Ket,

Thanks for the updated bios for the K-3. Made a few changes and running at 3.2 GHz and 3200 on Ram. Hope the board lasts as long as my Asus Rampage formula with the Qx9650. It is still running strong after all the years of 24/7 service.

MAK

Zitat von ket im Beitrag #548

---

Zitat von anito9999 im Beitrag #547

---

Hi Ket do you think the Bios for my MOBO can be modded

its a Gigabyte GA-AB350M-D3H.

---

Almost certainly but I'm already at capacity with the number of firmwares I'm modding, drivers being modded, and compiling packages with the latest drivers. Someone else would need to step in and agree to pick up on what I don't have the time for.

---

Zitat von ket im Beitrag #569

---

I hope you're all sitting down for this news but... GB have actually done some work for F23d, I'm yet to test how much actual work they have done but for the first time in a damn year advanced memory subtimings are displayed correctly in the UEFI.

---

Zitat von @JSebastian im Beitrag #572

---

***FORCES AMD SATA AHCI Drivers v1.3.1.267***

---

Zitat von POE_UK im Beitrag #585

---

Correct, while the x470 taichi will be every bit as good as the x370 theres absolutely no reason to get one, the topology is almost identical, asrock added bios support for ryzen2 cpu's ages ago, in fact i really cannot see the point in getting an x470 board if you own a good x370

---

Zitat von POE_UK im Beitrag #585

---

Correct, while the x470 taichi will be every bit as good as the x370 theres absolutely no reason to get one, the topology is almost identical, asrock added bios support for ryzen2 cpu's ages ago, in fact i really cannot see the point in getting an x470 board if you own a good x370

---

Well for me its GB presenting the middle finger , the F23d bios just brings in more crap
cant go over 3200Mhz with memory even different sets rated 3200, 3600, 4200Mhz , place it in a different mobo (not Gb) no problem
Geardownmode should be working on 2666 (does also on F7b and yes the above mentioned mem ratings )
But the crapness is the the softbrick is back (or atleast by mine GB-K7) lucky i dont have a battery seated so 5min powerof and dustoff and try again but nomore with this F23 or whats to come
For now i stick with F7b , no problems i cant handle
And to be honest its working great , so thxz ket for trying to fix things (f22d-m bios i believe) but still the bugs are infested deep
So i say stay away from Gigab

Zitat von redrain85 im Beitrag #588

---

Zitat von ket im Beitrag #569

---

I hope you're all sitting down for this news but... GB have actually done some work for F23d, I'm yet to test how much actual work they have done but for the first time in a damn year advanced memory subtimings are displayed correctly in the UEFI.

---

This is what I was hoping for . . . and what I was afraid of, at the same time.

I figured that if Gigabyte was actually going to take the AM4 platform seriously in any way, they'd finally make an effort with the UEFI when X470 debuted. And this might trickle down to X370/B350.

If the UEFI didn't improve now: then we'd know it would never, ever improve.


That being said, a couple of folks on the Gigabyte forum are claiming P-state overclocking might be broken again on F23d.

http://forum.gigabyte.us/post/19378
http://forum.gigabyte.us/post/19388

Zitat von @JSebastian im Beitrag #572

---

***FORCES AMD SATA AHCI Drivers v1.3.1.267***

---

Huh! I didn't know, until now, that the AMD chipset drivers package doesn't install an AMD AHCI driver. They leave it to the generic Microsoft drivers to do the job.

Tempted to give the AMD AHCI driver a try, but does it really make much of a difference in performance? According to this post, the MS and AMD drivers are about equal:

Recommended AHCI/RAID and NVMe Drivers (63)

Zitat von POE_UK im Beitrag #585

---

Correct, while the x470 taichi will be every bit as good as the x370 theres absolutely no reason to get one, the topology is almost identical, asrock added bios support for ryzen2 cpu's ages ago, in fact i really cannot see the point in getting an x470 board if you own a good x370

---

If you choose to believe CanardPC (and I do, in this case): the X470 chipset is really just a rebrand of the X370 chipset. There is no difference between them, whatsoever.

All the differences are in the new Ryzen 2000 series CPUs (IMC, for improved memory support), the new X470 PCBs (better VRMs, and improved memory trace layouts in some cases), and the updated UEFIs.

So if your board already has decent VRMs, good memory overclocking results, and really good UEFI support from the manufacturer: simply replacing your Ryzen 1000 series CPU with a 2000 one will get you almost all the benefits of X470 without needing to replace your board.

---

Zitat von anito9999 im Beitrag #587

---

Zitat von ket im Beitrag #548

---

Zitat von anito9999 im Beitrag #547

---

Hi Ket do you think the Bios for my MOBO can be modded

its a Gigabyte GA-AB350M-D3H.

---

Almost certainly but I'm already at capacity with the number of firmwares I'm modding, drivers being modded, and compiling packages with the latest drivers. Someone else would need to step in and agree to pick up on what I don't have the time for.

---

How bout showing a guide and the tools your using I think most of us are willing to learn and try it ourselves.

---

Zitat von POE_UK im Beitrag #578

---

Agreed ket, somethings wrong with that corsair cooler that belongs to mach5, i can hit 3.75Ghz on the stock cooler! 4ghz should be no problem peaking at 68c on that corsair, there is a setting in CBS/Power States that enables/disables throttling try experimenting with that m8.

---

Zitat von POE_UK im Beitrag #579

---

Good news for you sir, i can indeed confirm the wonderful soundblaster Z soundcard works awesome in the taichi board, i tried it last weekend for some FL studio work. support for the SBZ was added with bios v2.30 which was 10th may 2017

here is a list of compatible soundcards for the asrock: https://pcpartpicker.com/products/sound-...ble_with=fZKhP6

And remember guys its taken gigabyte over a year to sort the most basic of problems, at that rate it will be another year before they are all fixed.

