[Bf-committers] float-upgrade to sequencer

[Robert Wenzlaff]

At 03:19 PM 2/28/06, you wrote:
>I'm working my way through the sequencer code and am just wondering what's
>the best way to decide, if we should do effect processing on rect_float or
>just rect.
>
>I'm fairly sure right now, there is no real automatic way to decide
>if the user prefers speed or quality. (If input and output is float then
>everything is easy, but if both are not? Maybe we want to do float
>processing in the middle and just convert at both ends?)

Well, under the current Intel architecture (and probably common under 
others as well) int-int math is faster that float-float math, but 
float-float is oddly faster
than float-int math.  The fastest way to do float-int math is to convert 
the ints to floats in hardware and use the FPU to do a float-float. 
(Int-int has it's own hardware in the ALU.)  But the conversion isn't free, 
so it's slower.    So if you have to do float anywhere along the line, you 
should do float everywhere along the line.

So at least in some cases you get speed AND quality.

Detective Thorn.