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

[bugzilla-daemon at libre-soc.org]

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

--- Comment #141 from Alexandre Oliva <oliva at gnu.org> ---
jacob,

sorry, I misspelled your name before.  I've worked closely with a jakub for
some 15 years.

I think we're talking about different Ns, but I think I also made a mistake.

thinking of insns issued per cycle, I was reasoning in terms of possibilities
of starting points for insn n: for each insn, the number of possibilities
(nearly) doubles, I reasoned.  since the max insn length is a given, and a
constant, say M, each increment in n adds at most M half-words to look at in
phase 1, so it is O(n) half-words to precompute indeed

it's still a lot of throw-away computation, but it's indeed not even close to
as bad as I misconcluded.  thanks for the correction.

I don't see that it invalidates the reasoning, though.

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