The V20 instruction set is interesting and has some very 1980s extensions. In addition to the 8080 emulation mode, it has instructions for binary-coded decimal arithmetic and bit fields.
However, NEC uses their own naming scheme which makes picking out the differences from the 8088 unnecessarily difficult. The bits are the same but the mnemonics are different. For example, they renamed all the registers (AX -> AW, SI -> IX, etc) and many opcodes (JMP -> BR, REP MOVSB -> REP MOVBK).
Mentioned in the article, same link too. The court documents included a diagram of the microcode format, and also a comparison of the reset sequence between Intel and NEC.
This is an amazing project, I'm always impressed by stuff like this. I'd be very curious to know, though, if there's a market for achievements like that.
What I remember of the V20 was that there was some benchmark (Norton?) that was way faster, but I never saw a difference for any real application that I tested.
I have a soft spot for the V20. The laptop I carried in high school was V20-based (a NEC Ultralite PC1701). I remember a CGA 3D billiards game I played on both the laptop and an XT clone desktop. The V20 was definitely faster for that game.
For most of what I actually used it for (typing notes and playing with QBASIC) I didn't notice a difference.
If anybody's interested in the V20 from a programming perspective, The User's Manual from NEC is here: https://archive.org/details/bitsavers_necV20V30U_11351331/mo....
The V20 instruction set is interesting and has some very 1980s extensions. In addition to the 8080 emulation mode, it has instructions for binary-coded decimal arithmetic and bit fields.
However, NEC uses their own naming scheme which makes picking out the differences from the 8088 unnecessarily difficult. The bits are the same but the mnemonics are different. For example, they renamed all the registers (AX -> AW, SI -> IX, etc) and many opcodes (JMP -> BR, REP MOVSB -> REP MOVBK).
Of course the V20’s microcode was the centre of an interesting legal dispute when NEC was sued by Intel.
https://thechipletter.substack.com/p/intel-vs-nec-the-case-o...
Mentioned in the article, same link too. The court documents included a diagram of the microcode format, and also a comparison of the reset sequence between Intel and NEC.
This is an amazing project, I'm always impressed by stuff like this. I'd be very curious to know, though, if there's a market for achievements like that.
My type of article! Thank you
Lovely work, thanks for sharing!
ha I remember that chip so well
instant speed boost for IBM PC
just pop out the 8088 and put in a $20 V20
only problems is some programs ran their clocks too fast
wow that article is amazing, I cannot believe people can reverse engineer at that level, brilliant
What I remember of the V20 was that there was some benchmark (Norton?) that was way faster, but I never saw a difference for any real application that I tested.
I have a soft spot for the V20. The laptop I carried in high school was V20-based (a NEC Ultralite PC1701). I remember a CGA 3D billiards game I played on both the laptop and an XT clone desktop. The V20 was definitely faster for that game.
For most of what I actually used it for (typing notes and playing with QBASIC) I didn't notice a difference.