̶F̶o̶r̶ ̶t̶h̶e̶ ̶s̶p̶a̶r̶s̶e̶-̶m̶a̶t̶r̶i̶x̶-̶m̶u̶l̶t̶i̶p̶l̶i̶c̶a̶t̶i̶o̶n̶ ̶e̶x̶a̶m̶p̶l̶e̶ ̶g̶i̶v̶e̶n̶,̶ ̶w̶o̶u̶l̶d̶n̶'̶t̶ ̶i̶t̶ ̶b̶e̶ ̶b̶e̶t̶t̶e̶r̶ ̶t̶o̶ ̶p̶r̶e̶-̶s̶o̶r̶t̶ ̶e̶a̶c̶h̶ ̶v̶e̶c̶t̶o̶r̶ ̶a̶n̶d̶ ̶d̶o̶ ̶n̶^̶2̶ ̶w̶a̶l̶k̶s̶ ̶o̶n̶ ̶p̶a̶i̶r̶s̶ ̶o̶f̶ ̶p̶r̶e̶s̶o̶r̶t̶e̶d̶ ̶v̶e̶c̶t̶o̶r̶s̶,̶ ̶r̶a̶t̶h̶e̶r̶ ̶t̶h̶a̶n̶ ̶(̶a̶s̶ ̶I̶ ̶t̶h̶i̶n̶k̶ ̶w̶a̶s̶ ̶i̶m̶p̶l̶i̶e̶d̶)̶ ̶d̶o̶i̶n̶g̶ ̶n̶^̶2̶ ̶h̶a̶s̶h̶e̶d̶ ̶s̶o̶r̶t̶i̶n̶g̶ ̶o̶p̶e̶r̶a̶t̶i̶o̶n̶s̶?̶ ̶(̶F̶o̶r̶ ̶t̶h̶a̶t̶ ̶m̶a̶t̶t̶e̶r̶,̶ ̶I̶ ̶w̶o̶n̶d̶e̶r̶ ̶w̶h̶y̶ ̶s̶p̶a̶r̶s̶e̶ ̶m̶a̶t̶r̶i̶c̶e̶s̶ ̶w̶o̶u̶l̶d̶n̶'̶t̶ ̶a̶l̶r̶e̶a̶d̶y̶ ̶b̶e̶ ̶r̶e̶p̶r̶e̶s̶e̶n̶t̶e̶d̶ ̶i̶n̶ ̶s̶o̶r̶t̶e̶d̶-̶a̶d̶j̶a̶c̶e̶n̶c̶y̶-̶l̶i̶s̶t̶ ̶f̶o̶r̶m̶ ̶i̶n̶ ̶t̶h̶e̶ ̶f̶i̶r̶s̶t̶ ̶p̶l̶a̶c̶e̶)̶ ̶

EDIT: ah no I'm being dense, you'd aggregate the union of all the set-columns indices across rows and the union of the set-row indices across the columns, keeping track of the source locations, and do the hashed sorting on those union vectors to find all the collision points. You could still get a small win I think by sorting the row-aggregation and column-aggregation separately though?