Time Nick Message 00:03 sapier hopefully those toolcaps do make any sense as of damage perspective as of digging times I don't actually understand why it's done that strange 00:04 sapier for what I understand we specify a time for each group and level 00:04 sapier and if there's a time we look at the difference between maximum level of the tool and the node 00:04 sapier if it's nonzero we divide the time we just specified for this level by the upper leveldifference 00:05 sapier so why not just specify a smaller time for that level instead of doing this strange division? 00:12 celeron55 that is one of the new designs i had to come up for early 0.4; it's designed to minimize issues in making a certain kind of level progression 00:18 sapier well but it's quite hard to understand what the result is ... and doc in lua_api doesn't help (at least not help me) 00:19 Fixer shader performance problems, interested? ----> https://forum.minetest.net/viewtopic.php?f=6&t=13763 00:19 Fixer test results 00:20 VanessaE Fixer: your tests are invalid unless you're viewing precisely the same scene in all cases. 00:20 Fixer VanessaE, it is the same scene 00:20 VanessaE can't be. one is on LinuxGaming, the other on just test? 00:20 Fixer yes 00:20 VanessaE with different MT versions yet. 00:21 VanessaE you need to test all possible combinations and chart them all 00:21 Fixer same scene for each servers 00:21 VanessaE otherwise your results are misleading 00:22 Fixer sorry for 0.4.12 j-t, client was 0.4.13-dev from 15 dec 2015 00:22 Fixer linuxgaming2 is 0.4.13-dev 00:22 Fixer VanessaE, big performance boost is evident in singleplayer too 00:23 est31 Fixer, what's the point precisely? 00:23 est31 you complain that the default settings of minetest should be more on performance 00:23 est31 ? 00:23 Fixer est31, default is shaders enabled, it is much slower 00:23 VanessaE you still need to re-test. if you have precisely, node-for-node, the same scene on both, the server's combination of mods can also cause differences in performance on the client (i.e. entities, machines/etc changing metadata just out of your view, etc) 00:23 Fixer yeah 00:23 est31 yeah, so you want it to be disabled? 00:23 Fixer more on performance 00:23 est31 but whats gained by that 00:23 Fixer est31, that is just my opinion 00:23 VanessaE now all of that said, having shaders enabled for me is also a little slower 00:23 est31 yes, fps is higher and viewing range is larger 00:23 VanessaE maybe 5% difference. 00:24 Fixer est31, much better fps and you can have larger view :} 00:24 VanessaE Xorg compositor also affects fps. 00:24 sapier we could disable textures too so noone needs to wait for them do be downloaded ;-) 00:24 est31 but still, idk, shaders make scene look better, no? 00:24 sapier just joking of course 00:24 est31 so its just a trade-off 00:24 Fixer yes 00:24 Fixer but having 50fps instead of 24 fps is a BIG difference on my setup 00:25 Fixer VanessaE, if you have some time, please test on your PC, you can try the same spots :}, i will edit post to add more configs 00:26 sapier as you say "on your setup" ... it's quite subjective and obviously you're capable to change the settings yourself 00:29 Fixer i would prefer 50 fps with shaders off :} If you are interested, you can try testing on your systems and compare 00:30 sapier well right now I'm more interested how "hand" can dig dirt ... as of groupcaps it shouldn't be able to 00:30 sapier wait misinterpreted 00:31 Fixer from reading the forum it feels like people complain about performance, so lets give them performance first, shaders off has no big impact on picture and you have BIG BOOST of fps 00:32 Fixer at least on AMD/ATI + Windows, not sure about Linux + ATI/nvidia/intel or whatever 00:32 sapier you have to read carefull "performance" ain't performance e.g. if someone complains about lag in inventory boosting frames to 100fps wont help anything 00:33 Fixer that is separate, yes 00:33 Fixer for example I have good fps without shaders, but bad stuttering is still there 00:33 Fixer don't get me started on crazy inventory lag on some servers 00:34 Fixer it is sadistic 00:46 sapier strange according to digtime a dirt node should be digged by hand in 0.7 seconds but I need about 3 seconds 01:11 RealBadAngel Fixer, even without any effects enabled shaders are handling light calculculations for whole scene. thx to that changing light level is immediate - without shaders you can see checkerboard and see light updated block by block 01:16 RealBadAngel also, they do fog too 20:29 Gael-de-Sailly Could it be possible to add a lacunarity parameter in minetest.get_perlinN 20:29 Gael-de-Sailly *? 20:34 Gael-de-Sailly https://github.com/minetest/minetest/issues/3477 20:46 Gael-de-Sailly Hi Paramat 20:46 paramat ^ hmmmm 20:47 paramat hi, yes that needs fixing 20:55 paramat actually looking at the code lacunarity should already work with 'get2d()' and 'get3d()' 20:59 Gael-de-Sailly https://github.com/minetest/minetest/blob/c2b5da735ea0c961d4f6521df9d96142c7143eee/src/script/lua_api/l_env.cpp#L682 ? 21:06 paramat i might test to try and reproduce your bug 21:24 Gael-de-Sailly no, that was my code that were broken 21:25 Gael-de-Sailly sorry 21:26 Gael-de-Sailly paramat? 21:27 Gael-de-Sailly ^ 21:27 paramat oh cool 21:28 paramat relief =) 22:01 paramat ( hmmmm ) i have a plan for spawn search. the current single method cannot cope with the new variety of mapgens (for example mgflat with ground level > 16 will cause a failure) so each mapgen needs its own dedicated spawn search method and function. i'll keep the method unchanged for mgv5/6/7 22:03 Gael-de-Sailly interesting