First post, by MastErAldo
- Rank
- Newbie
This post is intended to provide an answer to my older post regarding hangs on my 486 setup with mainly EGA games.
You can find the original topic here: Cirrus Logic GD5424 (maybe)timing problems on some games
But also to ask for opinions about what I discovered.
Just for recap, I have a 486 system and the problem was that every single EGA game randomly hangs within seconds after started (games like doom, descent, monkey island 2 and in general VGA games run fine). The system has the following specs:
Motherboard: Micronics Gemini VLB, like this one: https://i.ebayimg.com/images/g/YBkAAOSwdJtevmle/s-l400.jpg part number 09-00144
CPU: am486dx4 NV8T
Video: Cirrus Logic GD5428 VLB equipped with 1MB of RAM
RAM: 16mb FPM ECC ram 60 ns
I recently managed to get a "new" system and what is important about it is that is equipped with an Intel 486 DX2 66mhz (sSpec sx750).
I placed the intel DX2 on my original system and every single game worked flawlessly.
Just to be sure that it wasn't my old CPU fault I also tried to use an SV8B dx4 (the amd writeback cache model) and the problem was still there.
Since the old motherboard is a 5v only, the amd CPUs used a voltage adaptor which also lets you set te multiplier to x3 or x2, I though that maybe dx4 was just too fast.
So I switched to x2, every benchmark confirmed that the actual speed was 66mhz, but the games still hang.
As a counter test I used my amd chips on the new board (also equipped with the same VLB card i used in the original one) and every EGA game run flawlessly.
So I'm pretty sure now that my original system has some incompatibilities with AMD chips regardless of their speed/cache/multiplier.
I'm happy to have found a solution to a problem that, searching on the internet, may be not that common, but now, just out of curiosity I wonder if anyone has an idea about what can cause this problem.
I mean, I know that even though they are compabile chips they are made by 2 different companies and the later am486 models had to come with their own microcode, but are you aware of some subtle differences between these 2 chip models that can in some circustances, cause timing issues?
I though about a particular instruction that on amd chips completes a cycle before/after intel ones, but so far I have not found something to support my theory.
My Projects on GitLab
Micronics Gemini VL-Bus | AMD 80486dx4 NV8T | Tseng Labs ET4000/W32i | Sound Blaster 16 Value CT2770
16mb FPM ram 60ns | IDE 504 MB | 12x CD-ROM Drive | MSDOS 6.22 & Win 3.11