| Time |
Nick |
Message |
| 00:01 |
VanessaE |
well now you can finally enable the death animation |
| 00:12 |
RealBadAngel |
or make chest opening anim |
| 00:12 |
VanessaE |
oh yes |
| 00:12 |
VanessaE |
of course that requires animated meshnodes |
| 00:12 |
VanessaE |
or jordach's hybrid thing |
| 00:12 |
RealBadAngel |
yup |
| 00:13 |
RealBadAngel |
for starters you can use caos, but the model will remain the same when animated meshodes come in |
| 01:27 |
Taoki |
https://github.com/minetest/minetest/issues/2813 Reported the weird attachment issue I discovered today. Hopefully somedev can take a look at it... it's a pretty obvious issue, and it being random is even weirder |
| 01:28 |
Taoki |
Basically, if you attach a Lua entity to a bone, it will follow its parent smoothly by random chances, which is weird and wrong |
| 01:33 |
VanessaE |
est31: ^^^^ |
| 01:33 |
VanessaE |
that makes three :P |
| 02:02 |
|
Hijiri joined #minetest-dev |
| 02:03 |
|
Wayward_Tab joined #minetest-dev |
| 02:36 |
RealBadAngel |
iirc est31 messed something with attachements lately |
| 02:38 |
est31 |
its not my fault |
| 02:38 |
est31 |
yes I changed them |
| 02:38 |
est31 |
but without my changes Taoki coulddnt even have tested it the way they tested |
| 02:38 |
RealBadAngel |
ok folks, here it comes, #2814 |
| 02:38 |
ShadowBot |
https://github.com/minetest/minetest/issues/2814 -- Add minimap feature. by RealBadAngel |
| 02:38 |
* RealBadAngel |
hides |
| 02:39 |
RealBadAngel |
seriously, it needs review for sure and lotsa tests |
| 02:41 |
est31 |
git sais the patch doesnt apply |
| 02:41 |
est31 |
wget https://github.com/minetest/minetest/pull/2814.patch |
| 02:41 |
est31 |
git am 2372.patch |
| 02:41 |
est31 |
this \r\n issue again I guess |
| 02:42 |
est31 |
https://help.github.com/articles/dealing-with-line-endings/ |
| 02:43 |
est31 |
seems we need a .gitattributes file |
| 02:43 |
RealBadAngel |
git is broken |
| 02:43 |
RealBadAngel |
use just git apply patch |
| 02:43 |
RealBadAngel |
i saw that already |
| 02:44 |
RealBadAngel |
it fills header data wrong |
| 02:45 |
RealBadAngel |
From df508bbc97b4834b9e842239fd98c6aee48ce945 Mon Sep 17 00:00:00 2001 |
| 02:45 |
RealBadAngel |
From: RealBadAngel <maciej.kasatkin o2.pl> |
| 02:45 |
RealBadAngel |
Date: Mon, 22 Jun 2015 04:34:56 +0200 |
| 02:45 |
RealBadAngel |
time is not set |
| 02:49 |
est31 |
minetest does have a .gitattributes file |
| 02:49 |
est31 |
I wonder how you could commit that with these line endings |
| 02:49 |
est31 |
what git do you use? |
| 02:52 |
RealBadAngel |
SmartGit 6.5.9 |
| 02:55 |
est31 |
what happens when you run in the console: git config --global core.autocrlf false |
| 02:55 |
RealBadAngel |
actually nothing |
| 02:55 |
est31 |
thats good |
| 02:56 |
est31 |
now it should honour line endings |
| 02:56 |
est31 |
perhaps restart smartgit |
| 02:56 |
est31 |
and then try to see whether "files changed" |
| 02:56 |
est31 |
usually now it wants to commit the changed files |
| 02:57 |
RealBadAngel |
no changes visible |
| 02:58 |
est31 |
ok next try whats your editor |
| 02:58 |
RealBadAngel |
geany |
| 02:59 |
RealBadAngel |
please do note that such gimmicks started lately |
| 02:59 |
RealBadAngel |
my config is yrs old |
| 02:59 |
RealBadAngel |
after i break i had i noticed that something changed on git side |
| 03:00 |
RealBadAngel |
i wasnt able to use for example "git push -f" before i answered "yes" to go back to old style (behiaviour?_ |
| 03:00 |
RealBadAngel |
) |
| 03:01 |
RealBadAngel |
geez, im thinking faster than im writing |
| 03:01 |
RealBadAngel |
sorry |
| 03:02 |
est31 |
ok the internet sais to run git rm --cached -r . and git reset --hard but be warned that deletes all files even the untracked ones, back up your worlds and mods folders if you have data in there |
| 03:03 |
est31 |
perhaps do before git add . -u |
| 03:03 |
est31 |
git commit -m "Saving files before refreshing line endings" |
| 03:03 |
est31 |
then you can put that to some distant branch |
| 03:03 |
est31 |
and use it as backup |
| 03:03 |
RealBadAngel |
for storing code i use just folder |
| 03:04 |
RealBadAngel |
no github tricks |
| 03:04 |
RealBadAngel |
so its all safe |
| 03:05 |
VanessaE |
oh GOd |
| 03:05 |
RealBadAngel |
what? |
| 03:05 |
VanessaE |
that minimap commit is a FPS killer |
| 03:05 |
VanessaE |
and we're back to that shit where another window can't overlap minetest |
| 03:06 |
RealBadAngel |
damn, copy my branch |
| 03:06 |
VanessaE |
the whole window is flickery when something moves around on it |
| 03:06 |
RealBadAngel |
i start to really hate this github shit |
| 03:06 |
est31 |
thats low fps isnt it? |
| 03:06 |
VanessaE |
yeah |
| 03:06 |
RealBadAngel |
fuck, 60+ fps here |
| 03:07 |
VanessaE |
lemme try with your branch |
| 03:07 |
RealBadAngel |
with far worse hardware than yours |
| 03:07 |
RealBadAngel |
just download the zip with code |
| 03:07 |
RealBadAngel |
without that git fucker |
| 03:08 |
VanessaE |
building yours... |
| 03:08 |
RealBadAngel |
it was the same as before when i commited parallax |
| 03:08 |
RealBadAngel |
you havent got all the files needed |
| 03:09 |
VanessaE |
which means you didn't `git add` all of them :) |
| 03:10 |
RealBadAngel |
if theyre in patch it means i did |
| 03:10 |
VanessaE |
I installed the patch with <download URL.patch>; git am <file> |
| 03:10 |
RealBadAngel |
theyre in my branch |
| 03:10 |
RealBadAngel |
why AM ???? |
| 03:10 |
RealBadAngel |
git apply |
| 03:11 |
VanessaE |
because github when you download a commit as <URL.patch> you get the 'am' format. |
| 03:11 |
RealBadAngel |
AM does not work at all |
| 03:12 |
RealBadAngel |
i never used "am" option, only "apply" |
| 03:13 |
est31 |
radar mode is weird for me |
| 03:13 |
est31 |
there are huge light squares |
| 03:14 |
est31 |
fps is high |
| 03:16 |
VanessaE |
ok, compiled and loaded all my usual crap into the tree |
| 03:16 |
VanessaE |
loading the game... |
| 03:17 |
VanessaE |
MUCH better |
| 03:17 |
VanessaE |
AND OF COURSE |
| 03:17 |
VanessaE |
I am on the wrong branch :-/ |
| 03:18 |
* VanessaE |
switches and rebuilds... |
| 03:19 |
RealBadAngel |
est31, radar is for use only underground |
| 03:20 |
RealBadAngel |
it works counting air nodes in colum |
| 03:20 |
RealBadAngel |
*column |
| 03:20 |
VanessaE |
loading... |
| 03:21 |
est31 |
ah I see |
| 03:21 |
VanessaE |
ok |
| 03:22 |
est31 |
and the more air it sees, the lighter it gets |
| 03:22 |
RealBadAngel |
yeah |
| 03:22 |
VanessaE |
RealBadAngel: as you predicted, it works fine. |
| 03:22 |
VanessaE |
