The program doesn't require a V20 at all .. it changes the timing of the DRAM refresh. Refreshing at slower rates makes the DRAM available more often for the CPU, thus improving the performance. But if you refresh the DRAM too slowly the bits start dropping - hence the warnings about parity errors in the readme.
Take a look at the ports it is manipulating: 0x43 and 0x41 - they are for the 8253 timer. Port 41 is used to set the timer that control the DRAM refresh interval.
@mile: Agreed. There were several magazine articles at the time about doing this--all of which mumbled something about perhaps compromising the integrity of your RAM data.
If there is something I have learned in almost 25 years of owning PCs its that the only way to make your compute faster is to buy a faster system. Overclocking, changing timings, etc. is just asking for pain later. The good thing about PCs now is that as long as you don't want to be on the bleedin edge it is pretty cheap to upgrade! If you want a faster XT then you need a 286 ....
Related thread at http://www.vintage-computer.com/vcforum/showthread.php?20593
There we worked out that, keeping within the timing specifications of the DRAM chips, the maximum gain in available CPU bandwidth for the 5160 would only be 0.22%
Well, if we are at it - what do you think of this one... (attached)
(these things might almost be worth its own thread- 'obscure speed enhancing utils and what they really do' )
Jorg, at least 386to486 is just a TSR hoax. It doesn't really do anything--most of the executable is taken up by Microsoft C library routines and messages.
I seem to remember that there was a similar hoax program that claimed to turn your 486 into a Pentium...