Problem with i7-14700K

[plutidan]

Hello community,

I've built a new computer recently and I'm experiencing some problems.

i7-14700K
Noctua NH-D15S cooler
Gigabyte Z790 AORUS Elite AX Rev. 1.x
Corsair CX750M 750W PSU
Corsair VENGEANCE DDR5 64GB 6400MHz CL32 (2x CMK64GX5M2B6400C32)
Samsung 990 PRO NVMe M.2 SSD, 4TB
GTX 1080
Windows 11

Everything worked very well for about a month at the "Optimised" perfdrive setting on BIOS which is the default, and no other changes except enabling XMP (FG version on BIOS at that point). Temperature in the system was optimal, with the CPU at around 35-50C in standard Windows tasks, increasing a bit when doing my work (Audio and Graphics related). 

Slowly, I started getting a couple of random "Clock_Watchdog_Timeout" blue screens which I didn't pay much attention to. This continued with increasing rate the next days, till the system was unbootable, as the blue screen would come during boot. Sometimes it would last till the Windows boot log-in screen, rarely until after log-in where it would crash within seconds, but most times it would crash earlier, often just before the log-in screen.

My first steps were to remove everything and leave just the M/B, CPU, RAM, SSD and internal VGA, which didn't have any results. I have reset the bios to factory defaults, disabling XMP, removed and re-inserted the RAM in various configurations (one or two modules). All didn't have any result.

After several boots I've managed to get into safe mode, run a repair on Windows, and also run a memory test. Memory test completed without problems and there was no change. In order to remove any suspicion from my Windows installation itself, I've removed the SSD and booted from a Windows 11 installation USB. It crashed with the same reason within few seconds, again and again, making it clear there is nothing wrong with my Windows 11 installation.

After a long day of playing with the BIOS settings, I found out that with the perfdrive "e-core disable" option the system will run perfectly good without any hiccups which is the option I've been using the past weeks to at least do my work. Obviously I would want my CPU to run properly with all its cores.

Yesterday I've spent the day giving this another go, trying various things. One thing I've tried was to run with just one of the EPS connectors (the M/B has input for two). Since my PSU doesn't have two EPS connectors, I'm using a converter from four PATA to EPS for the second, so I wanted to make sure there is no problem there. I've tried with either of those just to confirm they work well, and indeed the system has no problems to start and work with either. I've reconnected both.

The next suspicion, despite what seems to be very good temperature performance, was the thermal paste between the CPU and the cooler. I've removed the cooler, cleaned up the older paste and applied new one, this time in a different application pattern (a larger central blob instead a thin square and an x). This didn't make any difference. I've updated the M/B Bios to the latest version (FJd), which includes an option for "Intel Default Settings". Again, there was no difference, the system would crash in the same fashion and timing.

Lastly, I've started trying various BIOS settings, enabling the e-cores but running the CPU at lower multipliers for example. It's stable at 20x with all cores enables, but of course that's not ideal. I've used both Prime95 and multi-core Cinebench tests to stress-test the CPU. I've also tried adjusting voltages a bit from some things I've read online, something that again didn't bring any luck. Since I'm not experienced with voltage tweaking, I didn't give this much time.

Most importantly, the fact that the PC was running perfectly stable for about a month, and having excluded something being wrong with the Windows installation, indicates that something happened at the hardware level, rather than some setting on BIOS that needs to be changed.

The CPU with the 8 p-cores only is powerful enough that I can do my work, but the reduction in performance is often perceivable since the applications I'm using are multi-core heavy, so it's sad to see it running without the e-cores. The only thing that I haven't tried yet was to buy a new PSU (I've been using this for 4 years on an old dual XEON setup without problems).

I'm currently out of options and ideas, so I've decided to write here hoping for some advice that will help me determine what's going on, and hopefully find a solution.

Thanks for reading!

[JeanetteC_Intel]

Hello plutidan,

Thank you for posting in Intel Communities.

I highly appreciate your sharing your system information and all the steps you've taken in trying to resolve the issue, however, there are some missing keys that I would like to clarify to know more about your system details and fully understand where the issue truly lies.

1. Please confirm if this is the actual system device that you have (GA Z790 AORUS ELITE AX (rev. 1.0)). 

2. If the MB specifications link we got is correct, the latest BIOS version available that I can see is F12d while the version you mentioned was FJd, the reason why I am trying to confirm if the link I pulled up is correct.

