if you want to raise multiple issues, please post multiple mails (with different subjects). it makes it much easier to find and follow interesting bits.
On 2017-08-16 00:25, Yoann Le BORGNE wrote:
# You can add compiler flags using CODECFLAGS, MORECFLAGS and/or MORELDFLAGS. # For example, you can turn off optimizing and enable debugging: CODECFLAGS=-g # or *compile with native double precision: MORECFLAGS=-DPD_FLOATSIZE=64*
Is this flag enabling double precision as explained here -> http://www.katjaas.nl/doubleprecision/doubleprecision.html ? If yes, is it supported/experimental/deprecated?
yes, and i'd say it's "experimental" :-)
If no, what is the current state
on this topic?
In a more general way, what are the best compilation options to get the most out of Pd on one side, and my computer on the other side?
this very much depends on your definition of "most" for both Pd and your computer. iirc, that the default arguments are pretty OK performance wise. they don't enable any fancy CPU instruction sets, which usually gives a very good compromise between performance and CPU compatibility.
if this doesn't fit your needs, you probably have to find out for yourself (there was some genetic algorithm program that would find the "optimal" compiler optimization flags, but this requires a representative test corpus (aka patch))
m,aft IOhannes