Commit 2022-11-20 07:14 1b6572f0
View on Github →perf: optimizations in norm_num
and ring
(#613)
With a little help from perf and hotspot, I managed to find and improve some expensive bits in norm_num
and ring
. As a result, the time spent in linear_combination
in the linear_combination test suite (which is spent almost 100% in ring1
, of which about 35% is in norm_num
) is decreased by just over 50%.