On Wednesday, 3 June 2020 at 16:39:38 UTC, 9il wrote:
On Wednesday, 3 June 2020 at 16:15:41 UTC, jmh530 wrote:
This is unclear:
The chart below shows matrix implementation times minus
ndslice times; negative means that ndslice is slower,
indicating that the implementation used here does not
negatively represent D’s performance.
This means that ndslice (in the way how it was used here) is
slower than the custom matrix type used.
I understood "negative means that ndslice is slower, indicating
that the implementation used here does not negatively represent
D’s performance."
The problem is that "The chart below shows matrix implementation
times minus ndslice times;" is not clear. The use of "times" in
particular makes the whole thing sound like a mathematical
equation. I think it would have been clearer if he had said, "The
chart below shows the elapsed time of running the matrix
implementation times subtracted by the elapsed time of an ndslice
implementation."
Regardless, a relative comparison, such as (matrix time / ndslice
time) - 1, would be better for controlling for the size of the
input.
