Comment #2 on issue 1663 by [email protected]: Float64Array Performance http://code.google.com/p/v8/issues/detail?id=1663
Also, floating point arrays will be faster when storing newly calculated floating point numbers to them, because a new number object does not need to be created on the heap. Reading floating point numbers will not show much of a performance difference. Something like array[i+1] = array[i] + 0.01; should show more of a difference, assuming the optimizing compiler has a chance to optimize it.
-- v8-dev mailing list [email protected] http://groups.google.com/group/v8-dev
