From RAJA on the ROG forums (original post Crosshair VI Hero: UEFI build 1401 - Page 53 :(
"Direct from Elmor:
AGESA 1006 RC4 official BIOS 1401
Just tested quickly 3600 memory and cold booting, seems good but you guys are going to have to help me test this before we have a judgement. Seems CPU temp reading from SIO now has -20*C offset for XFR enabled CPUs."
Just leaving this here so you don't have to hunt it down buried in forums. It's not posted yet on their site.
well if the 1401 has the bios actually showing -20c then 1401 is correct since 1201 bios is incorrect. well done and thanks. bump ..
if anyone has 4000mhz sticks to try post it here if its achievable with this bios yet.
I'd say the temperatures are pretty close! As for 4000MHz RAM, I'd love to see that too.
The highest I've been able to take it is 3866 with G.Skill's TridentZ RGB RAM (F4-3866C18Q-32GTZR). I don't have the stats to show for it, but there was a noticeable difference in synthetic benchmarks with my 3200MHz TridentZ RGB's scoring faster - both in AIDA64 and Cinebench. To be fair I didn't have a lot of time with the 3866MHz stick as I just had 2 days to play around with timings.
The latency difference between the 3200MHz (14ns) vs the 3866MHz (18ns) consistently showed the most drastic difference in multicore tests for instance on Cinebench (Score #1 is with the 3200MHz RAM, #2 is with 3866MHz):
There was a negligible difference in single-core tests, consistently. If I can find a 32GB x 4000MHz set with a lower latency (<16), I'd gladly pick it up for my #Threadripper build. For now I'm happy with 3200MHz.
I am very curios before i make a new thread if anyone knows for curtain if the new agesa 1006a fixes the problem of using a base overclock on cpu. the problem is/was when any cpu clock over 104.8 from stock 100 would be max to keep the gen 3 3.0 pcie. anything over 104.8 on bclk (fsb) will automatically drop pcie to gen 2 2.0 , even if you have it set too gen3 in eufi bios it would still only be pcie 2.0 ? may need engineering support on this. ?
This still happens. I did a quick test and just looking at Radeon settings showed this:
Even at 105 BCLK this drops down to Gen 2 on the bus settings. Would be nice to get engineering input on this on a new thread rebelyell
I don't like the idea of the drop, but I do understand the need for it. This is why I opt for PState OC'ing as opposed to going through BCLK/modifiers. Updated
thanks for testing that whiskey-foxtrot . with that being said as shown above you may be surprised how many thousands of people never knew this was an issue.
since base clocking is the most common for solid frequency stabilization and experienced oc'ers . aswell as trying to get those ram sticks clocked higher, this is most unacceptable, so it looks like engineers need to get there programming fingers on and build another bios. on that note i also prefer base clocking since it speeds up entire system.