[Libre-soc-bugs] [Bug 238] POWER Compressed Formal Standard writeup

bugzilla-daemon at libre-soc.org bugzilla-daemon at libre-soc.org
Sun Nov 22 23:19:27 GMT 2020


https://bugs.libre-soc.org/show_bug.cgi?id=238

--- Comment #62 from Luke Kenneth Casson Leighton <lkcl at lkcl.net> ---
alexandre i have such a massive headache i'm not really able to concentrate. 
brief comments below, can you start a 2nd section in the wiki page, "compressed
2nd version"

(In reply to Alexandre Oliva from comment #60)

> To me, the 10-bit encodings come across as complexity I'd much rather
> get rid of.

the entire SimpleV concept is based around context tagging.  it's not going
away.

trying to drop this one tool in an arsenal of tools all based around the same
state-based concept would be... fruitless.

i.e. to try to get 10bit to not exist because it requires state, when the
entirety of SimpleV requires state, so that objdump will have an easier time,
this is not going to happen because for objdump to support disassembly of
SimpleV *will* require the state tracking i mentioned.

if however, disregarding state as being a "disadvantage", there are *better
schemes* than the 10bit one, i'm happy to hear about them.


> I don't really feel that an extend-next opcode that supplies an extra
> operand and/or an extended immediate range for the subsequent insn that

SV Prefix (which you've not been introduced to yet) is already adding 2 bits
per register.

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


More information about the libre-soc-bugs mailing list