Re: Optimizing my 128 channel IOProc.
Re: Optimizing my 128 channel IOProc.
- Subject: Re: Optimizing my 128 channel IOProc.
- From: Doug Wyatt <email@hidden>
- Date: Wed, 19 Oct 2005 17:25:35 -0700
On Oct 19, 2005, at 17:13, Mark Gilbert wrote:
Thanks to those folks who made suggestions on how to speed things
up in my heavy laden 128 audio channel IOProc.
I delved into shark, and within 20 minutes I had shaved huge
amounts off the tight loops.
Now, I have done what I think is practical in my own code, and I am
left with something unusual.
In Shark, alongside my main IOProc for the input and my render proc
for the output, I see 'AUMultibandCompressorEntry' taking a
whopping 10%.
I have a feeling this must be related to the AUGraph Matrix Mixer I
have on the back end of my application, but I am not using any
compressors to my knowledge.
Can anyone shed any light ?
Today I was heard to exclaim in the hall: "Never, ever believe any
symbol you see in CoreAudio.component!"
Only the entry points have symbols, the order of the functions within
the binary has been optimized, and the various tools that translate
addresses to symbols are simply picking the nearest one. They're more
often wrong than right.
Your feeling about the matrix mixer is a good guess; it's probably
copying memory.
--
Doug Wyatt
Core Audio, Apple
_______________________________________________
Do not post admin requests to the list. They will be ignored.
Coreaudio-api mailing list (email@hidden)
Help/Unsubscribe/Update your Subscription:
This email sent to email@hidden