ok.
so which values is right for strength ?
bios or RTC ?
because 120 ohms is very big on auto compare to ryzen dram calculator
thanks 
As a rule whatever you set in the firmware is what you should take in to account the only time what you set in the firmware would be wrong is if GB fudged something.
ok
in all case, your bios is better yhan gigabyte.
with gigabyte bios > F10, i couldnât boot more than 3000 on the ram
with your f23f, iâm @ 3280 
thanks a lot
what is latest GA-AX370-Gaming-5 bios?
or better: when will F23 with 1.0.0.4 be ready?
X370 Gaming 5 F23fM-D is still 1.0.0.2
Aah yes well its fixed now had it on makeing the screenies
Also tried with hpet disabled
yes, thatâs why when I see 120 in RTC, I was frightened !
I enter 20 in bios, but itâs always 120 in RTC
Hi Ket,
HPET has been fixed in F23
EDIT: Will send later screen captures of F23 UEFI
yes, thatâs why when I see 120 in RTC, I was frightened !
I enter 20 in bios, but itâs always 120 in RTC
I know seems to not working in rtc , ive set 24 in bios and also on all four 120 in rtc
with the Gigabyte F7b i have before, RTC was ok.
itâs with ket bios thatâs it doesnât work
No its a âgbâ feature , we see it as a new bug thx to the âgbâ biosteam
Its in the F23f and the F23 from âgbâ
Same as tWRRD in memory-settings of F23 after restart and with XMP its suddely a - and not the value of 3
So again âgbâ likes to mess around and acts like âwe do the best we can to fuck this upâ
From what Iâve seen GB donât bother checking the differences between AGESA revisions and just slap things together which is why so many things get borked.
and you canât correct this fuck*** errors from gigabyte in your modded bios ?
@Floyd
As example from the F23 âgbâ bios wether u OC by P-state or old fashion numbersetting even default cpu (fillin the numbers instead of auto) , sleepmode in windows or linux
just makes the mobo reset the cpu settings
Memory seems not to be afected , altho the screenies gives the âgb biosboysâ a glimpse of whats wrong this time
Left settings are the ones from the XMP profile , the right ones are suddenly differnt after a reboot while they where correct (once upon a time )
Even voltages get screwedup with auto like DDRVtt a/b it walks between 0.660 and 0.684 , while dram voltage is steady 1.356v
And then Vsoc in bios set at 1.031250 but its stuck at 1.104
https://ufile.io/ws04p
So each time theres a new agesa , u can be sure theres trouble and weird bugs
my english is very bad, so i donât understand exactly what you say. 
all that i know, is that i succeed to push my ram @3300 now but the vsoc enter in the bios is not exactly the same in windows
i grope, but i succeed to pass over all the bug, and thatâs the principal.
X370 Gaming 5 F23fM-D
https://mega.nz/#!z04y1SBA!sXBq-RhR2tWqiâŚDNyv1pe8Wd1_WBo
is still AGESA 1.0.0.2
Well today iam back at F7b for the simple reason from F22 till F23 it seems that under linux u get spammed by the kernellogs about AER problems from the PCIE & NVME
2018-08-30T01:00:04.196029+02:00 linux-5wdk kernel: [ 20.232653] nvme 0000:08:00.0: PCIe Bus Error: severity=Corrected, type=Physical Layer, id=0800(Transmitter ID)
2018-08-30T01:00:04.196032+02:00 linux-5wdk kernel: [ 20.232656] nvme 0000:08:00.0: device [144d:a804] error status/mask=00001081/00006000
2018-08-30T01:00:04.196034+02:00 linux-5wdk kernel: [ 20.232658] nvme 0000:08:00.0: [ 0] Receiver Error (First)
2018-08-30T01:00:04.196037+02:00 linux-5wdk kernel: [ 20.232660] nvme 0000:08:00.0: [ 7] Bad DLLP
2018-08-30T01:00:04.196039+02:00 linux-5wdk kernel: [ 20.232662] nvme 0000:08:00.0: [12] Replay Timer Timeout
2018-08-30T01:00:04.196042+02:00 linux-5wdk kernel: [ 20.243445] pcieport 0000:00:03.2: AER: Multiple Corrected error received: id=0000
2018-08-30T01:00:04.196044+02:00 linux-5wdk kernel: [ 20.243460] pcieport 0000:00:03.2: PCIe Bus Error: severity=Corrected, type=Data Link Layer, id=001a(Transmitter ID)
2018-08-30T01:00:04.196047+02:00 linux-5wdk kernel: [ 20.243463] pcieport 0000:00:03.2: device [1022:1453] error status/mask=00001040/00006000
2018-08-30T01:00:04.196049+02:00 linux-5wdk kernel: [ 20.243466] pcieport 0000:00:03.2: [ 6] Bad TLP
2018-08-30T01:00:04.196052+02:00 linux-5wdk kernel: [ 20.243468] pcieport 0000:00:03.2: [12] Replay Timer Timeout
2018-08-30T01:00:04.196054+02:00 linux-5wdk kernel: [ 20.243478] nvme 0000:08:00.0: PCIe Bus Error: severity=Corrected, type=Physical Layer, id=0800(Transmitter ID)
2018-08-30T01:00:04.196057+02:00 linux-5wdk kernel: [ 20.243481] nvme 0000:08:00.0: device [144d:a804] error status/mask=00001081/00006000