---

Zitat von POE_UK im Beitrag #585

---

Correct, while the x470 taichi will be every bit as good as the x370 theres absolutely no reason to get one, the topology is almost identical, asrock added bios support for ryzen2 cpu's ages ago, in fact i really cannot see the point in getting an x470 board if you own a good x370

---

Check that the pump is at 100% they are inaudible so theres no reason not to let them run 100% really. I'd also consider tipping the rad a few times there might be a trapped air bubble impeding the flow.

Zitat von ket im Beitrag #593

---

Check that the pump is at 100% they are inaudible so theres no reason not to let them run 100% really. I'd also consider tipping the rad a few times there might be a trapped air bubble impeding the flow.

---

Zitat von Mach5 im Beitrag #594

---

Zitat von ket im Beitrag #593

---

Check that the pump is at 100% they are inaudible so theres no reason not to let them run 100% really. I'd also consider tipping the rad a few times there might be a trapped air bubble impeding the flow.

---

The fan header is set to 100%/maximum. Because the rest of my system is normally rather quiet, I can hear the pump when I put it in non-Quiet profile (whatever it is called), so I normally keep it in Quiet profile unless I stress test or encode videos. I have a 200mm and 120mm intake, and a 120mm exhaust fan other than the two on the H100i, so I think it is getting enough air. Also, it is on the exhaust side, so it is drawing (probably warmer) air from inside the case. Would it make such a huge difference? I don't think I can mount it anywhere else though.

---

Is anyone else having problems with F23d firmwares on AX370 K7 boards? After one reboot or two, my machine won't POST.

Official gets stuck with debug code 78, and ket's one with F2 code.

Clear CMOS helps, but I set up stuff, and after one or two reboots, won't POST again.

That looks like you are overclocking with P-States. GB keep breaking that feature so general consensus is just don't use it OC with multipliers instead.

Actually I don't use P-states. I use fixed 39x multiplier and that's it.

What I usually do is:

1) Cpu multiplier to 39X
2) Enable XMP (3200Mhz CL16)
3) Disable geardown mode (CR2)
4) Loadline calibration to High
5) Vcore at 1.38750v
6) Vcore SOC at 1.10v
7) Enable SVM
8) Enable Fast boot
9) Disable CSM
10) Disable HPET (in GB's firmware)

Anything else remain as-is.

I'll try to test settings one by one, hopefully I find the culprit.

Zitat von AndreDVJ im Beitrag #598

---

Actually I don't use P-states. I use fixed 39x multiplier and that's it.

What I usually do is:

1) Cpu multiplier to 39X
2) Enable XMP (3200Mhz CL16)
3) Disable geardown mode (CR2)
4) Loadline calibration to High
5) Vcore at 1.38750v
6) Vcore SOC at 1.10v
7) Enable SVM
8) Enable Fast boot
9) Disable CSM
10) Disable HPET (in GB's firmware)

Anything else remain as-is.

I'll try to test settings one by one, hopefully I find the culprit.

---

Well, I just had the most bizarre experience.

Long ago, I flashed ket's modded F7b to the Main BIOS and left the Backup BIOS on the factory default of F3. I also left the BIOS mode switch in Dual BIOS mode, again the factory default.

When the Main BIOS decided to corrupt itself several months ago, the board automatically switched to the Backup BIOS. Then, just to make sure I was flashing the correct BIOS, I toggled the BIOS switch back and forth between Backup and Main. I made sure the LED indicator showed Main before I re-flashed F7b.


Not too long ago, I decided to set the BIOS mode switch on the K7 to Single BIOS mode in position 2 as a safety precaution. (At least I hope that's right, because I question whether or not the manual is actually correct.) And I've always had the Main BIOS selected for boot, in BIOS switch position 1.

A few days ago, I flashed the Main BIOS with ket's modded F22b. Then, I thought, I'd switch to the Backup BIOS and flash F7b as a safe BIOS to fall back on. To my surprise, the Backup already contained F7b. Hmm. So I guess back when I re-flashed F7b to the Main, and still had Dual BIOS mode enabled, at some point it copied F7b to the Backup and replaced the factory F3. Not happy about that, but okay. Whatever.


Then, yesterday, I noticed the board was suddenly booting up with BIOS F7b. WTF?!

I decided to flick the BIOS switch from position 1 (Main BIOS) to position 2 (Backup BIOS). And what did I get? F22b!

In other words, the board randomly decided to swap BIOSes. I checked the BIOS LED indicator and, yes, it showed Main BIOS was active with F7b. I flicked the switch to Backup, and it was active with F22b.

So I switched back to Main, and flashed it with ket's modded F23d. The Backup still contains F22b, when it should have been F7b.


I'm at a loss for words, here. This should never, ever happen. NEVER! I'm really regretting buying this board. This will be my first, and last, Gigabyte motherboard.

So, bottom line, is that the board seems to have arbitrarily decided to reverse the behaviour of the BIOS selection switch. Selecting Main now sets Backup, and selecting Backup now sets Main. Even though the LED indicators show otherwise.


P.S. On this F23d BIOS, the POST code indicator is frequently changing codes after boot. It usually shows 33, but sometimes 35, and sometimes D3. With prior BIOSes, it would always be a steady 40 unless you restarted, and then it would change to 24. Don't know what to make of this, but it's strange behaviour.

Oh, and just to make it clear: I know these issues have nothing to do with you, ket. You're doing your best to fix Gigabyte's fuck-ups, and we all appreciate it! Your modded BIOSes are far better than the originals.