good performance, and there's the map on my screen now |
| 03:22 |
VanessaE |
you still need to fix the lighting direction in the map :) |
| 03:22 |
RealBadAngel |
up to be decided |
| 03:22 |
VanessaE |
(the "sun" is left of the map instead of above and left) |
| 03:23 |
RealBadAngel |
just play with it and suggest better position |
| 03:23 |
RealBadAngel |
its being set with just a vector |
| 03:24 |
VanessaE |
move the source to about the 300-degree position on the circle |
| 03:24 |
RealBadAngel |
https://github.com/RealBadAngel/minetest/commit/df508bbc97b4834b9e842239fd98c6aee48ce945#diff-02288f7b8e4ece019108f9afbc4444a5R24 |
| 03:24 |
RealBadAngel |
this is light vector |
| 03:25 |
RealBadAngel |
0.0.0 is the center of the minimap |
| 03:25 |
VanessaE |
what's the range? |
| 03:25 |
RealBadAngel |
-1 1 |
| 03:25 |
RealBadAngel |
x y z |
| 03:26 |
VanessaE |
X/Y relative to screen height, Z relative to depth into the screen? |
| 03:26 |
RealBadAngel |
to the minimap plane |
| 03:26 |
VanessaE |
which means? |
| 03:26 |
RealBadAngel |
minimap texture |
| 03:27 |
RealBadAngel |
its the scene |
| 03:27 |
VanessaE |
well anyway |
| 03:27 |
RealBadAngel |
xz is the plane, y>0 is above texture |
| 03:27 |
VanessaE |
aw fuck |
| 03:27 |
RealBadAngel |
y <0 below |
| 03:27 |
VanessaE |
that nothing-can-overlay-the-minetest-window glitch is still there |
| 03:28 |
VanessaE |
but you have to turn on aniso_mipmap to see it |
| 03:28 |
VanessaE |
and performance is shit when I do that |
| 03:28 |
VanessaE |
(and the minimap isn't even turned on) |
| 03:28 |
RealBadAngel |
pretty weird |
| 03:28 |
VanessaE |
this is a serious performance regression |
| 03:28 |
est31 |
I think position should be updated smoothly like rotation |
| 03:28 |
VanessaE |
every time the minetest window gains or loses focus, it flickers and shows me what's under the window |
| 03:29 |
VanessaE |
(my browser in this case) |
| 03:29 |
RealBadAngel |
that really starts to look that you have something broken with your system |
| 03:29 |
VanessaE |
bullshit |
| 03:29 |
est31 |
you can perhaps render a bit larger area, for movement |
| 03:29 |
VanessaE |
it works perfectly fine in minetest HEAD |
| 03:29 |
est31 |
so radius * 1.1 |
| 03:29 |
RealBadAngel |
est31, you want to see more in minimap? |
| 03:30 |
RealBadAngel |
its 256 nodes in mode#1 already |
| 03:30 |
est31 |
no just when you fly around, the map position updates in steps not smoothly like movement |
| 03:30 |
VanessaE |
and the auto-tuner is "breathing" as well |
| 03:30 |
VanessaE |
distant...far...back and forth |
| 03:30 |
VanessaE |
and about half my usual performance |
| 03:30 |
RealBadAngel |
they do update as soon as blocks get mesh |
| 03:30 |
RealBadAngel |
no sooner no later |
| 03:31 |
RealBadAngel |
minimap shows all the block that are actually visible in the world |
| 03:31 |
est31 |
for me it update about twice a second |
| 03:31 |
est31 |
even when I move around in rendered blocks, there is still a lag |
| 03:31 |
est31 |
not for rotation |
| 03:32 |
RealBadAngel |
hmmm |
| 03:32 |
RealBadAngel |
can you record a vid with that? |
| 03:32 |
VanessaE |
confirmed - performance is fine in HEAD |
| 03:32 |
VanessaE |
but in your branch it honestly sucks :( |
| 03:33 |
RealBadAngel |
so tell me hows that possible im getting 4x better fps than you do? |
| 03:33 |
VanessaE |
it's almost like you're drawing the scene twice for every frame |
| 03:33 |
VanessaE |
fuck if I know :( |
| 03:33 |
RealBadAngel |
with low end cpu and older gupi? |
| 03:33 |
RealBadAngel |
*GPU |
| 03:34 |
VanessaE |
for reference: RBA's nVidia 550 GPU gets a passmark score roughly 1/4 what my R9 280X does, yet he gets 3-4 times the performance I get. |
| 03:34 |
VanessaE |
I want to know what the actual fuck is going on,. |
| 03:34 |
VanessaE |
because now it's starting to piss me off. |
| 03:37 |
est31 |
RealBadAngel, https://www.webrtc-experiment.com/Pluginfree-Screen-Sharing/#12GPVRWV-DZM |
| 03:37 |
est31 |
you getting it? |
| 03:39 |
RealBadAngel |
no |
| 03:39 |
RealBadAngel |
it asks to install some plugins and doesnt change |
| 03:39 |
RealBadAngel |
maybe i should restart browser, hold on |
| 03:40 |
RealBadAngel |
no go, same page |
| 03:40 |
est31 |
ok app might be broken |
| 03:42 |
RealBadAngel |
VanessaE, we shall compare vanilla without shaders |
| 03:42 |
RealBadAngel |
and no fps cap |
| 03:42 |
|
zat joined #minetest-dev |
| 03:42 |
RealBadAngel |
it really looks like your system/drivers are broken |
| 03:47 |
RealBadAngel |
est31, i dont have any idea why android build is broken |
| 03:47 |
est31 |
again RealBadAngel https://www.webrtc-experiment.com/Pluginfree-Screen-Sharing/#2ZX9JP5X-R5Q |
| 03:48 |
est31 |
RealBadAngel, if you add new files, they have to be added to the android build script again |
| 03:48 |
est31 |
wait a bit you should see my screen |
| 03:48 |
est31 |
does it work? |
| 03:49 |
RealBadAngel |
yup |
| 03:49 |
est31 |
great |
| 03:49 |
est31 |
now I fly |
| 03:50 |
est31 |
see how the map moves in steps? |
| 03:50 |
est31 |
rotation is fast |
| 03:50 |
RealBadAngel |
rotation is always the same as your FPS |
| 03:51 |
RealBadAngel |
but block appearing rate is the same as they do appear in the world |
| 03:51 |
est31 |
yea thats ok |
| 03:51 |
est31 |
but the map moves at a slow rate |
| 03:51 |
RealBadAngel |
if block is not loaded minimap has no info bout it |
| 03:51 |
est31 |
see here |
| 03:51 |
est31 |
all blocks are loaded |
| 03:52 |
est31 |
I can see them in map |
| 03:52 |
est31 |
and in screen |
| 03:52 |
est31 |
so they *are* loaded |
| 03:52 |
est31 |
now I move |
| 03:52 |
est31 |
ok slow rendering |
| 03:52 |
est31 |
now I try it |
| 03:52 |
est31 |
the flying is smooth |
| 03:52 |
est31 |
but the map moves in steps |
| 03:53 |
RealBadAngel |
i see |
| 03:53 |
est31 |
even though all blocks are there |
| 03:53 |
RealBadAngel |
on my box the map fps is almost the same as world fps |
| 03:53 |
RealBadAngel |
whats your setup? |
| 03:53 |
VanessaE |
est31: what GPU do you have? |
| 03:53 |
est31 |
onboard |
| 03:53 |
RealBadAngel |
gpu doesnt count here |
| 03:54 |
RealBadAngel |
shaders do texture here just one small texture |
| 03:54 |
RealBadAngel |
all is pure cpu work |
| 03:55 |
RealBadAngel |
VanessaE, btw, if hes getting 25 fps with onboard? |
| 03:55 |
RealBadAngel |
your system is completely fucked up then ;) |
| 03:55 |
VanessaE |
or your code is doing something the DMA driver can't handle. |
| 03:55 |
RealBadAngel |
est, what bout zoom x2? |
| 03:56 |
RealBadAngel |
do you get it smoother? |
