PYTHON · NUMPY
Silicon anyone can make

Count the gates in two datapaths for the same 8-bit operands: a classic int8 array multiplier (~n^2 partial-product cells) versus a ternary add/sub/skip lane (~n). Fill the one gap to compute how much silicon the multiply costs. Pure numpy, deterministic, no numbers invented — you count them.

You read

the arrays and values already in scope

You change

the code you write in each cell

Fixed

the dataset and the checks that grade you