| Time |
Nick |
Message |
| 00:00 |
|
tumeninodes joined #minetest-hub |
| 00:00 |
tumeninodes |
https://www.youtube.com/watch?v=ZR3EiIO22GQ |
| 00:03 |
Jordach |
tumeninodes: prepare yourself |
| 00:03 |
* Jordach |
renders again with VEGAS |
| 00:03 |
Jordach |
humble bundle had it cheaaaaaap |
| 00:06 |
tumeninodes |
how cheap? isn't pro like $600? |
| 00:06 |
Jordach |
$20 |
| 00:06 |
Jordach |
on humble bundle |
| 00:07 |
tumeninodes |
??? whaaaaa? |
| 00:07 |
Jordach |
https://www.humblebundle.com/software/vegas-pro-creative-freedom |
| 00:08 |
red-001 |
could someone that knows more about srp in minetest tell me, can server doing mitm use the data sent to it during sign up to then authenticate as that user on the server it's attacking or would it need the user to login every time it wants to authenticate as that user? |
| 00:08 |
tumeninodes |
Ah they have Xara too. If you have never tried it, I recommend you do, I think youd like it |
| 00:08 |
red-001 |
and can login data be used to enter sudo mode? |
| 00:09 |
sfan5 |
red-001, to do this kind of active mitm you would need to internet with the real client everytime you want to log into an attacking server |
| 00:11 |
sfan5 |
not sure what you mean by 'auth data' |
| 00:11 |
Megaf |
Hi |
| 00:11 |
Jordach |
tumeninodes: https://www.youtube.com/watch?v=rNZ0qKhB2tg |
| 00:11 |
sfan5 |
s/internet/interact/ btw |
| 00:11 |
Megaf |
Jordach: is the render ready yet? |
| 00:11 |
Jordach |
Megaf: ^^^ |
| 00:11 |
sfan5 |
also s/attacking/attacked/ |
| 00:12 |
Megaf |
oh boy |
| 00:12 |
red-001 |
ah so the data send during initial sign-up can't be used generate further logins, just verify them? |
| 00:12 |
Megaf |
Jordach: you made that? |
| 00:12 |
red-001 |
what about entering sudo mode? |
| 00:12 |
Jordach |
yes |
| 00:12 |
Jordach |
i may or may not have wrote a top down shooter |
| 00:12 |
Jordach |
:^))) |
| 00:13 |
Megaf |
Jordach: If I had money I'd contract you man |
| 00:13 |
Megaf |
brilliant |
| 00:13 |
Jordach |
MT isn't a goddamn Minecraft engine |
| 00:13 |
Jordach |
nor a voxel engine if you try hard enough |
| 00:13 |
Jordach |
and yes, that sucka supports the damn wardrobe API |
| 00:14 |
sfan5 |
red-001, the data exchanged during login authenticates the user exactly once; sudomode requires a login-equivalent SRP exchange to be entered |
| 00:14 |
tumeninodes |
nice Jordach. How long did it take to render? |
| 00:14 |
red-001 |
huh so a server would only need two logins for active mitm |
| 00:14 |
Jordach |
6 mins on CPU |
| 00:14 |
Jordach |
at a whopping 32mbps |
| 00:14 |
red-001 |
get one valid login tell the user the password is invalid |
| 00:15 |
red-001 |
they enter it again, then assuming they used the same password thinking they typed it wrong, it can enter sudo mode and change the password |
| 00:17 |
tumeninodes |
my rig sucks, takes 15 min to render a cube in Blender |
| 00:17 |
Jordach |
kektus |
| 00:18 |
Megaf |
my workstation has a burned GPU |
| 00:18 |
Megaf |
more like the IC that make the outputs actually |
| 00:18 |
Megaf |
GPU still works, but no output at all |
| 00:19 |
Megaf |
theres a hole in the midle ot the IF |
| 00:19 |
Megaf |
of the IC* |
| 00:19 |
tumeninodes |
Megaf, sell it on ebay as a custom |
| 00:19 |
Megaf |