| 03:56 |
est31 |
yes |
| 03:56 |
RealBadAngel |
VanessaE, test out vanilla |
| 03:56 |
est31 |
x4 is even smoother |
| 03:57 |
RealBadAngel |
est31, what cpu/ram ? |
| 03:57 |
est31 |
6 cores 8 gb |
| 03:57 |
est31 |
that should be no problem |
| 03:57 |
RealBadAngel |
ouch? |
| 03:57 |
est31 |
is there an extra thread which generates? |
| 03:57 |
RealBadAngel |
thy the fuck i do have worst hardware here |
| 03:57 |
est31 |
yea but low paced cores |
| 03:57 |
RealBadAngel |
and get fastest working game? |
| 03:57 |
RealBadAngel |
rotfl |
| 03:58 |
est31 |
dunno 2 ghz or so |
| 03:58 |
RealBadAngel |
mt just loves me ; |
| 03:58 |
RealBadAngel |
;) |
| 03:58 |
est31 |
hehe |
| 03:58 |
VanessaE |
28-30 FPS @ 240m (drawtime = 31-32) with default textures on my vanilla minetest_game world |
| 03:58 |
VanessaE |
no shaders, no filters, no nuthin |
| 03:59 |
VanessaE |
even THAT is too slow |
| 03:59 |
RealBadAngel |
VanessaE, for me its instant 60fps |
| 03:59 |
est31 |
minimap costs me ~8 fps |
| 03:59 |
RealBadAngel |
even when i apply 256x haven fps doesnt drop |
| 04:00 |
VanessaE |
turn on shaders and I get 26-27 FPS @ 169m, drawtime of 34 |
| 04:01 |
RealBadAngel |
est31, 8fps can be because of shaders |
| 04:01 |
RealBadAngel |
please do compare it without shaders |
| 04:01 |
VanessaE |
the minimap alone costs me a drawtime of +10 |
| 04:01 |
RealBadAngel |
map will get just flat |
| 04:03 |
RealBadAngel |
est31, i can see it gets better fps now |
| 04:03 |
RealBadAngel |
but the drawtime jumps |
| 04:03 |
est31 |
yea |
| 04:04 |
est31 |
its smoother now with shaders turned off |
| 04:04 |
RealBadAngel |
its integrated gpu i think |
| 04:04 |
RealBadAngel |
here i do have drawtime 12-13 |
| 04:04 |
RealBadAngel |
no matter the textures size |
| 04:05 |
VanessaE |
have you tried it with HDX>? |
| 04:05 |
VanessaE |
HDX-512 that is |
| 04:05 |
VanessaE |
+ dreambuilder |
| 04:05 |
est31 |
I think that shouldnt matter should it? |
| 04:05 |
VanessaE |
acid test the G*d damned thing already |
| 04:05 |
est31 |
because it calculates average size |
| 04:05 |
est31 |
and at load |
| 04:05 |
est31 |
if it doesnt calculate it at load its a bad program |
| 04:05 |
RealBadAngel |
not the size |
| 04:05 |
est31 |
should be change |
| 04:05 |
est31 |
d |
| 04:06 |
RealBadAngel |
it gets color |
| 04:06 |
RealBadAngel |
a single pixel |
| 04:06 |
est31 |
yea seen that |
| 04:06 |
RealBadAngel |
only solution for that is a nodedef definition |
| 04:06 |
RealBadAngel |
i do have a code for that already |
| 04:07 |
VanessaE |
fix the aniso+mip regression first |
| 04:07 |
RealBadAngel |
but then texture pack could not override minimap colors |
| 04:07 |
RealBadAngel |
but anyway that should be used instead of sampling textures |
| 04:08 |
RealBadAngel |
thats my opinion, but VE was against it |
| 04:08 |
RealBadAngel |
VanessaE, fix your drivers first :) |
| 04:09 |
VanessaE |
my drivers are not the problem |
| 04:09 |
VanessaE |
I just told you HEAD works fine with the same settings. |
| 04:09 |
RealBadAngel |
hmmm |
| 04:10 |
RealBadAngel |
i will just repack the sources and send you them |
| 04:10 |
RealBadAngel |
but wait a minute |
| 04:11 |
* VanessaE |
waits |
| 04:11 |
RealBadAngel |
do you claim to have 60 fps with mainstream? |
| 04:11 |
VanessaE |
no |
| 04:11 |
VanessaE |
though sometimes it reaches that |
| 04:11 |
VanessaE |
on VERY plain scenes |
| 04:11 |
VanessaE |
(like an open field with few trees and no structures, with shaders turned off) |
| 04:11 |
RealBadAngel |
so my code is not the reason for you are getting low fps |
| 04:12 |
RealBadAngel |
you should have steady 100-120 fps with your gpu |
| 04:12 |
VanessaE |
your minimap_pr branch is slower than the master branch, by a huge degree |
| 04:12 |
RealBadAngel |
ive tested mt some time ago on gaming pc |
| 04:13 |
RealBadAngel |
and saw there 300-400fps on vanilla |
| 04:13 |
est31 |
RealBadAngel, fix android by adding the file to build/android/jni/Android.mk |
| 04:13 |
RealBadAngel |
est31, ok |
| 04:14 |
|
blaise joined #minetest-dev |
| 04:16 |
|
blaise joined #minetest-dev |
| 04:18 |
RealBadAngel |
VanessaE, could you possibly try another distro? live maybe? |
| 04:19 |
RealBadAngel |
to get another setup |
| 04:20 |
RealBadAngel |
i really do remember issues with radeon drivers, some were fast some not working at all or very slowly |
| 04:20 |
RealBadAngel |
i had to manually select which ones to use |
| 04:20 |
RealBadAngel |
those proposed by system never worked for me |
| 04:21 |
VanessaE |
just for shits and grins, I just checked Open Arena.. .constant 90fps there |
| 04:22 |
VanessaE |
(at 1600x1200 window size) |
| 04:22 |
VanessaE |
with everything turned way up |
| 04:23 |
VanessaE |
like 8x aniso, trilin filter, max texture detail, max geometric detail, flares turned on..... |
| 04:23 |
VanessaE |
still think my drivers are broken? |
| 04:23 |
|
blaise joined #minetest-dev |
| 04:25 |
VanessaE |
and it was only using about half of a single CPU core... |
| 04:26 |
RealBadAngel |
so idk |
| 04:26 |
RealBadAngel |
i wish i had amd gpu right now |
| 04:27 |
VanessaE |
you didn't answer my question :) |
| 04:28 |
RealBadAngel |
something is broken with display |
| 04:28 |
RealBadAngel |
atm i dont have any idea what |
| 04:28 |
RealBadAngel |
pity is that evrything is working just fine for me |
| 04:29 |
RealBadAngel |
and i do not have a setup on which it could be broken |
| 04:38 |
cheapie |
Yay, texture corruption: https://github.com/minetest/minetest/issues/2815 |
| 04:39 |
cheapie |
Let me know if you need more information, ideally within the next few hours while I'm still awake... |
| 04:41 |
RealBadAngel |
cheapie, parallax and relief mapping is about texture displacements |
| 04:41 |
RealBadAngel |
its the way how it works |
| 04:42 |
RealBadAngel |
i havent figured by now a way how to make it look good with textures that are not seamless |
| 04:44 |
cheapie |
It didn't cause corruption on the commit before that one... |
| 04:45 |
RealBadAngel |
this commit removed faulty hack to avoid the problem |
| 04:45 |
RealBadAngel |
https://github.com/minetest/minetest/commit/53efe2ef4272c4ea8c347059e096f1ebd210d9a2#diff-13d708d8b99ea3c4d320fe0b2eccba9cR38 |
| 04:46 |
RealBadAngel |
please do read commit comments |
| 04:46 |
cheapie |
I have no clue what that comment is trying to say. |
| 04:47 |
RealBadAngel |
that means by now i have no idea how to make it work with textures that are not tiling |
| 04:47 |
cheapie |
So... what was it doing before, then? |
| 04:48 |
