[Cryptography] Finding undocumented opcodes

[Howard Chu]

Dave Horsfall wrote:
> On Sat, 5 Aug 2017, Bob Wilson wrote:
> 
>> With all the apparently random opcodes I saw generated, on different 
>> processors, by students in my CS classes and by the programmers I managed in 
>> industry, it is hard to believe there can be many (on those processors) that 
>> have not already beet tested!
> 
> The classic undocumented instruction set would have to be that of the Z-80, as 
> in e.g. http://www.z80.info/z80undoc.htm.  I can't believe that people went to 
> so much trouble to explore every bit pattern!

Reminded me of a 1975 article in Byte magazine on undocumented 6502 opcodes.

> But speaking of weird opcodes, there is the "population count" instruction on 
> the CDC series; it counted the number of "1" bits in a (60-bit) word, and was 
> known as the "NSA Instruction" (at first glance, it appeared to have no use at 
> all, but Seymour Cray wouldn't have included it for no reason).

So did the NSA knock on AMD's door?

https://en.wikipedia.org/wiki/SSE4#POPCNT_and_LZCNT

-- 
   -- Howard Chu
   CTO, Symas Corp.           http://www.symas.com
   Director, Highland Sun     http://highlandsun.com/hyc/
   Chief Architect, OpenLDAP  http://www.openldap.org/project/