https://megaf.imgur.com/all/ |
| 00:20 |
tumeninodes |
Jordach, I love the flying position though (wish that were same for MTG) |
| 00:22 |
|
DI3HARD139 joined #minetest-hub |
| 00:22 |
tumeninodes |
afk briefly |
| 00:23 |
Megaf |
Enjoy |
| 00:26 |
* Megaf |
opens a hacker website |
| 00:27 |
Megaf |
website says, please turn on javascript and reload the page |
| 00:27 |
Megaf |
how about now? |
| 00:27 |
Megaf |
I mena |
| 00:27 |
Megaf |
how about no? |
| 00:34 |
Jordach |
REEEE |
| 00:34 |
Jordach |
https://i.imgur.com/uBm9vKF.jpg |
| 00:34 |
Jordach |
minetest flight sim coming soon |
| 00:40 |
red-001 |
Jordach, what engine for that shooter? |
| 00:41 |
Jordach |
red-001: custom built |
| 00:41 |
red-001 |
including rendering? |
| 00:41 |
Fixer |
Jordach: lol |
| 00:41 |
Jordach |
i mean the actual *flight* part |
| 00:41 |
red-001 |
so what you use? |
| 00:42 |
Jordach |
something i whipped up on a christmas evening |
| 00:42 |
red-001 |
oh is this in-minetest? |
| 00:42 |
Jordach |
yes |
| 00:45 |
Megaf |
I wont be rendering much stuff in my workstation so soon https://imgur.com/a/BH4va |
| 00:50 |
tumeninodes |
This is why we don't cook on circuit boards |
| 00:51 |
tumeninodes |
Megaf, no worries... that'll buff right out |
| 00:55 |
tumeninodes |
on a side note, fired up my old XP desktop today https://www.youtube.com/watch?v=IPJJ0yWhDkg |
| 00:55 |
|
sniper338 joined #minetest-hub |
| 00:57 |
red-001 |
Megaf, what happened there |
| 00:57 |
red-001 |
looks like there was a minor fire or something |
| 00:58 |
|
sniper338 joined #minetest-hub |
| 01:05 |
|
ThomasMonroe joined #minetest-hub |
| 01:09 |
shivajiva |
middle hdmi socket looks like it was shorting +5V to ground |
| 01:10 |
Megaf |
I'd say 20 KV to ground :) |
| 01:10 |
Megaf |
red-001: blown transfromer on the street |
| 01:10 |
Megaf |
it sent several thoudands of volts instead of 220 V |
| 01:10 |
red-001 |
ouch |
| 01:11 |
ThomasMonroe |
what happened Megaf? |
| 01:11 |
Megaf |
It vaporized some stuff |
| 01:11 |
Megaf |
if you look at the HDMI conector photos on the PC |
| 01:11 |
Megaf |
under it, theres a chunk of metal missing |
| 01:11 |
red-001 |
how is your whole electric setup not fried? |
| 01:11 |
Megaf |
ThomasMonroe: I'd say a miracle from God, all that voltage and an arc/bolt and yet no fire in my home |
| 01:12 |
red-001 |
I wouldn't expect something made for 220 v to not burn with 20kv |
| 01:12 |
Megaf |
just above the PC a lot of wood and books |
| 01:12 |
ThomasMonroe |
O.O glad He kept you safe |
| 01:12 |
tumeninodes |
they were just giving you all the power you paid for, and they held back... at one shot |
| 01:13 |
Megaf |
I just added photos of the UPS, just took it apart https://imgur.com/a/BH4va |
| 01:13 |
shivajiva |
they blow with a nice green flash as the copper vaporizes, there are a couple of vids of pole pigs blowing on yt |
| 01:13 |
Megaf |
this once was a fuse https://i.imgur.com/VCGrphb.jpg |
| 01:14 |
tumeninodes |
happens frequently in my area. |
| 01:15 |
ThomasMonroe |
wow |
| 01:15 |
ThomasMonroe |
Hey, at least you will remmeber this Christmas for a while |
| 01:16 |
tumeninodes |
one would think that the power industry would have an upgrade since 1903 |
| 01:16 |
Megaf |
Notice how useless a fuse is |
| 01:16 |
