Yep some reworks are needed to ensure no bricked boards as the file @barisuraz sent me caused the board to not POST. I suspect that issue was caused by the microcode updates according to @Lost_N_BIOS the latest Biostar firmware has the new CPU microcode updates anyway. Me and Lost are working on it and should have something soon. Also, Biostar can be sneaky I've noticed them in the past making improvements to their firmware but uploading it with the same filename and release date so it might be worth occasionally downloading the latest file again to check for obvious changes.

The last file I sent you boots fine on my board though. Are you sure you tried the right file? I believe it was named mod_bios.rom

@barisuraz yep. The file with just the AMD CBS changes works fine, but not mod_bios. No rush at the moment anyway me and Lost need to track down some updated UEFI drivers once thats done I'll test the changes Lost makes and perhaps by then I'll have what appears to be some kind of lock on the Biostar setup module and AMD PBS module figured out.

That's very weird the mod_bios works perfectly with my board. Maybe the CBS modded BIOS that didn't work for me might boot for you.

Thats another theory I have which is why I've started looking at the CBS module already in Biostars firmware. I'm having trouble tracking down how to unlock specific features though there should be a line with "True {46 02}" but there isn't.

Test theory, (Random example, intended change/edit in bold) -

Suppress If {0A 82}
QuestionId: 0x14F equals value 0x0 {12 06 4F 01 FF 00}

That change in Hex will change QuestionId: 0x14F equals value 0x0 to QuestionId: 0x14F equals value 0xFF in IFR output, but this I believe it effectively changing the original suppress or question ID to where it's not working/not suppress anymore. But I cannot test this, I only have older AM3 boards.

I wonder if there's any way to improve memory OC on this board. I'm literally getting cancer from trying to reach 3200Mhz with my Hynix MFR ram.

@barisuraz that problem should at least partially resolve itself if Biostar put the time in to fixing all the current issues I reported to them but yes they do need to do a lot more extensive testing with memory. A big problem is that (during my testing anyway) disabling Gear Down which should enable the 2T command rate outright resulted in the board refusing to POST. The added irony here is that I can get better performance results out of my kit of G.Skill RipjawsV at 2T than I can 1T.

Ohh noooooos, back to 2T with DDR4? We're going back in time, should be new command rate 0.5 for DDR4

BIOS sent back with updated vBIOS/GOP

@Lost_N_BIOS The difference between 2T and 1T performance-wise isn't that big, about 3GB/s on write commands and thats about it ;) besides when I'm switching to 2T my kit of RipjawsV (F4-3200C15D-16GVK, Samsung B-Die) is running at almost 3700MHz with only 1.38v vs. 1.5V or so with 1T

Interesting it's not such a big of a difference like it is with DDR3, that voltage difference would make it well worth-while too

Haha I was just checking biostar's URL for the bios updates and found a new unreleased BIOS for our board. It was built yesterday.

https://www.biostar.com.tw/upload/Bios/X37AG919.BST is the link

Edit: I haven't used it yet and wanted to tag you @ket

Time to re-redo our edits now

Nope, turns out I'm an idiot. It's actually last years BIOS. 19/09/ 2 0 1 7 :( @ket @Lost_N_BIOS sorry for the false alarm guys.

lol yep I was about to say thats an ancient firmware :p