3. Moreover, if you may, please download Intel® System Support Utility for Windows* software. When the download is complete, launch SSU.exe.

1. Scan: Check the box Everything.
2. Click Scan.
3. Review: When finished scanning, click Next.
4. Click Save (as .txt file).

4. Please also share the power scheme used (Balanced, Power saver, High Performance, Ultimate Performance).

I will wait for your reply, so we can proceed further.

Best regards,

JeanetteC.

Intel® Customer Support Technician

[plutidan]

Hello Jeanette and many thanks for your reply.

1+2. The M/B is actually the "rev.1.x" version of the GA Z970 AORUS ELITE AX, thus the different bios version.
4. The power scheme used in Windows currently is "Balanced".

I've included the results from SSU.

Thank you for your time.

[JeanetteC_Intel]

Hello plutidan,

Thank you for sharing these details. It is best to coordinate this with our team for further investigation. We will post an update once it's available. 

Best regards,

JeanetteC.

Intel® Customer Support Technician

[JeanetteC_Intel]

Hello plutidan,

Good day to you.

Your motherboard manufacturer released another BIOS version (FJe). Please contact your OEM support to get full assistance with this update and make sure to select the Intel Default Settings. Should the issue still exist after this update, let us know so we may proceed with the next steps.

I will wait for your reply.

Best regards,

JeanetteC.

Intel® Customer Support Technician

[plutidan]

Hello Jeanette. I have updated the BIOS to the latest version (FJe). There is no change in behaviour. I had to revert back to "e-core disable" to be able to boot.

Thank you.

[14700k]

I have the same processor and board as you do. I'm also having issues with BSODs about once every one to two days. They even happen when the PC is idle. I would wake up and find out that it crashed and restarted overnight. When I first build this PC in December it ran well until my first bsod in January. Then a few months later it started to crash much more often. I'm not running any overclocking on it. I've now reduced the clock frequency to 48x using Intel xtu and will now wait and see if it still crashes. If it does then I will disable all ecores. If it still bsod then I will downclock Corsair ddr5 2x16gb ram to 4200 (currently 4800). I'm also on beta bios fjd (this bios didn't help it still crashes) and will also keep trying the new bioses. I see beta bios fje just came out. Keep me updated on how you go. All my PC components are only 8 months old.

[JeanetteC_Intel]

Hi plutidan,

Thank you for sharing this update. Kindly check your email on how to proceed.

Best regards,

JeanetteC.

Intel Customer Support Technician

[JeanetteC_Intel]

Hello 14700k,

Thank you for posting in Intel Communities.

I understand that you are experiencing the same issue with the same processor and motherboard. I kindly request that you create a new thread for your case. This will allow us to focus on your specific problem and provide you with better assistance. 

Best regards,

Jeanette C.

Intel® Customer Support Technician

[JeanetteC_Intel]

Hello Reaper,

As discussed over email, I will now proceed in closing this thread.

Should you need assistance in the future, please submit a new question as this thread will no longer be monitored.

Sincerely,

JeanetteC.

Intel Customer Support Technician

[plutidan]

Although this thread is now resolved thanks to Intel support's help and my new CPU is in post, I wanted to point out some new information that might be helpful to other people experiencing the same problem and for future reference.

After understanding the issue better, I realised that it was very likely that a specific core was responsible for causing the stability issue, rather than some other mystical setting. Since the CPU was operating well with all e-cores disabled, the issue would obviously be one or more e-cores. So I went to BIOS and manually enabled each of the e-cores one by one, booting into Windows and running stress tests after each core was enabled. I found out indeed that it was a single e-core (#2 specifically) that was causing all these problems. When enabled, Windows would crash at boot. With it disabled and all other cores enabled, the system will run perfectly good. It seems like it was damaged at some point within the first month of operation for whatever reason. Whether the older BIOS version and how it was working the CPU on its default settings was problematic causing this problem on the CPU core, or this unit just happened to have a fault is unknown to me, but it doesn't matter at this point.

Currently running like this (8 p-cores and 11 e-cores) on the latest BIOS version at "Intel default settings" for the past couple of days, and everything is very stable, including stress tests, benchmarks etc. I will of course replace the CPU when the new CPU arrives, but it's good that I now know where exactly the problem was, and I hope this information can be helpful.

Thanks.

[CoolBook]

@plutidan 

Well performed investigation sir!