[Libre-soc-bugs] [Bug 798] fix Ripple to use parallel prefix sum rather than a serial gate chain

[bugzilla-daemon at libre-soc.org]

https://bugs.libre-soc.org/show_bug.cgi?id=798

--- Comment #2 from Jacob Lifshay <programmerjake at gmail.com> ---
I implemented both the non-work-efficient and the work-efficient parallel
prefix sum algorithms.

I also added code to render the graph of operations done by the algorithms.

graphs in tests:
https://git.libre-soc.org/?p=nmutil.git;a=blob;f=src/nmutil/test/test_prefix_sum.py;h=63aa68ee94c998102f40d3c16ab161c40f08dcb3;hb=7ebe1cc2294de816b2d2809ce84492ae6aecb5d1

I ran out of time today so I'll get to adding more tests and using it in Ripple
another day.

Note that Ripple can't actually be used in TestEqualLeadingZeroCount since it's
not general enough. prefix_sum is though.

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