NOTE: this speedtest is no longer relevant for Scratch 3 Furthermore, the broken timer block in Scratch 3 means the timings are pretty wrong anyway... :( ======================================= MORE SPEEDTESTS: http://scratch.mit.edu/studios/795672/ ------- This one shows that, despite the extra work Scratch has to do when interpreting use of a custom block, it *can* be worth using an 'inner' custom block. In this case it avoids creating a temporary variable (xorig), avoids performing some fairly expensive operations more than once (sine & cosine here), and the values used in its script can be referenced as parameters rather than referenced as variables (and we already know that referencing custom block parameters is slightly quicker than variables: http://scratch.mit.edu/projects/40989674/ ). Note: I've ignored the fact that "ang" is fixed for all of the repeat 100000 loop (and so we could calculate sin(ang) and cos(ang) before the repeat). I've assumed that in real life the angle would likely be different for all the different objects that might be rotating.
