Thanks for bringing me up to date on the issue dandan. It certainly is an annoying issue for those experiencing it.

Strange how some have issue with the chip and others don't. I agree the chip is unstable regarding temp sensors and maybe it's location is an issue.
As I understand it the I/O chip ITE IT8665E lies somewhere below the GPU and is subject to heat dissipation from the card, Some have had success (not a fix) blowing a desktop fan at the chip.
Maybe the chip becomes unstable when under heat stress, sensors become none operational (fans stop functioning) and software stops working correctly.

As I said , I don't have an issue and don't use AISuite nor any other third party product to interact with Bios. Maybe it's because my GPU is an EVGA Water Cooled Hybrid card which does not expel heat as a none reference card.
I use AIDA64 and HWInfo64 registering correct temp readings (almost identical to my temp probe)
Bios fan curves are manually set in Bios without issue and my case fans and GPU ramp up correctly when under load, so I must be one of the Lucky ones.

dandan1108 wrote:
...it seems like the super I/O chip ITE IT8665E is what makes the board unstable....

Problems related to that chip and external monitoring software has been a non issue for me since October 2018ish for me. I would recommend not using AI Suite III for anything under any circumstance.

As far as WiFi and BT acting the proverbial "foo" there are several things that I have found, depending on Bios rev to a limited extent that can be problematic and it all relates to voltages.

If I am going to be of any help I am going to need to know some things:

CPU
Memory configuration (exact model preferred)
LLC settings (if custom set)
All voltage settings.

Without that info for starters there really isn't much aid I can render but I'll pro tip you with some anecdotal discoveries for my 2700x, 2x16Gb @3400MT 14-14-14-14-28-42 TRFC256, C7H rig.

The wifi and not so much the BT would often be the manifestation that I had my negatvie core offset entirely too low but stable enough to boot if it managed to post. Using the CPU USB port(s) required me to reduce the negative offset. Using an M.2 NVME, same thing. Together, even more reduction to the negative offset was required. It took me a couple of weeks to figure that one out because it was a mild but obvious symptom of a bigger problem with my configuration that, once sorted out, served as a bell-weather for my future tweaking attempts on the CPU using manual or XFR tuning methods alike.

I haven't personally had a problem with HWinfo probing the "super chip" since I want to say October 2018 give or take. Initially the problem was solved seemingly by disallowing access and then later it could stably poll my VRM temps, SVI 2, Fan Extension card fans speeds, etc without potetially crashing the system. I did not note the approximate date when that was fixed in my steno pad.

CAM, while I still had that NZXT pile of garbage, despite countless revisions never could resist trying to poll the super chip and was problematic until I switched out to a different cooler. After HWinfo was fixed CAM was my only source of software problems until the x62's removal.

I would recommend uninstalling AI Suite III and finding the AI3Cleaner executable on these forums at the very least. Asus software has been an abomination since 2004 or so... it's been a long time of problematic sucktitude and although I don't think it is your problem best to be rid of it. Just one old ****er's opinion...

That said, if you hit me back with some details I would be happy to see if I can be of help. Fwiw, my negative offsets with the 2700x worked out like this:

Just a video card: -0.100 was stable, -0.1125 would cause Wifi to go awry + et al
Video card and NVME m.2 usage (single or dual) OR any CPU USB port usage: -0.08125 stable, -0.0875 would be wifi problematic et al
Video card, NVME m.2, and CPU USB in use: -0.0625 stable, -0.08125 blah blah blah.