2018-08-30T01:00:04.196060+02:00 linux-5wdk kernel: [ 20.243483] nvme 0000:08:00.0: [ 0] Receiver Error (First)
2018-08-30T01:00:04.196062+02:00 linux-5wdk kernel: [ 20.243485] nvme 0000:08:00.0: [ 7] Bad DLLP
2018-08-30T01:00:04.196065+02:00 linux-5wdk kernel: [ 20.243487] nvme 0000:08:00.0: [12] Replay Timer Timeout
2018-08-30T01:00:04.196110+02:00 linux-5wdk kernel: [ 20.255422] pcieport 0000:00:03.2: AER: Multiple Corrected error received: id=0000
2018-08-30T01:00:04.196112+02:00 linux-5wdk kernel: [ 20.255438] pcieport 0000:00:03.2: PCIe Bus Error: severity=Corrected, type=Data Link Layer, id=001a(Transmitter ID)
2018-08-30T01:00:04.196115+02:00 linux-5wdk kernel: [ 20.255440] pcieport 0000:00:03.2: device [1022:1453] error status/mask=00001040/00006000
2018-08-30T01:00:04.196118+02:00 linux-5wdk kernel: [ 20.255443] pcieport 0000:00:03.2: [ 6] Bad TLP
2018-08-30T01:00:04.196120+02:00 linux-5wdk kernel: [ 20.255445] pcieport 0000:00:03.2: [12] Replay Timer Timeout
2018-08-30T01:00:04.196123+02:00 linux-5wdk kernel: [ 20.255455] nvme 0000:08:00.0: PCIe Bus Error: severity=Corrected, type=Physical Layer, id=0800(Transmitter ID)
2018-08-30T01:00:04.196125+02:00 linux-5wdk kernel: [ 20.255457] nvme 0000:08:00.0: device [144d:a804] error status/mask=00001081/00006000
Its solveable by u can guess disabling the messages âduhâ
So i just flipped the bios switch and 3 things where visible ,faster boot no AER messages and overal faster system linux and win10 (ok F7b no spectre bug walkaround)
So for me this closes the issue âgb gaming K7â it may happen but i lost that faith âgbâ comes with a bios like F7b (wich has also few flaws like P-state and non X cpuâs)
But for a internet and dedicated gameserver its a better solution then the so called effort of 'gbâ
Maybe you Ket can solve a few things with F23 but soon the âgbâ biosfloppers come with a new fluke with agesa 1.0.0.5 so its like âa never ending storyâ
Firstly, thanks so much, ket, for doing this! I really appreciate the effort youâre putting in to fix Gigabyteâs slipshod work. Is there anything I can do to help? It sounds like you already got the needed screenshots of F23, yes? Let me know if thereâs anything I can do to assist!
My previous system was based on the GA-EP45-UD3R, and I really liked that board and appreciated its quality and longevity. That experience is what lead me to choose Gigabyte again when finally upgrading. Itâs disappointing that their BIOS/UEFI quality has declined so much.
Anyway, thank you so much, again. I had planned to keep this board through a Zen2-based CPU upgrade down the line, and your work is giving me hope that I might actually be able to do that with my AX370-GAMING K7.
Cheers!
Is anyone here running B-Die 2x16GB dual rank memory (exact model: G-Skill F4-3200c15D-32GVK)?
I´m getting crazy with my K7, I can run my Ryzen 1700 stable at 3,9 GHz @ 1.28v or even 4 GHz at 1.375V -1.4 , but I just can´t get my RAM stable at all.
So typhoon burner is telling me my RAM modules are b-die and I tried nearly everything to get them stable at 3200. Even @2933 it´s unstable with relatively loose timings. I also tried populating different RAM Slots, but it got just worse.
Things I´ve tried so far:
Different procODT Values mostly 60, 68.6 and 80. Best resluts with 68.6. 60 would simply not even boot on most bios
All kind auf RTT_NOM, RTT_WR and RTT_Park combinations most succssessfull was RTT_NOM disabled *_WR RZQ/3 and lastly PARK RZQ/1
CAD_BUS Settings made nearly no difference between 20/20/20/20 and 24/24/24/24 all other values caused error after some minutes
Different CLDO_VDDP Voltages seems to have no impact too.
I flashed every Bios from F7a to offically F23 and modded F23 one and there are slight differences but still can´t get it to 3200 MHz. Booting is no problem at all but the most of the time Memtest ist throwing errors already after 10 min
Other related Settings:
DRAM Voltage: 1.35-1.4
Vcore -0.725 offset
LLC both on Turbo ( gives me the most stable voltages)
Screenshot of ryzen timing checker with all values
GOODBOI, try using Ryzen DRAM Calculator to figure out the timings for your memory and set them all manually with XMP off (GB screwed up XMP a while ago), especially tRFC - it looks to be set way too low (GB firmware always shows 312 as default which is way off).
I have 2x16GB 3200 CL16 Hynix memory running at 2933 CL14 on my Ryzen 1700, F22bM-Dv2, and Iâve found that setting tRC to at least 53, tRAS to at least 33, and tRFC to at least 476 is required. You have different memory than me so YMMV.
Can You Make One For GA-AB350M Gaming 3 Or Can I Just Use GA-AB350 Gaming 3âs BIOS?