Performance improvement of get_detcost() function. #110
Description
What is your feature request?
It appears that get_detcost() function relies heavily on fdiv/vdiv instructions. Swapping with fmul instructions should show improvement in CPU cycles.
What solution or approach do you envision?
Here is the associated change -- #108. The description shows the performance improvement seen in a benchmarks. LMK what you think.
How urgent is this for you?
P3 – not blocked; it's an idea