[Libre-soc-isa] [Bug 532] discuss iterative approach for statistical analysis of effectiveness of Compressed PowerISA encoding

bugzilla-daemon at libre-soc.org bugzilla-daemon at libre-soc.org
Sat Nov 28 23:41:36 GMT 2020


--- Comment #17 from Alexandre Oliva <oliva at gnu.org> ---
I had not noticed before writing the above, but the 16-bit immediate compressed
encoding is very very similar, in function and possibilities, to the
extend-next-insn mode I'd suggested before.

I'm not sure I see why one is too CISC-ish while the other isn't (one vs 5/6
bits to detect is probably a good hint ;-) but it seems to me it could very
well be used to significantly extend the range of registers accessible in
compressed mode, whether or not we go for 2-reg encoding in 16-bit insns, which
is just what I'd hoped to accomplish with an extend-next opcode.

This bit can even be used to make room for some extended opcodes if there are
any compressed primary opcodes that have no use for additional operands, e.g.
nop or unary operations, and even to reserve room for future extensions.

You are receiving this mail because:
You are on the CC list for the bug.

More information about the Libre-SOC-ISA mailing list