red-001 |
do they have to pay damages or do they have indemnity? |
| 01:16 |
Megaf |
my PC PSU doesnt have a fuse, it's totally fine, and most of my PC too, only one broken IC |
| 01:16 |
Megaf |
red-001: we have an issurance that covered half |
| 01:17 |
Megaf |
my mom didnt know the PC was blown too so I didnt get any money for it yet |
| 01:17 |
shivajiva |
why the hdmi socket as opposed to other links in the chain |
| 01:17 |
tumeninodes |
this is why (surge protectors) |
| 01:17 |
red-001 |
shivajiva, who knows |
| 01:17 |
Megaf |
fuse should protect for surges... |
| 01:18 |
Sokomine |
Jordach: will the player fall down if the backpack fails? :-) |
| 01:18 |
Megaf |
well, my PSU has a built in surge protection, hence it been fine |
| 01:18 |
Jordach |
Sokomine: unlimited use |
| 01:18 |
shivajiva |
tranzorbs are fitted as standard on most equipment so the path to ground should have occurred much earlier in the path |
| 01:18 |
Jordach |
it's also unaffected by the anti-cheat for speed |
| 01:18 |
|
CBugDCoder joined #minetest-hub |
| 01:20 |
Sokomine |
Jordach: beeing unaffected by anti-cheat is good. that is mostly just annoying |
| 01:20 |
Sokomine |
megaf: poor ups |
| 01:22 |
Megaf |
poor nothing, it didnt do anything to protect my 38¨ monitor nor my workstation |
| 01:22 |
Megaf |
everything blown |
| 01:23 |
Megaf |
and all the surge protectors are totally unharmed |
| 01:23 |
tumeninodes |
Some years ago, I had a slew of clients ordering surge protector equipment and peripherals (even for individual phone lines), when a freak electrical storm hit the area and many businesses were hit with massive power surges |
| 01:25 |
shivajiva |
wondering how big the emp would be |
| 01:25 |
tumeninodes |
some of those companies estimated a loss of up to 1 mil $ usd per hour their phone systems were down |
| 01:27 |
tumeninodes |
I recall one companies Avaya Definity, every port was damaged similar to yours... charred, warped boards, ports fully blown |
| 01:33 |
|
Jordach joined #minetest-hub |
| 01:56 |
|
sniper570 joined #minetest-hub |
| 02:04 |
Megaf |
MinetestBot! |
| 02:04 |
MinetestBot |
Megaf! |
| 02:48 |
Megaf |
check the last 3 albums for a full coverage to the deep fry episode https://megaf.imgur.com/ |
| 02:51 |
|
ThomasMonroe joined #minetest-hub |
| 02:59 |
Megaf |
00:48 <Megaf> check the last 3 albums for a full coverage to the deep fry episode https://megaf.imgur.com/ |
| 02:59 |
Megaf |
ThomasMonroe: ^^^^ |
| 03:01 |
ThomasMonroe |
wow, if you knew the shock was coming, you could have cooked dinner instantly it looks like :D |
| 03:02 |
Megaf |
hah, indeed |
| 03:02 |
ThomasMonroe |
how was your Christmas? |
| 03:03 |
Megaf |
It was alright |
| 03:03 |
Megaf |
I traveled to Brazil to spend Christmas and New Year with my family |
| 03:03 |
Megaf |
weather is good too |
| 03:03 |
ThomasMonroe |
I have an Aunt from Brazil |
| 03:03 |
Megaf |
nothing wrong really, all good |
| 03:03 |
ThomasMonroe |
:) |
| 03:04 |
Megaf |
ThomasMonroe: https://imgur.com/a/0Tapd This is a wallpaper I created around 4 months ago. It represents the Holy Trinity |
| 03:05 |
ThomasMonroe |
:D niceley designed. Clean, yet effective |
| 03:06 |
|
EvergreenTree joined #minetest-hub |
| 03:06 |
Megaf |
Thanks :) |
| 03:06 |
ThomasMonroe |