RealBadAngel |
before there were no parallax effect |
| 04:48 |
cheapie |
Ah, so it was basically just broken, and now it still is, just not as much? |
| 04:48 |
RealBadAngel |
no |
| 04:49 |
RealBadAngel |
now mt supports relief mapping and parallax mapping |
| 04:49 |
RealBadAngel |
and those effects just "moves" textures by definition |
| 04:49 |
cheapie |
But it looks horrible, so it's still at least sort of broken in that aspect. |
| 04:49 |
cheapie |
At any rate, I guess I should probably just turn it off until it isn't. |
| 04:51 |
RealBadAngel |
at some point i will find a solution to this issue |
| 04:52 |
cheapie |
OK. |
| 04:52 |
RealBadAngel |
other games just dont use non tiling textures |
| 04:53 |
RealBadAngel |
while i forced to look for solutions and hearing how terrible advanced gfx is :P |
| 04:53 |
RealBadAngel |
*im |
| 04:53 |
RealBadAngel |
thats just a side note ;) |
| 05:09 |
VanessaE |
... |
| 05:12 |
est31 |
RealBadAngel, how are tiling textures advanced? |
| 05:50 |
|
Hunterz joined #minetest-dev |
| 06:02 |
VanessaE |
RealBadAngel: fix it yet? :)( |
| 06:03 |
est31 |
VanessaE, what was the original issue you had? |
| 06:03 |
est31 |
wasnt it also something like that? |
| 06:04 |
VanessaE |
that his minimap commit really slows down the rendering performance with high-def textures, particularly when aniso+mipmap are enabled. |
| 06:04 |
VanessaE |
in current minetest upstream it is acceptable. |
| 06:04 |
VanessaE |
in his branch, it's.. .well.. not :) |
| 06:05 |
VanessaE |
the screen flickers when it gains or loses focus and it overwrites any window placed over it, like it's drawing the scene twice |
| 06:05 |
VanessaE |
when it flickers, you can see whatever's under the minetest window, e.g. my background or browser or whatever |
| 06:05 |
VanessaE |
s/background/desktop background/ |
| 06:06 |
VanessaE |
these things aren't happening in upstream. |
| 06:11 |
|
Naked joined #minetest-dev |
| 06:15 |
est31 |
VanessaE, also when minimap is turned off? |
| 06:15 |
VanessaE |
yep. |
| 06:15 |
est31 |
thats bad |
| 06:16 |
VanessaE |
in fact I didn't even bother turning it on when I noticed the regression |
| 06:16 |
VanessaE |
(at first that is) |
| 06:16 |
VanessaE |
when I DID turn it on, it added about +10 to my drawtime. |
| 06:16 |
VanessaE |
(from low 30's to low 40's) |
| 06:19 |
est31 |
add it as comment, thats a reason to not merge yet |
| 06:20 |
VanessaE |
copied this discussion to the PR. |
| 06:24 |
VanessaE |
cheapie: you've got a card very similar to mine, yes? |
| 06:24 |
cheapie |
Huh? |
| 06:24 |
cheapie |
Video card? Radeon HD 7770. |
| 06:24 |
VanessaE |
yeah |
| 06:25 |
VanessaE |
mine's an R9 280X which is basically a 7970... |
| 06:25 |
cheapie |
VanessaE: Does that work with the radeon driver, or does it need fglrx? |
| 06:25 |
VanessaE |
mine rates about twice the speed of yours on passmark. Can you reproduce the performance regression I just discussed? |
| 06:26 |
VanessaE |
well, I know mine "works" with the open source driver, at about half the performance of the binary driver |
| 06:26 |
cheapie |
Noted. |
| 06:26 |
* cheapie |
reads the scrollback |
| 06:27 |
cheapie |
What's this thing you're wanting me to test? |
| 06:27 |
VanessaE |
load HDX 512, go do a simple test world with lots of trees, similar to this: http://digitalaudioconcepts.com/vanessa/hobbies/minetest/screenshots/random/Screenshot%20-%2006202015%20-%2009%3a15%3a02%20PM.png |
| 06:27 |
VanessaE |
turn on aniso+mipmap |
| 06:28 |
VanessaE |
and turn on shaders |
| 06:28 |
cheapie |
It'll take a bit for me to download HDX-512, hold on. |
| 06:28 |
VanessaE |
compare minetest upstream against RBA's branch |
| 06:28 |
VanessaE |
s/go do/go to/ |
| 06:28 |
cheapie |
And the branch in question is...? |
| 06:29 |
VanessaE |
https://github.com/RealBadAngel/minetest , minimap_pr branch |
| 06:30 |
cheapie |
OK. Still downloading HDX, this'll take a while. |
| 06:30 |
VanessaE |
ok |
| 06:30 |
VanessaE |
you'll need a fuckton of RAM, unless you can reproduce that ^^^ screenshot adequately in minetest_game (maybe a dense jungle or something will do, idk) |
| 06:31 |
cheapie |
I do have that. |
| 06:31 |
cheapie |
(16 GB counts, right?) |
| 06:31 |
VanessaE |
heh yep |
| 06:32 |
VanessaE |
I was using Dreambuilder here, one of your recent downloads of my Creative or Survival worlds should be adequate if you can find a dense area similar to the screenshot |
| 06:32 |
cheapie |
Once I have *all* of the parts for the new computer (meaning no longer sharing with the old one), I'll have 32 GB. |
| 06:32 |
VanessaE |
nice |
| 06:32 |
VanessaE |
gimme. :) |
| 06:32 |
cheapie |
I'll probably create a new dreambuilder world and find something. |
| 06:32 |
cheapie |
(not hard with dreambuilder) |
| 06:32 |
VanessaE |
heh |
| 06:33 |
cheapie |
Receiving objects: 30% (1300/4227), 173.23 MiB | 665.00 KiB/s |
| 06:33 |
VanessaE |
I guess you could just pull the seed and coords from my screenshot, come to think of it |
| 06:33 |
cheapie |
OK, I guess I'll do that now, and let it generate. |
| 06:33 |
VanessaE |
yeah |
| 06:35 |
VanessaE |
I don't know how plantlife's defer cache will affect the test, so I guess generate a bunch of land all around, and then return to those coords and quit so plants_lib will empty its cache |
| 06:36 |
cheapie |
I'll have to quit to change TPs anyway. |
| 06:36 |
cheapie |
(I'm currently running around the area) |
| 06:36 |
VanessaE |
ok |
| 06:37 |
VanessaE |
note the block cache can take a couple minutes to empty |
| 06:37 |
VanessaE |
(on my box, I've never managed to get more than a dozen or so blocks into the cache) |
| 06:40 |
cheapie |
Heh, it was empty when I went to quit. |
| 06:40 |
cheapie |
The most I've seen in there at exit was 7, by the way. |
| 06:40 |
cheapie |
91% done with HDX. |
| 06:41 |
cheapie |
In the meantime: here's what I'm seeing with the default TP, mipmap+aniso on: http://cheapiesystems.com/media/images/screenshot_20150622_013940.png |
| 06:42 |
|
nore joined #minetest-dev |
| 06:43 |
cheapie |
OK, HDX-512 enabled, mipmap+aniso on, waiting for it to stabilize... |
| 06:43 |
VanessaE |
cheapie: your screenshot is consistent with mine - your drawtime is 106, mine is 71 |
| 06:44 |
VanessaE |
lemme re-check that, I forgot to turn shaders et al. on to match yours. |
| 06:44 |
cheapie |
drawtime=75, FPS=12 now, with HDX on, and everything that's within range of that spot loaded (I didn't move around) |
| 06:45 |
cheapie |
VanessaE: I have parallax occlusion off (since it's broken) and all shaders, bumpmapping, etc. on. |
