Subject: Re: [linux-audio-dev] New "plugin"
From: Paul Barton-Davis (pbd_AT_Op.Net)
Date: Wed Nov 29 2000 - 15:43:02 EET
>All I've been doing is running applyplugin under time and then
>dividing the user time by the length of the input sample and multiplying
>by the clock speed of the machine.
>
>Awful metric but it gives you some idea of how much CPU power is used by
>the plugin when its running realtime. Of course the reults vary wildly
>with the chunk size and moon phase.
OK, can i suggest a better way, one that would give us really hard
numbers ? Modify applyplugin to do this:
#include <asm/msr.h>
.
.
.
unsigned long then, now;
rdtscl (then);
plugin->run (nframes);
rdtscl (now);
plugin_cycles = now - then;
/* you know have an close-to-exact cycle count for the
execution of the plugin. collect the results in an array
with a spread of about 1000 cycles, then at program end,
average the array and print the result.
*/
This would be a much more useful and generic number (though its
obviously intel/x86 specific).
>Cheers, I'm not sure I'd compare it to TDM's though! It's a logical
>extrapolation of a really cool Swiss sound mangling "synth" that a
>friend has. Thier quite hard to get hold of and I was just thinging, hang
>about all it is is a bunch of oscilators and modulators, its all piss
>easy really. Theres lots of tweaks in there to make it sound more
>like analogue parts, but it was basicly quite simple.
heh. i was over at the studio last night, and the band/friends that
were in (excellent brooklyn gogo/funk stuff) had an electro-harmonix
bass synth a lot like this. they were all ooh-ing and ah-ing it. the
brochure for e-h notes "16,20,24 bit. all real musicians hear the
irregularities [sic] of digital processing, which is why most of them
want the warmth of electro-harmonix analog gear" or something very
close to that. bah! humbug! :)
--p
This archive was generated by hypermail 2b28 : Wed Nov 29 2000 - 17:35:34 EET