wb EvergreenTree |
| 03:06 |
ThomasMonroe |
yw Megaf |
| 03:08 |
Megaf |
basically the shapes reassemble the cross, the circle the union, as in all is the same, all together. The golden colour means precious, important, holy, the white for peace |
| 03:08 |
ThomasMonroe |
:) cool |
| 03:08 |
Megaf |
So very Christian :P |
| 03:08 |
* ThomasMonroe |
sighs |
| 03:08 |
ThomasMonroe |
this next week is going to be hectic |
| 03:08 |
ThomasMonroe |
for me at least |
| 03:08 |
Megaf |
why? |
| 03:09 |
ThomasMonroe |
we are moving house on Saturday :) |
| 03:09 |
Megaf |
Ah |
| 03:09 |
Megaf |
know the feeling |
| 03:09 |
EvergreenTree |
I found an actually decent IRC client on Android |
| 03:09 |
ThomasMonroe |
the sad part is that we have to leave our church :( |
| 03:09 |
Megaf |
moved 6 times since 2014 |
| 03:10 |
ThomasMonroe |
heh |
| 03:10 |
Megaf |
well, I will be online on my server, playing in fullscreen, cya later :) |
| 03:10 |
ThomasMonroe |
cya Megaf |
| 03:11 |
EvergreenTree |
ThomasMonroe: Wherever you are moving to I hope you will be able to find one you like |
| 03:13 |
ThomasMonroe |
thanks, we have one that we are going to attend where our former Pastor grew up at |
| 03:14 |
|
tumeninodes left #minetest-hub |
| 03:16 |
|
EvergreenTree joined #minetest-hub |
| 03:17 |
* benrob0329 |
regrets everything |
| 03:18 |
benrob0329 |
making 256 transparency masks, why did I think this was a good idea |
| 03:18 |
|
EvergreenTree joined #minetest-hub |
| 03:19 |
CBugDCoder |
https://www.youtube.com/watch?v=2Fe11OlMiz8 My favorite 12 days of Christmas rendition XD |
| 03:23 |
CBugDCoder |
https://www.youtube.com/watch?v=7E-47VmFopE another hilarious song by the same people |
| 03:31 |
benrob0329 |
30 down, 125 left to go |
| 03:32 |
CBugDCoder |
XD have fun |
| 03:33 |
benrob0329 |
I've got a method going, should get the rest done in an hour or so |
| 03:34 |
CBugDCoder |
well if you need a break watch the above videos they are hard not to end up laughing to |
| 03:34 |
benrob0329 |
anyone have a kick-ass synth playlist or something? |
| 03:39 |
benrob0329 |
http://freemusicarchive.org/music/Lee_Rosevere/Synthcronicity/ |
| 04:01 |
ThomasMonroe |
Night all |
| 04:21 |
|
compunerd joined #minetest-hub |
| 04:33 |
sofar |
FYI https://forum.minetest.net/viewtopic.php?f=10&t=19148 |
| 04:41 |
sofar |
may be an interesting alternative for people looking for hosting |
| 06:16 |
|
nerzhul joined #minetest-hub |
| 07:04 |
sofar |
geesh, holy cow |
| 07:04 |
sofar |
someone is running a minetest server with over 7200 media files |
| 07:05 |
nerzhul |
sofar, omg |
| 07:05 |
sofar |
and they pointed their clients to the PRMS I run |
| 07:05 |
sofar |
of course, the PRMS has like none of the files they need: |
| 07:05 |
sofar |
159/7330 |
| 07:05 |
sofar |
96/6429 |
| 07:05 |
sofar |
137/7248 |
| 07:05 |
sofar |
lol |
| 07:06 |
sofar |
funny thing is, I have zero data to tell me what server they came from |
| 07:07 |
sofar |
I don't think the HTTP request from the media server includes a referrer URL/URI |
| 07:53 |
|
CWz joined #minetest-hub |
| 10:41 |
|
nore joined #minetest-hub |
| 11:26 |
|
lisac joined #minetest-hub |
| 12:06 |
|
Fixer joined #minetest-hub |
| 12:39 |
|
ThomasMonroe joined #minetest-hub |
| 12:46 |
|
