It's certainly possible, but it would be prohibitively slow to do anything except examine the behaviour in order to write a more traditional accurate simulator. I take it you're familiar with http://www.visual6502.org/ - that's 8 times less complicated than an 8088 (even excluding the support chips, as you said).

There's also the problem that nobody outside of Intel has the exact schematic of the 8088. I made contact with the Visual 6502 guys in 2013 to see what needs to be done to get some good die photos, but they just told me to be patient. I found some hi-res die photos of an 8086, but on close inspection they don't seem to show enough information to be able to reverse-engineer (I guess because they just show the top layer). In particular, only half of the bits in the microcode ROM and PLA seem to be visible.

So in the meantime I think that probing the 8088 from the outside (with my ISA bus sniffer) is likely to be the more fruitful route to get 8088MPH emulated correctly.

fpga?

FPGA would definitely solve the first of those two problems, and indeed somebody's done exactly that: http://www.microcorelabs.com/mcl86.html. It's not cycle exact, though, and doesn't run 8088MPH correctly.

reenigne wrote:

FPGA would definitely solve the first of those two problems, and indeed somebody's done exactly that: http://www.microcorelabs.com/mcl86.html. It's not cycle exact, though, and doesn't run 8088MPH correctly.

Wow, they even made a DOS Plus video! 😁

https://www.youtube.com/watch?v=EmloLviZIIg

(..and I thought I was about the only fool who takes an interest in this relic .)