| 06:46 |
cheapie |
Testing with aniso+mipmap off... |
| 06:46 |
cheapie |
(still stock client) |
| 06:47 |
cheapie |
drawtime=82 FPS=12 |
| 06:47 |
VanessaE |
ok, with the same settings as your screenshot, default textures, with full range view turned on, 8 FPS @ 79.3m, drawtime of 113. |
| 06:48 |
VanessaE |
turning full-range off, it gives me 40 fps @ 74.6m, drawtime around 24 |
| 06:48 |
cheapie |
OK, so should I test with that branch now? |
| 06:48 |
VanessaE |
yeah |
| 06:48 |
cheapie |
OK, one moment. |
| 06:50 |
VanessaE |
(it's interesting that my card doesn't perform any better than yours on the basic tests sand HDX. |
| 06:50 |
VanessaE |
sans* |
| 06:50 |
cheapie |
Compiling... |
| 06:51 |
cheapie |
Remember, those last two sets of results were with HDX. |
| 06:51 |
VanessaE |
right |
| 06:51 |
VanessaE |
I'm just comparing to your non-HDX screenshot at the moment. |
| 06:51 |
cheapie |
Done compiling, in 54 seconds. |
| 06:52 |
* cheapie |
can't wait to try out the FX-9590... |
| 06:53 |
cheapie |
"Initializing nodes" seems to take way longer this time. |
| 06:54 |
VanessaE |
yeah, it will |
| 06:54 |
VanessaE |
on my box it takes a little over a minute to load this test world when HDX is enabled. |
| 06:54 |
VanessaE |
(without, perhaps 25 seconds) |
| 06:54 |
cheapie |
drawtime=81, FPS=12, with mipmap+aniso off. |
| 06:54 |
VanessaE |
now turn on aniso |
| 06:54 |
VanessaE |
+mipmap |
| 06:54 |
cheapie |
Loading... |
| 06:54 |
VanessaE |
this is where it failed for me |
| 06:55 |
cheapie |
What's different about this branch, anyway? |
| 06:55 |
VanessaE |
it adds that minimap feature |
| 06:55 |
VanessaE |
the one he's been working on |
| 06:55 |
cheapie |
What minimap feature? |
| 06:56 |
cheapie |
Ah, I see, I'm supposed to press F9 to make it do something? |
| 06:56 |
VanessaE |
http://digitalaudioconcepts.com/vanessa/hobbies/minetest/screenshots/random/screenshot_20150622_025604.png |
| 06:56 |
VanessaE |
this |
| 06:56 |
cheapie |
radeon: Failed to allocate a buffer: |
| 06:56 |
cheapie |
(snip a bunch of that) |
| 06:56 |
cheapie |
Segmentation fault |
| 06:56 |
VanessaE |
however, you don't have to turn the minimap on to see the regression |
| 06:56 |
VanessaE |
yeah, that happened to me at least once as well |
| 06:56 |
cheapie |
Try again? |
| 06:56 |
VanessaE |
yeah |
| 06:59 |
cheapie |
That worked, drawtime=80 FPS=12 |
| 07:00 |
cheapie |
Actually, wait a second. |
| 07:00 |
RealBadAngel |
cheapie, parallax effect isnt broken |
| 07:00 |
cheapie |
It ate my settings when it crashed. |
| 07:01 |
cheapie |
RealBadAngel: It produces a non-optimal result. It doesn't matter whether it's the implementation's fault, a bad idea, whatever. It doesn't work properly, so it's broken. |
| 07:01 |
VanessaE |
cheapie: yeah, I have that happen from time to time too, so what I do is flip whatever settings, then quit and restart the game |
| 07:02 |
cheapie |
Trying again with mipmap+aniso on... |
| 07:02 |
RealBadAngel |
it does exactly as it is supposed to work |
| 07:02 |
VanessaE |
cheapie: you got enough swap to handle MT pushing everything else out? |
| 07:02 |
RealBadAngel |
just dont use it if you dont like it but stop complainin its broken |
| 07:02 |
cheapie |
RealBadAngel: ...which results in bad results... |
| 07:03 |
cheapie |
VanessaE: Heh, I have no swap. Maybe that's it. |
| 07:03 |
VanessaE |
yeah, you'll want that :) |
| 07:04 |
VanessaE |
for me, MT takes about 12 GB of resident RAM for this extreme test |
| 07:04 |
* cheapie |
tries again, but logging the memory usage this time |
| 07:05 |
cheapie |
No, I only hit 10 GB in use total when it crashed. |
| 07:06 |
cheapie |
Something with VRAM maybe? I think it's only a 1GB card. |
| 07:06 |
VanessaE |
mine's 3GB |
| 07:06 |
VanessaE |
don't think that matters too much here though |
| 07:07 |
cheapie |
At any rate, it won't start, and it's not running out of RAM. |
| 07:08 |
VanessaE |
turn off your mesh cache |
| 07:08 |
VanessaE |
see if that helps |
| 07:08 |
VanessaE |
enable_mesh_cache = false |
| 07:08 |
VanessaE |
it'll save about 800MB anyway |
| 07:08 |
VanessaE |
and for this test, any performance improvement it should give by having it turned on is irrelevant right now anyway |
| 07:09 |
* cheapie |
tries |
| 07:11 |
cheapie |
Nope, still crashes around 70% into "initializing nodes", or whatever it's actually doing behind that. |
| 07:11 |
RealBadAngel |
can you run mt in gdb? |
| 07:11 |
VanessaE |
wow |
| 07:11 |
RealBadAngel |
i would like to know where it crashes |
| 07:12 |
cheapie |
OK, lemme rebuild in debug mode... |
| 07:12 |
RealBadAngel |
no |
| 07:12 |
RealBadAngel |
just gdb minetest and then after crash type "bt full" |
| 07:12 |
cheapie |
Oh, OK. |
| 07:12 |
RealBadAngel |
and copy the output |
| 07:13 |
VanessaE |
he won't get much output without debug symbols... |
| 07:13 |
RealBadAngel |
i will know where it crashed |
| 07:13 |
VanessaE |
true |
| 07:13 |
* cheapie |
waits for it to get to that point |
| 07:14 |
RealBadAngel |
VanessaE, can i see screenshot with minimap and shaders turned on? |
| 07:15 |
VanessaE |
RealBadAngel: moment. |
| 07:15 |
VanessaE |
with or without HDX? |
| 07:16 |
RealBadAngel |
whatever |
| 07:17 |
VanessaE |
loading... |
| 07:17 |
cheapie |
Well, that was interesting. |
| 07:17 |
VanessaE |
what? |
| 07:17 |
cheapie |
It started printing the crash messages to the terminal, but parts of the screen weren't updating... |
| 07:17 |
cheapie |
Then X crashed. |
| 07:17 |
VanessaE |
ouch |
| 07:18 |
cheapie |
A quick "startx" later and everything's back to normal, at least. |
| 07:18 |
cheapie |
Spewed a ton of "[drm:radeon_gem_object_create [radeon]] *ERROR* Failed to allocate GEM object (430080, 2, 4096, -12)" into dmesg. |
| 07:19 |
cheapie |
[89493.102238] [TTM] radeon 0000:01:00.0: Unable to get page 2 |
| 07:19 |
cheapie |
[89493.102313] [TTM] radeon 0000:01:00.0: Failed to fill wc pool (r:-12)! |
| 07:19 |
VanessaE |
ah you're on the open source driver, that's right |
| 07:19 |
VanessaE |
what's a GEM object? |
| 07:19 |
cheapie |
No idea. |
| 07:20 |
VanessaE |
RealBadAngel: http://digitalaudioconcepts.com/vanessa/hobbies/minetest/screenshots/random/screenshot_20150622_032014.png |
| 07:20 |
cheapie |
I also see "segfault at 10 ip 00007fb992d06f9c sp 00007ffd2564cae0 error 4 in radeonsi_dri.so[7fb992738000+778000]" in my logs. |
| 07:21 |
cheapie |
(oh, and X didn't give me *any* messages when it crashed, nor did it log anything unusual) |
| 07:21 |
VanessaE |
