weird thing. Copying a file to external USB3 HDD spikes all cores to 100%?!?

 

[citay]

Hmm. If you could list all your hardware in detail, i could see if there's something that catches my eye.

Enabling XMP usually raises some secondary voltages like IO and SA2, but that should only cause up to a 10W increase in power draw overall, so it's nothing that should make the temps go wild.
I talked about that here before, for example, https://forum-en.msi.com/index.php?...-but-now-xmp-doesnt-work.367744/#post-2083048

I haven't seen what the MSI BIOS raises those voltages to when you use four dual-rank modules, which is the most challenging setup for the memory system, as i describe in my thread about RAM.
It would be interesting to know what voltages you have active now (you can press F12 to make a screenshot in the BIOS and save to a FAT32 drive, you might wanna convert BMP to PNG later).

The behaviour of the board not being able to recover from failed OC unless you force-flash it is unusual indeed. MSI boards are usually renowned to recover pretty well from failed OC attempts, meaning you can often get by without a CMOS Clear even. But this... weird.

Another weird thing. Copying a file to external USB3 HDD spikes all cores to 100%?!?

It can be something like the antivirus checking the files. Full load is easy to reach, the task manager is quite "dumb" and doesn't show what kind of "effort" the CPU is using for the load. You can have vastly different power consumption with different types of full load. This is visible in HWinfo64 , by the amount of CPU power draw. If you open the sensors and double-click on CPU Package Power, it will show a graph that tracks CPU power draw over time, or you can just keep an eye on the number.