Time |
Nick |
Message |
00:05 |
|
Eragon joined #minetest-dev |
05:00 |
|
MTDiscord joined #minetest-dev |
05:49 |
|
SFENCE joined #minetest-dev |
06:43 |
[MatrxMT] |
<Zughy> Can we remove built-in knockback for punch() in 6.0? |
09:08 |
|
SFENCE joined #minetest-dev |
11:41 |
|
ivanbu joined #minetest-dev |
12:50 |
|
jonadab joined #minetest-dev |
13:31 |
sfan5 |
sure |
13:59 |
|
SFENCE joined #minetest-dev |
14:06 |
|
BuckarooBanzai joined #minetest-dev |
14:45 |
|
hwpplayer1 joined #minetest-dev |
15:26 |
|
AwesomeAdam54321 joined #minetest-dev |
15:38 |
AwesomeAdam54321 |
Hi, would patches to use Irrlicht's particle system still be accepted if the 'vertical' particle effect problem was solved? |
15:40 |
MTDiscord |
<luatic> There is no more Irrlicht particle system in Luanti. What benefits do you think the Irrlicht particle system would bring? |
15:42 |
AwesomeAdam54321 |
Better performance |
15:42 |
MTDiscord |
<luatic> Why? |
15:43 |
AwesomeAdam54321 |
Is it not a problem anymore? |
15:43 |
AwesomeAdam54321 |
If it's already optimised enough then I guess there's nothing to be done |
15:44 |
MTDiscord |
<luatic> "enough" is always debatable, but the glaring lack of batching has been solved; particle rendering is batched |
15:59 |
|
Desour joined #minetest-dev |
18:15 |
MTDiscord |
<exe_virus> The spatial map candidate results are in (got enough data over 7 days), updated the issue so core devs can take a look: https://GitHub.com/minetest/minetest/issues/14613 |
18:46 |
|
wrrrzr joined #minetest-dev |
18:46 |
wrrrzr |
Can i send PR with remove lambda and create constexpr functions |
18:46 |
wrrrzr |
*function |
18:47 |
MTDiscord |
<luatic> What would the advantage of a constexpr function be? C++ compilers are pretty good at getting rid of lambdas. |
18:47 |
MTDiscord |
<luatic> And even if there is a runtime cost: Why do you think the runtime cost is significant? |
18:48 |
celeron55 |
i think lambdas make things very readable and optimizers are good at optimizing them, so in general that doesn't sound like a good idea |
18:48 |
wrrrzr |
But lambda created in big loop |
18:48 |
celeron55 |
(optimizers? i mean compilers. well, given most of the compiler does optimizations it would be the better name...) |
18:49 |
celeron55 |
wrrrzr: it's likely inlined. but feel free to try it out and do benchmarking |
18:50 |
wrrrzr |
How i can benchmark drawBlockBounds? |
18:50 |
wrrrzr |
or Draw3D pipeline? |
18:51 |
MTDiscord |
<luatic> Going by the name, drawBlockBounds is entirely insignificant for performance. It just draws a couple lines at block bounds for debugging purposes if the feature is enabled (as the name implies). |
18:51 |
MTDiscord |
<luatic> And even if it had a performance problem it wouldn't matter much since it's a purely debug tool. |
18:52 |
wrrrzr |
Ahh, yes |
18:54 |
|
wrrrzr left #minetest-dev |
18:54 |
MTDiscord |
<luatic> have you looked at https://github.com/minetest/minetest/blob/master/doc/developing/misc.md ? if you want to measure time spent e.g. in "draw3d pipeline" (?), you can do that using tracy. |
18:54 |
MTDiscord |
<luatic> you can also look at flamegraphs to see what takes how much time globally. |
18:56 |
MTDiscord |
<luatic> and finally you can also write benchmarks (we use catch2 for this, you can find some examples in the sources): usually relatively small snippets of code to test the performance of a very specific function (or in this case language construct) which can be isolated cleanly. |
19:22 |
|
BuckarooBanzai joined #minetest-dev |
22:30 |
|
TheCoffeMaker joined #minetest-dev |
22:41 |
|
TheCoffeMaker joined #minetest-dev |
23:33 |
|
panwolfram joined #minetest-dev |