cheapie: looks similar to this: https://bugs.launchpad.net/ubuntu/+source/xserver-xorg-video-ati/+bug/1335612 |
| 07:22 |
VanessaE |
(and fittingly, that person was using Minecraft) |
| 07:22 |
RealBadAngel |
cheapie, that looks like bug in your drivers |
| 07:22 |
cheapie |
Yeah, probably. |
| 07:22 |
RealBadAngel |
ive also found that page on mc |
| 07:22 |
cheapie |
At any rate, I'm *NOT* installing fglrx. |
| 07:22 |
cheapie |
(fglrx is on my "never again" list, after how horribly it broke the last ~10 times I tried to use it) |
| 07:23 |
RealBadAngel |
cheapie, with that you loose about half of speed if not more |
| 07:23 |
cheapie |
Slow X is better than no X. |
| 07:23 |
RealBadAngel |
you just have to know how to install PROPER drivers thats all |
| 07:24 |
RealBadAngel |
best way is to download package and to compile them |
| 07:24 |
VanessaE |
for me, I just `sudo apt-get install fglrx fglrx-amdcccle` (I guess you will have to then run the "additional drivers" app to switch it on) |
| 07:24 |
cheapie |
I don't think a package is available in anything other than stable and maybe testing for it. |
| 07:24 |
RealBadAngel |
and not using aviable speed and at the same time agreeing to use faulty and buggy ones is just purely stupid |
| 07:24 |
VanessaE |
(which uses DKMS to compile the sources components) |
| 07:25 |
VanessaE |
cheapie: can you try the two tests (master, then RBA's branch) using HDX-256 maybe? |
| 07:25 |
VanessaE |
that should still cause enough load to be noticable |
| 07:26 |
cheapie |
AMD's installer for it either gives a cryptic error, whines that I don't have exactly the right kernel version, can't find a random file that does in fact exist, breaks X completely, or several random other things, depending on its mood. |
| 07:26 |
* cheapie |
really doesn't want to do any more "testing" tonight |
| 07:27 |
cheapie |
If I do somehow manage to get fglrx working, it almost always breaks whenever the kernel or any part of X gets updated. |
| 07:27 |
VanessaE |
RealBadAngel: by the way, after letting Minetest settle on that same scene for a few minutes, turning the minimap on costs me 11 to 12 FPS -- that is, my FPS goes down from 28-29 FPS to 17 |
| 07:28 |
cheapie |
At the very least, my experience with that branch should tell you something about its performance... |
| 07:28 |
RealBadAngel |
does turning off minimap by pressing F9 cause fps to go up? |
| 07:28 |
VanessaE |
yes. |
| 07:28 |
|
Fritigern joined #minetest-dev |
| 07:29 |
cheapie |
With HDX, it takes much longer to "Initialize nodes" than the stock client with HDX, and with mipmap+aniso on, it crashes my driver too, which the stock client doesn't do. |
| 07:29 |
cheapie |
That should tell you something. |
| 07:29 |
VanessaE |
28-30 FPS (it's trying to inch a bit higher) with the minimap off, 17-18 fps with it turned on. |
| 07:30 |
RealBadAngel |
please do compare zoom levels, like 1x and radar x1 |
| 07:30 |
VanessaE |
oooo I just found something |
| 07:30 |
RealBadAngel |
cheapie, when using much larger textures it just take longer |
| 07:31 |
RealBadAngel |
VanessaE, what? |
| 07:31 |
VanessaE |
that flickering screen / can't put a window on top of Minetest bug only happens if I move the minetest window at least once after starting (again, this does not happen with mainline) |
| 07:31 |
cheapie |
RealBadAngel: "longer" apparently means almost twice as long as the stock client with the same settings... |
| 07:31 |
RealBadAngel |
cheapie yes and thats normal with that codfe |
| 07:32 |
RealBadAngel |
*code |
| 07:32 |
cheapie |
And that's a problem, since it already takes a relatively long time. |
| 07:32 |
VanessaE |
RealBadAngel: and when I mouse in/out of the minetest window, not only can I see my desktop background during the flicker, I ALSO SEE the original 800x600 blue background up in the corner! |
| 07:32 |
VanessaE |
so Minetest IS drawing the view twice |
| 07:33 |
RealBadAngel |
lol |
| 07:33 |
VanessaE |
I suspect another AMD-specific uninitialized variable somewhere |
| 07:34 |
cheapie |
AMD CPU or GPU specific? |
| 07:34 |
VanessaE |
not unlike the ones that caused the glsl compile failures during some tests |
| 07:34 |
VanessaE |
GPU |
| 07:34 |
cheapie |
I was going to say... CPU specific would be "interesting", to say the least. |
| 07:34 |
RealBadAngel |
VanessaE, that doesnt explain why you have such low fps without shaders |
| 07:35 |
RealBadAngel |
btw whats the fps with minimap on and without shaders? |
| 07:35 |
VanessaE |
I'm about to check that now. |
| 07:36 |
* cheapie |
looks at his client that is exhibiting no fewer than three bugs at the moment, and wishes that people would fix them instead of constantly adding new features instead |
| 07:39 |
VanessaE |
without shaders, with HDX-512, and having only aniso+mipmap turned on 30-31 fps @ 35m, drawtime of 32-34 with the minimap disabled. With it turned on, 16-18 FPS @ 35m, drawtime of 52-54. |
| 07:39 |
VanessaE |
so it still costs FPS even without shaders. |
| 07:39 |
RealBadAngel |
yes, texture is being made each frame |
| 07:40 |
VanessaE |
and my overall performance in this particular situation is about the same with or without shaders |
| 07:40 |
VanessaE |
the flickering bug is still there even without shaders. |
| 07:40 |
RealBadAngel |
thats actually good |
| 07:40 |
VanessaE |
also: when I quit the game, I get a blue window filled with clouds for a second ot two |
| 07:41 |
* VanessaE |
turns off aniso+mipmap |
| 07:41 |
RealBadAngel |
at least we do know that shaders are not the reason |
| 07:41 |
VanessaE |
in theory yeah |
| 07:42 |
RealBadAngel |
can you grab the screenshot of that blue window? |
| 07:42 |
VanessaE |
I'll try |
| 07:42 |
RealBadAngel |
second or two should be enough time :) |
| 07:44 |
VanessaE |
ok with aniso+mipmap off, so literally about as low of settings as I can get, 43 FPS @ 89.6m, drawtime of 22, with the minimap disabled. With it turned on, 29-30 fps @ 89.6m, drawtime of 31-32. |
| 07:47 |
VanessaE |
waiting for it to load... |
| 07:47 |
VanessaE |
(quit from the game in that ^^^^ state did not produce any unusual appearance) |
| 07:49 |
VanessaE |
having some trouble figuring out the pattern of settings and actions to cause the flicker effect/ |
| 07:52 |
VanessaE |
ok, the pattern is load the game, resize the window once or twice and then drag between screens. That causes the flicker to start happening. |
| 07:53 |
VanessaE |
but the world has to be visible when you do that |
| 07:53 |
VanessaE |
it won't do it while it's still loading. |
| 07:54 |
VanessaE |