ThomasMonroe joined #minetest-hub |
| 13:20 |
|
Aerozoic joined #minetest-hub |
| 13:23 |
|
ThomasMonroe joined #minetest-hub |
| 13:47 |
|
EvergreenTree joined #minetest-hub |
| 13:49 |
|
nerzhul joined #minetest-hub |
| 14:01 |
Megaf |
building a very non optimized build here -march=k8 -mtune=generic -mfpmath=sse -O0 |
| 14:03 |
VanessaE |
-funroll_loops -fracing_stripes -ftype_r_sticker -fred_paintjob -fflame_job |
| 14:03 |
VanessaE |
;) |
| 14:04 |
Jordach |
kek |
| 14:05 |
VanessaE |
^^ my standard dig at gentoo a;0 |
| 14:05 |
VanessaE |
:) |
| 14:05 |
Megaf |
my flags are flags to not make it fast |
| 14:05 |
VanessaE |
(and I"m allowed to make that joke because I ran gentoo at one time) |
| 14:05 |
benrob0329 |
-fred_paintjob gives all nodes a red tint |
| 14:05 |
VanessaE |
Megaf: dude. joke. |
| 14:05 |
Megaf |
I know :) |
| 14:06 |
VanessaE |
benrob0329: nono that's -fhlsl_shaders :P |
| 14:06 |
benrob0329 |
Haha |
| 14:06 |
benrob0329 |
This is true :P |
| 14:09 |
Megaf |
I'm basically comparing O0, Os and Ofast |
| 14:10 |
Megaf |
I want to see if Os will use less memory than O0 and if Ofast will be any fast than Os and O0, without using more memory |
| 14:11 |
sfan5 |
-Os is mostly about code size |
| 14:11 |
Megaf |
So the size of the binary, so it fits in a smaller media. |
| 14:16 |
VanessaE |
Megaf: a few MB size difference probably. |
| 14:19 |
sfan5 |
hey, those are savings in the order of several floppy disks! |
| 14:23 |
VanessaE |
heh |
| 14:31 |
Megaf |
8207064 bytes for O0 and Os bin/minetest |
| 14:41 |
Calinou |
did you run `strip` on the binary? |
| 14:42 |
Calinou |
also, you could use UPX if you really want to be frugal on file size, but it's not really worth it |
| 14:42 |
Calinou |
(it's a tad slower to open, not noticeable on modern CPUs) |
| 14:42 |
Calinou |
(distribution is going to be compressed anyway, so that's why there is no point) |
| 14:54 |
Fixer |
back in my day i had to fit jabber messanger into the floppy |
| 15:24 |
Jordach |
))) |
| 15:27 |
* Jordach |
hits render again |
| 15:31 |
|
srifqi joined #minetest-hub |
| 15:32 |
Jordach |
https://youtu.be/sWjy8pfTjog |
| 15:32 |
Jordach |
Megaf: ^ |
| 15:32 |
Jordach |
benrob0329: ^^ |
| 15:34 |
benrob0329 |
Jordach: need some interpolation on that role animation though |
| 15:34 |
Jordach |
>minetest doesn't have interpol |
| 15:34 |
Jordach |
:thinking: |
| 15:34 |
benrob0329 |
Interpolate in blender, set to keyframes :^) |
| 15:34 |
Jordach |
ew |
| 15:35 |
benrob0329 |
It'd work |
| 15:35 |
Jordach |
you can just see the server struggle with mapgen |
| 15:35 |
benrob0329 |
Are you using something other than sqlite? |
| 15:36 |
Jordach |
it's actual LuaJIT mapgen trees causing it |
| 15:43 |
* Jordach |
doubles anim speed to 60fps |
| 15:49 |
|
srifqi joined #minetest-hub |
| 16:22 |
|
IhrFussel joined #minetest-hub |
| 16:24 |
IhrFussel |
Hi guys ... I'm running the mod profiler for 10 hours now and I'm not sure if the server lag is really just caused by map loading ... after 10 hours I see many ABMs taking 10ms+ and some high globalsteps ... but it's not easy to see how much lag the globalstep actually causes when the code is running (most of the time it's just a return since X secs aren't over yet) |
| 16:25 |
IhrFussel |
