Re: Cheers for GoGo.

[Al Hudson <nospam [at] york.ac.uk>]

On Mon, 13 Mar 2000, Chris J/#6 wrote:

> Well, gogo is at least 40 to 50% faster than lame on my machine (which it is 
> based on), probably down to a mix of the assembler and the MMX instructions. 

Probably not ;) LAME is based on shoddy 'Hypothetical Reference Decoder'
ISO code, IIRC. Quite why you need faster than real-time encoding is also
something I don't understand ;)) LAME sped up the original HRD by around
an order of magnitude, yet was written in the same language ;). C/Asm has
very little to do with these things, you certainly won't get an order of
magnitude improvement simply by coding in a different language.

> I don't know if you remember a program called Fastcad for DOS...it was the 
> fasted CAD I'd ever seen running on a 286, mostly because it was 100% 
> assembler. You could *really* tell the difference.

How do you know it was faster because it was written in assembler?

> I used to have and code on (so it's not a sea-change for me). Curiously 
> enough as well, you still do a lot of assembler in engineering fields (which 
> I have done), with Z80's and 68000's being used as embedded controllers.

Even embedded controllers don't use asm any more ;)) I'd like to see you
write software for an embedded system that dealt with real-time issues and
concurrency ;))

Cheers,

Alex.


---------------------------------------------------------------------
Sheffield Linux User's Group - http://www.sheflug.co.uk
To unsubscribe from this list send mail to
- <sheflug-request [at] vuw.ac.nz> - with the word 
 "unsubscribe" in the body of the message. 

  GNU the choice of a complete generation.