that ^^^ common pattern of actions for me, because my launcher (and the terminal pointing at your branch) are on my middle screen, while I've been moving the game window to my left screen to use it. |
| 07:55 |
VanessaE |
that's* |
| 07:55 |
VanessaE |
in fact, moving it between screens is unnecessary, just resizing the window while the world is being shown is enough. |
| 07:55 |
VanessaE |
bingo, captured the screenshot |
| 07:56 |
VanessaE |
this is right after clicking "exit to menu": http://digitalaudioconcepts.com/vanessa/hobbies/minetest/screenshots/random/Screenshot%20-%2006222015%20-%2003%3a56%3a00%20AM.png |
| 07:57 |
VanessaE |
opinions? |
| 07:57 |
VanessaE |
<RealBadAngel> your driver is broken :) |
| 07:58 |
VanessaE |
note that the window was maximized when I exited. |
| 07:58 |
VanessaE |
it switched to the contents of this screenshot a second or so after "exit to menu", then went a normally-displayed menu. |
| 07:59 |
VanessaE |
(which filled the still-maximized window) |
| 07:59 |
RealBadAngel |
that looks like loading screen |
| 07:59 |
VanessaE |
it does, yes. |
| 07:59 |
RealBadAngel |
we can see simply check that |
| 07:59 |
VanessaE |
but this is after exiting from the world. |
| 07:59 |
RealBadAngel |
hold on a sec |
| 08:00 |
VanessaE |
notice the crosshairs are still visible? |
| 08:00 |
|
Yepoleb_ joined #minetest-dev |
| 08:00 |
VanessaE |
lower right corner of the clouds/sky section |
| 08:00 |
RealBadAngel |
https://github.com/minetest/minetest/blob/master/src/drawscene.cpp#L502 |
| 08:00 |
RealBadAngel |
insert: return; here |
| 08:01 |
|
Darcidride joined #minetest-dev |
| 08:01 |
RealBadAngel |
to just dont draw loading scene |
| 08:01 |
VanessaE |
ok, moment... |
| 08:01 |
VanessaE |
that's line 505 in your branch... |
| 08:01 |
RealBadAngel |
you wont be able to see loading screen but we will check if it affects your performance |
| 08:01 |
RealBadAngel |
ah, can be |
| 08:02 |
RealBadAngel |
ive modified that file too |
| 08:02 |
RealBadAngel |
that should be #505 exactly |
| 08:02 |
ShadowBot |
https://github.com/minetest/minetest/issues/505 -- Add seed parameter for default and L-system trees by RealBadAngel |
| 08:02 |
RealBadAngel |
rotfl |
| 08:02 |
VanessaE |
heh. |
| 08:02 |
VanessaE |
anyway |
| 08:02 |
VanessaE |
lemme try it |
| 08:03 |
VanessaE |
<insert XKCD "compiling!" reference here> |
| 08:04 |
VanessaE |
ok, loading... |
| 08:04 |
VanessaE |
and sure enough, it isn't drawing the loading screen. |
| 08:04 |
* VanessaE |
waits... |
| 08:07 |
RealBadAngel |
still waiting? |
| 08:08 |
VanessaE |
making a couple of tests.. |
| 08:08 |
VanessaE |
ok, it's loaded. 27-28 FPS @ 35m, drawtime of 33-34. Exiting to the menu takes a couple of seconds after clicking, but goes straight to the menu. |
| 08:08 |
|
nore joined #minetest-dev |
| 08:08 |
VanessaE |
no intervening half-drawn screen like in my screenshots. |
| 08:08 |
VanessaE |
-s |
| 08:09 |
VanessaE |
also, it takes less time to exit to the menu from my minetest_game world than from dreambuilder - time spent freeing memory maybe? |
| 08:09 |
VanessaE |
(roughly half a second, versus around 2 seconds) |
| 08:11 |
RealBadAngel |
thats pretty weird. with loading screeen it takes circa 1s to launch mt game. without it world loads instantly |
| 08:11 |
VanessaE |
also, I can't reproduce the flickering problem with minetest_game, so it's probably a RAM usage issue triggering it |
| 08:12 |
VanessaE |
or it's.. wait. |
| 08:12 |
VanessaE |
I'm gonna re-enable the loading screen and see what happens. |
| 08:13 |
|
Etzos joined #minetest-dev |
| 08:14 |
|
Lunatrius joined #minetest-dev |
| 08:14 |
VanessaE |
nope, can't reproduce it with the plain old minetest_game world |
| 08:14 |
VanessaE |
so you have to test it like I've been testing to see the problems I'm seeing :) |
| 08:15 |
|
Anchakor joined #minetest-dev |
| 08:15 |
VanessaE |
dreambuilder + HDX + various combinations of settings/options. |
| 08:20 |
VanessaE |
RealBadAngel: next? :) |
| 08:20 |
RealBadAngel |
open drawscene.cpp |
| 08:21 |
VanessaE |
open. |
| 08:21 |
RealBadAngel |
http://pastebin.com/wYBwY0hd |
| 08:21 |
RealBadAngel |
line 419, replace whole function |
| 08:21 |
VanessaE |
eep |
| 08:21 |
VanessaE |
ok |
| 08:22 |
VanessaE |
done. |
| 08:22 |
RealBadAngel |
recompile and check |
| 08:23 |
VanessaE |
compiling... |
| 08:23 |
VanessaE |
intersting, you took out all the 3d mode calls? |
| 08:23 |
VanessaE |
still compiling... |
| 08:24 |
VanessaE |
loading... |
| 08:28 |
VanessaE |
interesting |
| 08:28 |
VanessaE |
that cause me a small boost: 34-35 FPS @ 35m, drawtime of 29-30, minimap still costs about 10 fps. |
| 08:28 |
VanessaE |
er that GAVE me a small boost. |
| 08:33 |
|
RealBadAngel joined #minetest-dev |
| 08:33 |
VanessaE |
RealBadAngel: see my result? |
| 08:34 |
RealBadAngel |
nope, my connection lagged |
| 08:34 |
RealBadAngel |
what you got? |
| 08:34 |
VanessaE |
replacing that function gave me a small boost: 34-35 FPS @ 35m, drawtime of 29-30, minimap still costs about 10 fps. |
| 08:35 |
RealBadAngel |
how much you gained? |
| 08:36 |
VanessaE |
that's 9-10 fps more. |
| 08:36 |
|
Amaz joined #minetest-dev |
| 08:37 |
VanessaE |
well, more like 8-9 |
| 08:37 |
VanessaE |
but anyway, a significant increase. |
| 08:39 |
RealBadAngel |
kinda impossible |
| 08:39 |
VanessaE |
beats me why but there it is. |
| 08:40 |
|
troller joined #minetest-dev |
| 08:40 |
RealBadAngel |
you know what i have disabled? |
| 08:40 |
VanessaE |
compiler optimization failure maybe? |
| 08:40 |
RealBadAngel |
sapier's 3d modes |
| 08:40 |
VanessaE |
yeah, you turned off all the 3d stuff (which I don't use anyway) |
| 08:41 |
RealBadAngel |
8-9 fps is damn much if you have circa 30fps after all |
| 08:41 |
VanessaE |
yes |
| 08:41 |
RealBadAngel |
ive started dreambuilder world too |
| 08:42 |
RealBadAngel |
for me its leaves that slow rendering down |
| 08:42 |
VanessaE |
yes, but by THIS much? makes no sense. |
| 08:42 |
* VanessaE |
turns fancy leaves off, just to see what it does. |
| 08:44 |
RealBadAngel |
here doubles the framerate |
| 08:44 |
VanessaE |
waiting for it to load... |
| 08:46 |
VanessaE |
no effect. in fact I'm back down in the 25 fps range again |
| 08:46 |
VanessaE |
(25 FPS @ 35m, drawtime of 37-38) |
| 08:46 |
VanessaE |
s/no effect/no improvement/ |
| 08:49 |
RealBadAngel |
but the replaced function gain is still in effect? |
| 08:50 |
VanessaE |
in theory it is, but that boost may be a transient thing :P |
| 08:50 |
VanessaE |
just checked with shaders turned off. still in the same general range - 28 fps @ 35m, drawtime of 34-35 |
