[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]
Re: Cheers for GoGo.
On Mon, 13 Mar 2000, Alastair Donlon wrote:
> > Finally, someone agrees with me! If I want portability then C is fine,
> > but I would rather have two streamlined packages for two separate
> > chips, because when you have the code for one chip converting is
> > relatively easy, due to you working out all the problems on the first
> > chip.
> Excuse the sarcastic tone, but what the hell are you writing that requires
> such a degree of streamlining? OS internals?
Even OS internals are written in C mostly... ;)
> Bear in mind that streamlining at the asm stage is the final stage of the
> process. It requires the most amount of physical work and gains the least
> amount of speedup.
I think we've had this argument already ;) I think perhaps those
interested in speeding up programs should first take a look at TAOCP
(Knuth), and then the NASM Guide, and decide which is going to get you a
faster program more easily and more quickly.
Never seen GoGo myself, I presume it's an MP3 encoder? Depending on which
algorithm it implements, you can't necessarily just compare it to other
MP3 encoders and say 'yep, it's faster'. MPEG does not define any encoding
schemes at all; and encoder may encode to MP3 any way it pleases, using
any algorithm. A lot of MP3 encoders use the same algorithm, but many
don't, and the fast ones generally are using different algorithms - the
MP3 coming out of these will be different to the MP3 coming out of the
other types of programs, so comparing them is like comparing apples &
oranges.
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.