Re: [Audacity-devel] doubles and floats

[Dominic M. <do...@au...>]

On Feb 23, 2006, at 11:47 PM, Steve Harris wrote:
> On Thu, Feb 23, 2006 at 08:09:06 -0500, Mar...@ao... wrote:
>>    Hi
>>
>>    Pardon my ignorance (yet again) but should I be using doubles  
>> or floats
>>    for variables that are, effectively, part of the UI and won't  
>> have much
>>    effect on a signal?  For examples, the frequency scale of an  
>> equalizer,
>>    the log of the frequency scale etc..  Does it make any performance
>>    differences to anybody?  My (uninformed) guess is that it would  
>> be OK to
>>    ditch float and use double for everything (on a modern  
>> machine), but I
>>    don't know what I'm talking about in this realm.
>
> Personally I doubt very much that it would make a difference, the main
> reason we use floats in DSP routines it to reduce the cache damage,  
> when
> streaming lots of data, but the data structures in GUI code are so  
> huge
> that the odd extra 32bits here and there probably doesn't matter.

Agreed.  I've done some performance comparisons and for large arrays,  
floats are about twice as fast as doubles, on a wide range of  
processors.  Basically, memory bandwidth is often more of a  
bottleneck.  For computation-intensive code, that difference  
disappears, so you should almost always use doubles.

For your application: use doubles.

- Dominic



> - Steve
>
>
> -------------------------------------------------------
> This SF.Net email is sponsored by xPML, a groundbreaking scripting  
> language
> that extends applications into web and mobile media. Attend the  
> live webcast
> and join the prime developer group breaking into this new coding  
> territory!
> http://sel.as-us.falkag.net/sel? 
> cmd=lnk&kid=110944&bid=241720&dat=121642
> _______________________________________________
> Audacity-devel mailing list
> Aud...@li...
> https://lists.sourceforge.net/lists/listinfo/audacity-devel