| 08:52 |
* VanessaE |
turns fancy leaves back on |
| 08:54 |
VanessaE |
no effect. |
| 08:55 |
VanessaE |
actually it ticked up to 30 fps after letting it sit. |
| 08:55 |
VanessaE |
so statistical noise there. |
| 09:29 |
|
nore joined #minetest-dev |
| 10:43 |
|
Ritchie joined #minetest-dev |
| 10:43 |
|
Robby joined #minetest-dev |
| 10:43 |
|
nore joined #minetest-dev |
| 10:43 |
|
nolsen joined #minetest-dev |
| 10:43 |
|
Darcidride joined #minetest-dev |
| 10:48 |
|
Amaz joined #minetest-dev |
| 10:51 |
|
realbadangel__ joined #minetest-dev |
| 10:55 |
|
daswort joined #minetest-dev |
| 11:10 |
|
RealBadAngel joined #minetest-dev |
| 11:27 |
|
Calinou joined #minetest-dev |
| 11:48 |
RealBadAngel |
hi Calinou |
| 12:32 |
|
blaze joined #minetest-dev |
| 14:13 |
|
zat joined #minetest-dev |
| 14:22 |
|
CraigyDavi joined #minetest-dev |
| 14:24 |
|
TenPlus1 joined #minetest-dev |
| 14:24 |
TenPlus1 |
Hi folks... |
| 14:25 |
TenPlus1 |
any joy on the std::bad_alloc ..or.. unrecoverable map errors as yet ??? |
| 14:29 |
TenPlus1 |
been getting a bundle of those today for some reason: http://pastebin.com/tUZdZDAt |
| 14:57 |
|
blaise joined #minetest-dev |
| 15:06 |
|
hmmmm joined #minetest-dev |
| 15:27 |
|
Dartmouth joined #minetest-dev |
| 15:30 |
|
neoascetic joined #minetest-dev |
| 15:31 |
neoascetic |
Voxus cancelled https://forum.minetest.net/viewtopic.php?p=182476#p182476 |
| 15:34 |
sfan5 |
kenney switches to a new engine due to minetest engine really not being ready for this would be a better tl;dr |
| 15:36 |
Calinou |
Half-Life 3 confirmed |
| 15:37 |
Calinou |
meanwhile, people recreate iOS' interface using Microsoft Word... |
| 15:39 |
rom1504 |
lolwat |
| 15:50 |
|
ElectronLibre joined #minetest-dev |
| 15:50 |
RealBadAngel |
Calinou, here? |
| 15:52 |
Calinou |
yes |
| 15:53 |
RealBadAngel |
can you try minimap pr? |
| 16:08 |
Calinou |
not now, sorry, gtg |
| 16:09 |
RealBadAngel |
hmmmm, are you here? |
| 16:13 |
|
Hunterz joined #minetest-dev |
| 16:21 |
|
rubenwardy joined #minetest-dev |
| 16:24 |
|
Krock joined #minetest-dev |
| 16:25 |
|
zat joined #minetest-dev |
| 16:30 |
|
est31 joined #minetest-dev |
| 16:31 |
est31 |
#2754 is ready for a second +1 |
| 16:31 |
ShadowBot |
https://github.com/minetest/minetest/issues/2754 -- Fix bug when craft input isn't replaced by TeTpaAka |
| 16:38 |
rubenwardy |
Take note: https://forum.minetest.net/viewtopic.php?p=182476#p182476 |
| 16:38 |
rubenwardy |
Voxus is gone |
| 16:38 |
rubenwardy |
Minetest confirmed for a terrible game engine |
| 16:39 |
rubenwardy |
I just wish I had the skills to meaningfully contribute. |
| 16:39 |
est31 |
seen it |
| 16:39 |
Krock |
wat |
| 16:49 |
est31 |
its a hard choice |
| 16:50 |
est31 |
giving freedom to people on one hand |
| 16:50 |
est31 |
and ensuring compatibility between mods on the other |
| 16:56 |
rubenwardy |
It's freedom to subgames |
| 16:56 |
rubenwardy |
A subgame can choose their mods |
| 16:56 |
rubenwardy |
Also, why does freedom take away compatibility? |
| 16:58 |
est31 |
the more freedom you have the more you have to care about how mods interact |
| 17:01 |
RealBadAngel |
est31, ive worked a lot with VanessaE to find out reason for her problems |
| 17:01 |
RealBadAngel |
her GPU drivers are busted |
| 17:02 |
RealBadAngel |
she went to sleep in rather bad mood |
| 17:02 |
est31 |
lol |
| 17:02 |
est31 |
ok |
| 17:02 |
|
Gethiox joined #minetest-dev |
| 17:02 |
rubenwardy |
If Minetest is a game engine, then you need higher levels of flexibilirt |
| 17:03 |
RealBadAngel |
est31, she should have far better fps when she fixes those drivers, atm my card which is way slower runs circa 2x faster than her R9 |
| 17:04 |
est31 |
ok |
| 17:05 |
RealBadAngel |
btw, funny thing |
| 17:05 |
RealBadAngel |
ive enabled hardware buffering of meshes |
| 17:05 |
RealBadAngel |
added removing buffers when mesh gets deleted |
| 17:05 |
RealBadAngel |
and everythin just runs fine |
| 17:06 |
|
JZTech101 joined #minetest-dev |
| 17:07 |
|
blaise joined #minetest-dev |
| 17:08 |
|
Gethiox left #minetest-dev |
| 17:08 |
|
JZTech101 joined #minetest-dev |
| 17:09 |
RealBadAngel |
huh, almost fine ;) |
| 17:13 |
|
Gethiox joined #minetest-dev |
| 17:21 |
|
blaise joined #minetest-dev |
| 17:21 |
|
Gethiox joined #minetest-dev |
| 17:26 |
|
VanessaE joined #minetest-dev |
| 17:43 |
|
MinetestForFun joined #minetest-dev |
| 17:54 |
Guest80272 |
woohoo, I have 54,223 total views and 76 subs :D |
| 17:55 |
nolsen |
erm wrong channel |
| 18:04 |
|
Elinvention joined #minetest-dev |
| 18:41 |
|
Elinvention joined #minetest-dev |
| 18:44 |
|
zat joined #minetest-dev |
| 18:49 |
|
est31 joined #minetest-dev |
| 19:13 |
|
Amaz joined #minetest-dev |
| 19:26 |
|
Robert_Zenz joined #minetest-dev |
| 20:48 |
|
ElectronLibre left #minetest-dev |
| 21:23 |
|
AnotherBrick joined #minetest-dev |
| 21:24 |
|
luizrpgluiz joined #minetest-dev |
| 21:27 |
|
luizrpgluiz left #minetest-dev |
| 21:29 |
|
luizrpgluiz joined #minetest-dev |
| 21:29 |
luizrpgluiz |
hi |
| 21:29 |
luizrpgluiz |
I liked the mod whitelist, although it could have on a next update of minetest :) |
| 21:47 |
|
TeTpAaka joined #minetest-dev |
| 21:53 |
|
deltib_ joined #minetest-dev |
| 21:56 |
|
Lunatrius joined #minetest-dev |
| 21:56 |
|
AnotherBrick joined #minetest-dev |
| 22:06 |
|
luizrpgluiz left #minetest-dev |
| 22:11 |
|
Lunatrius joined #minetest-dev |
| 22:15 |
|
Lunatrius` joined #minetest-dev |
| 22:36 |
RealBadAngel |
somebody remember why VBO patch wasnt merged? |
| 22:36 |
RealBadAngel |
kahrl? |
| 22:39 |
kahrl |
memleaks I think, but I don't really remember the details |
| 22:41 |
RealBadAngel |
i found your patch |
| 22:41 |
RealBadAngel |
its basically ok but you forgot one thing for it to be working |
| 22:41 |
RealBadAngel |
when camera offset has changed mesh in hw buffer has to be reloaded |
| 22:42 |
RealBadAngel |
m_mesh->setDirty(); is enough |
| 22:43 |
RealBadAngel |
http://irc.minetest.ru/minetest-dev/2014-03-10#i_3612932 |
| 22:43 |
RealBadAngel |
that fixes above issue |
| 22:43 |
kahrl |
oh good point |
| 22:44 |
kahrl |
I might have assumed that was already fixed in SN's branch |
| 22:44 |
RealBadAngel |
it wasnt |
| 22:44 |
RealBadAngel |
freeminer also doesnt care bout reloading meshes |
| 22:45 |
RealBadAngel |
and yet has vbo in ;) |
| 22:51 |
RealBadAngel |
im testing now VBO with dreambuilder and it looks stable, doubles the fps |
| 23:35 |
|
Lunatrius joined #minetest-dev |
| 23:56 |
|
VanessaE joined #minetest-dev |
