Hi all, I've been away a bit from OC because I was quite busy but just recently getting back into it. Changed to a custom water cooling from a massive D-14 simply because I wanted to play around with custom water cooling. Those all in one solutions really don't give me much satisfaction as oppose to getting all the parts together and assembling them in good old DIY style. So I've successfully put together a 240 rad simple CPU only loop right now, planning to go GPU water cooling as well in the future. (Right now my GPUs don't really need it, nor are there any good water blocks for them because they are heavily customised.)
Anyway now onto the good stuff, I have re-done my previous submission of 5ghz with HT off, stablised with slightly lower vcore, and a fresh new setup @ 5ghz with HT turned on.
Here's the proof for the new HT off setup. As usual, I only feel my computer to be truly stable after a full cycle of custom blend at 18 hours:
(PS: Before anyone ask, I forgotten to delete the line "HT off for better gaming". done my share of testing and I no longer stand by that claim
Bios templates is below:
One thing I've tested getting this stable is try to use lower LLC, and the conclusion is that the minimum stable max load Vcore required does not change no matter what LLC level I use. The frequency of Vcore fluctuation did greatly reduced when I was using a High LLC as oppose to Ultra High or Extreme, however the overall range of the fluctuation (max Vcore recorded minus the lowest Vcore recorded) doesn't change. Somehow I am not seeing over-shoots which some people have mentioned. In the end I decided to tolerate a more variable Vcore by using Extreme LLC Simply because Extreme sounds nice......
Posting this first, uploading the HT on results and bios in a bit.