I have a Genoa X4 with Nov 94 BIOS date. But I use a VLB EIDE card with it's own BIOS and only have a 2.5GB Quantum Fireball connected so I haven't encountered the disk geometry limits. What IDE interface are you using? Maybe XT-IDE BIOS would help?

Can't Linux bypass BIOS limits by using it's own IDE driver? Maybe it has the same problem that WinNT does, but I've never understood why an OS that doesn't use BIOS functions for disk I/O still needs to be subject to BIOS limitations.

BTW, my board was not Y2K compliant so I have an add-on ROM card to fix that issue.

It's an ISA Multi IO with a Winbond chip. It has one IDE port, two serial, one floppy, a game port and a parallel. Not sure of the model number but I don't think it has it's own bios. I've been using one of those cheapo CF card adapters with an 8GB flash card as a HDD. It works well in DOS, I probably don't need anything bigger really. I haven't tried linux on it yet. I usually use Arch linux, which won't work on a 486, but a bleeding edge OS doesn't seem a good fit for it anyway. It's been years since I made an LFS build though. Seems like a fun (lengthy) project.

You wouldn't have a dump of your bios would you. I'm not sure what all Y2K compliant entails but mine will accept 2019 as a year date in the BIOS. Here's the dump of my bios.

EDIT: I just realized I put X2 in the title when is should have been X4.

Oh man, I dumped that bios last night and everything seemed to go well. This morning I booted up the computer and it can't recognize any partitions, the bios sees the drive, but fdisk can't see the partitions. I tried to start fresh and fdisk makes the partitions but they don't save and are gone after reboot. I don't know what could have happened it's not like I wrote to the eprom. I just dumped it.

I have no clue what caused the issue but it's resolved now. I re-seated my IO board several times and changed the IDE cable and that didn't fix it. I put in a 16GB Sandisk CF card instead of the one of the two 8GB Transcend I had been using and it started working. But now the two Transcend CF cards work too.

Anyway, I was able to locate a couple more bios roms. They are actually for a Freetech F38xg but that board seems identical to my Genoa by appearance. Maybe one or the other is a rebrand. I've tried both of these on my board. One is just a hack of the other. The unhacked one makes my 8GB cards report properly in the bios splash and FDISK (the Genoa bios always reports 1592MB, though FDISK sees it as 7.6GB), and the hacked one makes my 32GB card report properly (more or less) in the bios splash but Win98 FDISK still only shows it as 8024MB.

I haven't tested them much but, I'll post them here, who knows maybe me or someone else will need them in the future.

here is my dump 😀

Thanks. I'm gonna try it out this weekend.

Hey overlord, any chance those freetech bioses you tried are newer than bakemonos? I'd like to run an AM5x86 133 in my board but no dice.SiS85C471 supports l1 write-back and the 5x86 but it won't work on our system (defaults to DX4 x3 w/ write-through). Trying to find out if a newer bios exists OR if a modbios can be done to add proper 5x86 support. I have the exact same 11/1994 bios he posted.

Also, since you're using the same exact board and bios as me, you should check out this thread I opened about an issue with L2 cache in write-back mode, you may have a massive drop in main memory performance: 486 cache/ram speed issue with write-back (very detailed post from user Megaherz on fixing) .. I don't have a programmer on hand so it'll be a bit before I can test it on my board, but it's a simple fix to boost performance.

The multiplier is controlled by a pin on the CPU, nothing to do with the BIOS.

To my knowledge the 5x86 internally multiplies x2 to x4. If the board would set 2x, it would work. Since the board thinks its a dx4 and forces x3, it won't work. That would make the bios have plenty to do with it, no? That pin should be the exact same pin that would set x2 on a dx2, how else would the cpu have worked on any boards launched before late 1995?

This board automatically sets the cpu, theres no difference in jumpers between a dx2, dx4, and Pentium overdrive. No manual controls. Bios must be in charge of deciding what cpu is in the socket.

Hey which EEPROM did the Genoa 486VLG have? Looks like the socket is DIP-28, I'm going to have to take the system down a good bit before I can peel the sticker back, would like to skip that if I can 😜

Would W27C512 do?