[FFmpeg-devel] [PATCH][WIP] lavc/cbrt_tablegen: speed up tablegen

[Ronald S. Bultje]

Hi,

On Fri, Jan 1, 2016 at 10:52 AM, Ganesh Ajjanagadde <gajjanagadde at gmail.com>
wrote:

> On Fri, Jan 1, 2016 at 7:51 AM, Ronald S. Bultje <rsbultje at gmail.com>
> wrote:
> > Hi,
> >
> > On Fri, Jan 1, 2016 at 10:46 AM, Ganesh Ajjanagadde <
> gajjanagadde at gmail.com>
> > wrote:
> >>
> >> On Fri, Jan 1, 2016 at 7:43 AM, Ronald S. Bultje <rsbultje at gmail.com>
> >> wrote:
> >> > Hi,
> >> >
> >> > On Fri, Jan 1, 2016 at 10:40 AM, Ganesh Ajjanagadde
> >> > <gajjanagadde at gmail.com>
> >> > wrote:
> >> >>
> >> >> On Fri, Jan 1, 2016 at 6:54 AM, Ronald S. Bultje <rsbultje at gmail.com
> >
> >> >> wrote:
> >> >> > Hi,
> >> >> >
> >> >> > On Thu, Dec 31, 2015 at 9:55 PM, Ganesh Ajjanagadde
> >> >> > <gajjanagadde at gmail.com>
> >> >> > wrote:
> >> >> >>
> >> >> >> This exploits an approach based on the sieve of Eratosthenes, a
> >> >> >> popular
> >> >> >> method for generating prime numbers.
> >> >> >>
> >> >> >> Tables are identical to previous ones.
> >> >> >>
> >> >> >> Tested with FATE. Does not work yet with --enable-hardcoded-tables
> >> >> >> due
> >> >> >> to the union and lack of proper WRITE_ARRAY for it. Want to get
> >> >> >> feedback
> >> >> >> on this; if we always dynamically init it this won't need
> >> >> >> addressing.
> >> >> >
> >> >> >
> >> >> > I think you're getting ahead of yourself here. Assume for now that
> >> >> > the
> >> >> > hardcoded-tables feature will continue to exist for a while.
> >> >>
> >> >> I was referring to just this one, not to the question in general.
> >> >
> >> >
> >> > I was also referring to this case specifically. Assume, for now, that
> >> > hardcoded tables for this specific case, will continue to exist.
> >>
> >> Then defend it technically please. For instance, you have not
> >> addressed the fundamental amortization of table init cost.
> >
> >
> > No. It is the status quo, and remains that until it changes. That's why
> it's
> > called the status quo. I've encouraged you several times to have this
> debate
> > in the open, but you keep jumping away when that comes up.
>
> What is this if not "open"? Where does this "jumping away" come from?


Open is an identifiable thread where people interested in that subject can
participate. "[PATCH] cbrt: improve" is barely anything like that. Call it
"[RFC] Disable compile-time tablegen if total cyclecount < $threshold".

Ronald