I decided to make a relatively long summary video of my recent LN2 session and all of the results and what happened, instead of making multiple short and boring clips of each hardware that was tested.
The Cascade Lake-X 10980XE, which you have been seen me testing on my channel ended up being relatively bad on cold. Being awesome on water cooling, it stops at around 5.6-5.7GHz for multicore tests on LN2, so it is not very good. Considering that my better 9980XE got the same tests done at over 5.8GHz, the target is much higher than where the 10980XE stands at. I learned a lot about lapping overall, and based on thermal paste spread the contact was awesome.
The 2080 Ti K|NGP|N LN2 tests ended up being just multiple mount tests, with and without the GPU Inferno back plate. I was working my way up to get colder temperatures, but for some reason the 12V input fuse blew on the card so I can't use it anymore before it is repaired. So far the GPU Inferno heater plate did not help cracking wise, but it really keeps condensation issues away. All of the thermal pastes stop at a certain point, and even the very dry Thermalright TFX was very bad on LN2.
Only X58 testing went somewhat OK. I managed to get multiple top scores with Core i7 930, 970 and X5675. I also tested the rare A0 stepping version of Gulftown 980X, but it was quite horrible on LN2 despite the cold temperatures it could run. Also my 7.2GHz 980X is pretty much done, what helped with the Q3QP degraded the retail very very quickly. I took the risk and disabled the physical cores from 6 to 2 with hopes of gaining more CPU frequency, but it degrades the CPU really fast. So I would not try this on any retail Gulftown as I have now seen it myself. On the Q3QP it helped a little bit, but even so it was still way too far from record capable speeds.
I wanted to try the DMI Pin mod on my last retail 7740X as I now got the required mod points to supply the lacking voltage manually from the motherboard. Option was to either use VCC PLL or CPU PLL. CPU PLL is a safer option as we always run that at no higher than 1.8v, usually 1.6-1.8v. Some CPUs need really high VCC PLL toget rid of cold bug, so running 2.8v+ on DMI would surely be dangerous for the CPU. Sadly I damaged the target cap while soldering, so it kinda failed haha.
Be sure to check my 3DMark Port Royal record video using the 2080 Ti K|NGP|N: https://www.youtube.com/watch?v=Mh5dIG1Lmq0
X58 Bloomfield scores can be found on HWBOT: https://hwbot.org/
