On Thu, 29 May 2025 20:59:31 +0200
ichthyo <[email protected]> wrote:
>> On 28.05.25 10:38, Will Godfrey wrote:
>>> Finally, I notice this is slightly slower to startup.
>
>On 29.05.25 17:42, ichthyo wrote:
>> Haven't noticed that, but I'll do a comparison with two release builds.
>
>Drilled down with timings.
>
>Culprit is Bank loading.
>Seems the new code somehow does not use the data
>from the persisted file and thus does a full rediscovery.
>
>
>The infosearch function takes roughly the same time in both versions,
>about 110µs. But it is called 1000 times in the new version, and not
>at all in the old version.
>So the new version needs about 150ms longer for boot-up.
Ouch!
The search is only supposed to be done if an instrument file doesn't have all
of ADD, SUB & PAD entries defined in the header.
--
Will J Godfrey {apparently now an 'elderly'}
https://willgodfrey.bandcamp.com/
http://yoshimi.github.io
Say you have a poem and I have a tune.
Exchange them and we can both have a poem, a tune, and a song.
_______________________________________________
Yoshimi-devel mailing list
[email protected]
https://lists.sourceforge.net/lists/listinfo/yoshimi-devel
