[Libre-soc-bugs] [Bug 1044] SVP64 implementation of pow(x,y,z)

[bugzilla-daemon at libre-soc.org]

https://bugs.libre-soc.org/show_bug.cgi?id=1044

Jacob Lifshay <programmerjake at gmail.com> changed:

           What    |Removed                     |Added
----------------------------------------------------------------------------
             Status|IN_PROGRESS                 |RESOLVED
         Resolution|---                         |FIXED

--- Comment #80 from Jacob Lifshay <programmerjake at gmail.com> ---
(In reply to Luke Kenneth Casson Leighton from comment #79)
> rats forgot about bug #1183, the locations where the assembler is
> sub-optimal (5 instructions instead of 1)

ok, icr all the locations, but I think the motivating examples are the spots
where Knuth's Algorithm D needs inputs with the most-significant word being
non-zero, so we have to find the number of leading zero words in our
fixed-length inputs so we know how many words to tell Algorithm D is in the
dynamically-sized inputs (basically, we just make Algorithm D ignore all the
high-order zero words and act as if the inputs are really shorter than they
are).

so, count-leading-zeros, except we're counting zero words, not bits.
https://git.libre-soc.org/?p=openpower-isa.git;a=blob;f=src/openpower/test/bigint/powmod.py;h=7fc794685bebb1f3c2451c64da041a0e81143e29;hb=d3c7a0ddc169877e8daf38a347091c366e85624b#l717
and
https://git.libre-soc.org/?p=openpower-isa.git;a=blob;f=src/openpower/test/bigint/powmod.py;h=7fc794685bebb1f3c2451c64da041a0e81143e29;hb=d3c7a0ddc169877e8daf38a347091c366e85624b#l734

I'll link bug #1183 to this comment, so I think that's all we need.

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