It doesn't seem like there is much love from others, but I would love to get a copy of 68k SoftWindows 2 to have the 386 emulation, as that unlocks a LOT more guest software options. With a SCSI2SD or BlueSCSI for a drive, the performance is decent enough on a 68040. I really hope the dev can find it or even better the old code to recreate it.
It's hard to see how SoftWindows 2.0 in 68K code can have fast enough 386 emulation even on a 486. See:
I tried using SoftWindows on my Performa 475. It was unusable. Click a button, and listen to the hard drive churn away for the next five minutes.
A Performa 475
is a 68040 computer. The major reason why it can't be fast enough, is because the emulator on even an LC 630, would have to execute a 386 instruction in about 4 clock cycles to achieve a 16MHz 386 performance (since the 386 executes instructions in >=2 cycles). This would only be vaguely possible with a dynamic recompilation emulator of the kind made possible by Rosetta, but those kinds of emulators didn't exist then. In practice, the SoftWindows 2.0 emulator will take at least 10 instructions (at least 20 cycles) to execute one 286 or 386 instruction, but probably up to 100 instructions. So, a 286 or 386 will be emulated at about 3.3MHz down to 0.33MHz. This is why
@s_pupp reports an unusably slow execution speed.
On my PowerBook 1400/166 I get vaguely usable performance on SoftWindows 3. It's subjectively about as fast as a 16MHz to 32MHz 386 with a graphics accelerator, which is probably about correct, because that's the same ballpark as the built-in 680LC40 emulator on the PowerBook itself.
These things are largely unavoidable I'm afraid.
