Dear Hugo, On Thu, 10 Jul 2003 hugo_lec@club-internet.fr wrote: [...]
There's an algorithm to use the less temporary variables as possible (in order to help the compiler to optimize).
[...] Could you please elaborate a little bit on that statement? As it stands I find it slightly irritiating. Is this assessment based on real experience? Given a compiler that transforms its parsed tree into SSA form at some stage (as GCC 3.5 will), besides those phi-expressions tons of temporaries are generated anyway in the IR. More or less by definition, later optimizations must cope with and reduce them. So: is there a point in trying to keep the number of temporaries low or is that just fiction? Maybe an example of the effect you are alluding at would be helpful. Cheers -richy. -- Richard B. Kreckel <Richard.Kreckel@GiNaC.DE> <http://www.ginac.de/~kreckel/>