Looks like I will have to measure the dtime myself inside my running code and cannot rely on the profiler for that |
| 16:26 |
IhrFussel |
Or rather µs before/after |
| 16:45 |
|
paramat joined #minetest-hub |
| 17:01 |
Megaf |
IhrFussel: Minetest is bad by design. The whole block/chunk think has to be redone from sratch |
| 17:01 |
Megaf |
do you volunteer? |
| 17:01 |
Jordach |
>bad by design |
| 17:01 |
Jordach |
LUL |
| 17:01 |
Jordach |
cubic chunks are the best |
| 17:02 |
Megaf |
Jordach: They way it was was very good long time ago |
| 17:02 |
Jordach |
current SQL forks can easily handle this |
| 17:02 |
Megaf |
when we didnt have mods and stuff |
| 17:02 |
Jordach |
it's simply DISK performance not upto scratch |
| 17:02 |
Megaf |
then you drop a pooprly implemented mod stack into a not very optimized world stuff stack and you hang everything |
| 17:02 |
Jordach |
running MT on an SSD helps |
| 17:03 |
Megaf |
disagree |
| 17:03 |
Megaf |
why is that Minetest Minimal runs without any problems? |
| 17:03 |
Jordach |
it does |
| 17:03 |
Jordach |
Megaf: there's nothing that ticks |
| 17:03 |
Megaf |
it's just too many stuff for a single threaded thing to handle |
| 17:03 |
Megaf |
the whole server hangs when loading chunks, period. |
| 17:04 |
Megaf |
it shouldn't |
| 17:04 |
Megaf |
That's bad design |
| 17:04 |
Calinou |
did you tweak server settings for performance? |
| 17:04 |
Calinou |
https://gist.github.com/Calinou/683cb0748efe867be6e18f86f206a6b7 |
| 17:04 |
Calinou |
try these |
| 17:04 |
Megaf |
very fast hardware mascarades the problem |
| 17:04 |
Calinou |
(starting from line 40) |
| 17:04 |
Megaf |
as I always say, coding for pwoerful hardware is easy |
| 17:04 |
Calinou |
also set num_emerge_threads to at least 2 |
| 17:05 |
Megaf |
Calinou: that's for generating map |
| 17:05 |
Megaf |
we are talking about map already generated and stored |
| 17:05 |
Calinou |
yes, but the Performance section also affects loading |
| 17:05 |
Calinou |
read it carefully :) |
| 17:05 |
Megaf |
Calinou: too busy right now downloading Jordachs render and watching it :) |
| 17:06 |
Megaf |
[download] Destination: Solar Plains Improved Flight System-sWjy8pfTjog.mp4 |
| 17:06 |
Megaf |
[download] 100% of 9.66MiB in 00:06 |
| 17:06 |
Calinou |
what quality is that? :P |
| 17:07 |
Megaf |
(+) Video --vid=1 (*) (h264 1280x540 27.823fps) |
| 17:07 |
Megaf |
-f 22 |
| 17:08 |
Megaf |
lets try 60 fps. -f 298 |
| 17:10 |
Megaf |
I never understood this fully full_block_send_enable_min_time_from_building = 0.0 |
| 17:11 |
Jordach |
Megaf: try the 1440p version |
| 17:11 |
Jordach |
from my copy of Vegas Pro |
| 17:11 |
Jordach |
:^) |
| 17:12 |
Megaf |
Jordach: using nouveau driver here, wont happen |
| 17:12 |
Calinou |
Megaf: it decreases block sending speed while someone is placing nodes. This predates block placement prediction |
| 17:12 |
Megaf |
:P |
| 17:12 |
Calinou |
it should probably be removed entirely |
| 17:12 |
Calinou |
as we have block placement prediction since mid-2012 |
| 17:14 |
Megaf |
emergequeue_limit_total = 64000 |
| 17:14 |
Megaf |
emergequeue_limit_diskonly = 2048 |
| 17:14 |
Megaf |
emergequeue_limit_generate = 2048 |
| 17:14 |
Megaf |
what about those ^^ |
| 17:19 |
Calinou |
I've never tweaked these |
| 17:19 |
