I wouldn’t say I did that much, just identified the problems, repeatedly tried to tell GB where the problems were (which I’ll add even for a cash-strapped R&D fixing would not be difficult for 90% of the problems if they had a competent R&D department) and cleaned a bunch of things up and optimised some other stuff. To me thats not really doing a whole lot but I get it is still going way above and beyond what I could have done.
Still your bios isnt crashing all the time like the ‘genuine’ gigabitter does
And after a last contact with their costumersupport it showed me that they only bring up
false hope even the us-forum gives false assumptions regarding timetables of a new bios
In wich i lost all faith , so i bought a taichi and a 2700 wich runs fine for the highend things
And that Gigash*t only for for internet and simple server , VM and so
A 1700 OC’ed to 3500Mhz low power consumption , all thx to Gigablabla who’s x370 K7 is a nono
Gaming5 and k7 are hysterically bad, in fact all the gigabyte x370 boards are, since ive used the x470 gaming7 wifi board ive had not a single issue everything works perfect, but gigashites poor reputation from the last 18 months is putting a lot of people off buying their x470 stuff though. you did right to buy the taichi they are awesome boards both the x370 and x470 variants, not much bling but very stable.
So F23 final was released today for X370, the changelog says "Update AGESA 1.0.0.4". Not expecting much else
On the gaming k5 they removed pob settings and ram estability is worst now
… Soo rip…
So Gigabyte released F23.
PinnaclePI AGESA 1.0.0.4
Memory latency is 1ns higher
SecureBoot has been fixed
Still no Windows 10 Options so UI looks stretched
Added CLDO_VDDP in Memory options
Will run few tests, but after F7a, this is their less-broken firmware to date.
@ket : will you make an F23 mod bios ? thx
for the “indiana jones” guys, there’s a f23 mod bios for k7 here : https://www.overclock.net/forum/27580184-post3420.html 
I will make a modded F23, I’ll fit it in when I can might start work on it later today.
thx
Thanks ket
Well my thoughts for the ‘new just broken F23 Gb bios crap’ is this
XMP i know wich settings it has standard (on F23 its a fake collection of numbers NOT read from XMP … its a ‘gb’ or ‘amd’ solution)
Sleep well dont put the system into sleepmode windows(whatever version) or linux , even at plain standard (auto settings ‘gb’) it just goes 8 of 10
into the famous triple beep and it reboots (hence byebye running/slumber session)
Then try to set Vsoc on 1.03 volt (1,02500 volt its floating 1,0008~1,0200) after reboot it sets on 1.1 volt (F10 till F23f had same issue)
due to pumpfans stuck at 800rpm i had to go back to F7b where the sys_fan_pump headers temp scale work flawless ??? (really ‘gb’ did it the good old “thats the right way” )
seems all other version have the oops 1 degree raise so lets do a 400rpm xtra , wich is annoying on a silent pc for www and household tasks
Or the setting for fan5 or fan6 just wont stick ? , else 1 does temp related up/down -+700rpm , the other goes haywire with result 1500rpm or more
Nah its now 18 months of zen , but ‘gb’ has not seen a enlightened path
hi just updated to your new bios and very good job there i found it works very well ,but its down clocked cpu from 3900 to 3850 -3800 and led code has changed from 24 to 33 i am using 2600 not overclocked
Modded F23 will take a bit longer than planned as I can’t port over string tables to F23, looks like a specific change GB have implemented to deterre modders. Apparently they REALLY don’t like us (probably specifically me) fixing up their terrible firmware. I could do with someone getting screen captures from the stock F23 then linking me to them in an archive so I can see which options are exposed and which aren’t. EDIT: If someone can test HPET with F23 that would be good too don’t want a repeat of last time where GB managed to break HPET.
hello ket, just install your bios f22b on my k7 (no problem for flashing…thanks
but, I don’t find cldo-cddp setting in.
normal ?
thanks
nb : with GDM disable with ram@3266 : boot, but bios freeze…
must enable to work properly.
nb2 : thanks to you for this bios. will test this now.
so, try your last F23f mod bios but no cldo_vddp too.
and think i find a problem.
look at this screen :
bad cad strenght and procodt…don’t know if the value I enter in the bios is really used.
?? 
find another problem in your f23f
if i enter vsoc 1.04375, it’s ok, i’ve got +/- this in windows
but if I enter 1.05 in bios, i’ve got 1.08-1.09 in windows
@floyd
When i put 1.02500 for Vsoc its 1.0200 when i add a notch up 1.031250 its after save/restart 1.1V in bios in windows or linux its 1.2~1.3 , same happens in the last release F23
After F10 and its ‘gb’ miracles i stopped updateting till F23f from ket
And try this Ryzen DRAM Calculator 1.3.1 for the Cad_Bus_Block normaly u should stay between 20-24
@Ket
Got the bios F23 screenies here "https://ufile.io/u70gn"
PS what was the Hpet error back then , vaguely remeber if on the bios would go bonkers
it seems working now tho got it on and only some pieces of of the gearbox came out 
ok for vsoc, so you’ve got the same problem from f23f from ket ?
Yes if i put 1.02500 it ranges from 1.0200~1.0320 according to Hwinfo in bios its 1.0008~1.0200 with llc on 'high for vsoc
Same as F23
Cheers for that, will get back to modding the firmware later today hopefully. HPET issue was that with it disabled the board entered an infinite reboot loop after the first boot to windows requiring HPET to be set back to "Enabled".
Thats normal, GB X370 boards have awful voltage stability. Regarding the drive strength RTC just doesn’t always read things accurately from my experiences with it. You’re likely better off leaving drive strength on "Auto" anyway rather than trying to fine tune it. CLDO_VDDP also went awol some time ago, I found the best people I could at GB to refer all of my findings to and putting that option back was one of them.