[FFmpeg-devel] [PATCH] avcodec/mpegaudio_tablegen: more dynamic initialization speedups

[Ganesh Ajjanagadde]

On Sat, Nov 28, 2015 at 8:48 AM, Ronald S. Bultje <rsbultje at gmail.com> wrote:
> Hi,
>
> On Sat, Nov 28, 2015 at 12:46 AM, Ganesh Ajjanagadde
> <gajjanagadde at gmail.com> wrote:
>>
>> I personally consider this net ~ 10x and overall perf numbers sufficient
>> for using dynamic initialization all the time here, especially since the
>> tables are large.
>
>
> This is a logical fallacy. By this logic, I could post some really shitty
> code, speed it up 10x and then say "I made it better so it should be
> default", regardless of what the actual quality of the current code is.
>
> If you want to argue for dynamic vs. static table default choice, you should
> compare only these two currents (dynamic vs static), not the improvement of
> an old dynamic vs current dynamic table generation.

What is the "logical fallacy here"? My point was that previously the
code was much slower, now it is much faster, and is fast enough for my
own personal tastes to be made the default.

Static vs dynamic is mostly a speed vs size tradeoff. I have increased
the speed, and have essentially made it a Pareto optimal point on the
speed vs size tradeoff as far as I can tell - at this level of space
usage, there is not really anything that can be done. If you don't
feel that way, please tell me and I will improve it further.

Whether that is sufficient to make it the default is of course subjective.

[...]