Subject: [linux-audio-dev] floats vs. ints
From: est_AT_hyperreal.org
Date: pe loka 15 1999 - 11:44:34 EDT
Paul Barton-Davis discourseth:
> In message <87zoxljw0v.fsf_AT_flophouse.localnet>you write:
>
> my main reason is the cost on the Pentium of converting back and
> forth. but i have also read compelling arguments on why
> float makes sense over on the Csound mailing list.
Can you summarize? I can certainly think of some arguments. :) I was
just checking out some code I wrote exactly a year ago. I must have
just got my soundcard working under linux as I was benchmarking
int/float/double operations. As soon as I saw float was quite viable,
I was happy to bid ints goodbye. That was a pleasant change in some
ancient (80s!) assumptions of mine.
> [ aside: are people here familiar with our benchmarking of short/float
> conversion on the Csound list earlier this year ?
You may remeber my posting to LAD a while ago reporting something like
a 10-fold speedup of float<->int16 conversion by doing it via 3dnow.
That was using float<->int32 operations + some regular x86 asm. The k7
has added float<->int16 operations which would speed it up even more.
Apparently the k6-2 has these as well, though they're `undocumented'.
Eric
This archive was generated by hypermail 2b28 : pe maalis 10 2000 - 07:27:59 EST