OpenPOWER Foundation ISA Working Group

TODO

Reminder of fundamental principle when putting forward proposals: the greater the change, the bigger the benefit needs to be. Complex changes with little benefit are likely to be rejected; simple changes with high benefit stand a better chance.

Also note that:

  • OpenPOWER is over 25 years old. This means a huge pedigree with a lot at stake (IBM in particular makes billions from it and will need to be reassured that its future is not compromised by proposals)
  • It also means that there are people within the community who really know what they are doing and know the ISA not just inside out but how it got to where it us now.

A good strategy to apply is one where LibreSOC proposals are useful well beyond our own use-case (broad as it is). For example, scalar transcendentals are useful outside of our domain; we happen to then vectorise them.