[Ffmpeg-devel] 2.95?

Rich Felker dalias
Fri Jul 14 07:19:21 CEST 2006

On Thu, Jul 13, 2006 at 09:29:38PM +0100, M?ns Rullg?rd wrote:
> Mike Melanson <mike at multimedia.cx> writes:
> > Rich Felker wrote:
> >> Yes, gcc devs want to keep us from discovering and using the
> >> infinitely superior gcc 2.95 !! :)
> >
> > Loathe as I am to admit it, you might be onto something here, at least
> > for compilation speed. gcc 2.95 compiles the tree in 15 minutes. 3.4.6
> > takes 21 minutes. ~30 minutes for 4.1.1.

I don't think they're actually worried that gcc 2.95 is superior, but
rather eagar to bury it so that people forget that compiling is twice
as slow as it used to be with minimal practical gains. If people
realize how disgustingly slow and bloated gcc4 is, the developers have
to justify it or fix it.

> In theory, the compiler could be spending that extra time optimizing
> the code.  Were that the case, I'd happily trade a little compilation
> time for faster execution.  Sadly, I'm afraid this is not the case
> here.

Yes, I'm quite sure it's not the case. Rather, they're using big fancy
theoretically-better algorithms that have little actual benefit on
real code, and proud that the "old nasty hacks" that worked just as
well and ran twice as fast are gone.

In fact I actually once heard a rumor that the developers
intentionally crippled gcc 3 because of patent fears over some of the
algorithms used in 2.95. Can anyone confirm that they intentionally
made it worse because of patents?


More information about the ffmpeg-devel mailing list