1 comment

[ 11.4 ms ] story [ 509 ms ] thread
Is there any generic way to preserve loss of numerical accuracy for any given function?