[Libre-soc-dev] efficient decoding algorithm for variable-length instructions

[Cole Poirier]

On Wed, Nov 25, 2020 at 6:34 PM Luke Kenneth Casson Leighton <lkcl at lkcl.net>
wrote:

> y'know what?  suddenly occurred to me: i don't think we're being
> radical enough.  how about: we design something more akin to RISC-V
> RVC 16/32/48/64 that uses 6 bits maximum to encode, and just do a
> total re-map of all OpenPOWER 3.1B Major opcodes?
>
> all bits *other* than the Major Opcodes would be left entirely
> unmodified.  EXT004 would map to 0b0001, EXT031 to 0b0010, EXT19 to
> 0b0011 etc. etc. etc.
>

Doesn't this run up against the OPF benefit vs complexity tradeoff
unfavourably like you were explaining yesterday i.e. remapping all of OPISA
v3.0B/3.1B?

Cole