Calinou |
try increasing these (at least doubling) and see what happens |
| 17:22 |
Megaf |
besides, server never gets starved of IO |
| 17:22 |
Megaf |
it's just a single thread using one full hardware thread |
| 17:22 |
Megaf |
and everything hanging |
| 17:22 |
Megaf |
again, bad design |
| 17:23 |
Megaf |
Do I know how to fix? no, no idea |
| 17:23 |
Calinou |
asynchronous design wasn't hip in 2010 |
| 17:23 |
Calinou |
I don't think anyone knows how to fix this |
| 17:23 |
Megaf |
Rewriting would be a good start |
| 17:24 |
paramat |
maybe open an issue to suggest block loading is in a separate thread? seems to me it should be |
| 17:24 |
Calinou |
it wouldn't be Minetest then |
| 17:24 |
paramat |
then you'll get feedback about the issues |
| 17:25 |
|
nerzhul joined #minetest-hub |
| 17:25 |
Megaf |
paramat: we do have separate threads fro map loading, dont we? |
| 17:25 |
Megaf |
though all other threads still waiting for this one to finish |
| 17:25 |
paramat |
don't know |
| 17:25 |
Megaf |
which kinda makes sense from some point of view |
| 17:26 |
Megaf |
!title |
| 17:26 |
MinetestBot |
Megaf: This Video Is 2D And 3D Simultaneously: the Pulfrich Effect - YouTube |
| 17:26 |
* Megaf |
youtube -f 18 https://youtu.be/Q-v4LsbFc5c |
| 17:26 |
benrob0329 |
Megaf: you'll want high quality |
| 17:27 |
Megaf |
not enough RAM for that |
| 17:27 |
Megaf |
Minetest taking all for itself |
| 17:27 |
benrob0329 |
Good luck with the effect then |
| 17:27 |
Megaf |
1.2 GB and still hasnt finishing loading the server |
| 17:27 |
Megaf |
I mean, connectint to the server |
| 17:28 |
benrob0329 |
Thats a lot of ram for MT |
| 17:28 |
benrob0329 |
Afaik anyways |
| 17:30 |
IhrFussel |
Megaf, if map loading was in its own thread mods couldn't cause "map pausing" I'd say |
| 17:31 |
IhrFussel |
Very simple to test: Do something HUGE with WE, walk around -> no map loading |
| 17:32 |
IhrFussel |
Or do a big loop in globalstep and try to load map parts |
| 17:32 |
benrob0329 |
Lua does have corutines, and does have threading libraries afaik |
| 17:34 |
benrob0329 |
Really I *think* we need to be able to spawn custom async functions from Lua, such as mobs or worldedit rather than trying to make the whole mod system be multithreadded |
| 17:34 |
IhrFussel |
To me it looks like map loading DEPENDS on the globalstep ... you cannot load map in between globalsteps AFAIK |
| 17:36 |
IhrFussel |
That's exactly why locked/protected nodes can be dug client side before the server makes them re-appear |
| 17:36 |
|
EvergreenTree joined #minetest-hub |
| 17:36 |
benrob0329 |
Protected nodes should just be undigable by clients |
| 17:37 |
IhrFussel |
Try to set the dedicated_server_step to 1 or so and dig a protected node...it will re-appear after 1 second and not right away |
| 17:37 |
Megaf |
benrob0329: that video is nuts |
| 17:37 |
Megaf |
very 3D |
| 17:37 |
benrob0329 |
Which if node groups can be edited could be done by giving the player a custom group, assuming we can apply it to more than just the hand |
| 17:38 |
benrob0329 |
Megaf: ikr |
| 17:38 |
benrob0329 |
I had to go find a pair of sunglasses though, I wear prescription glasses so I dont really own any |
| 17:39 |
benrob0329 |
Found an old pair of clipons though, worked ok |
| 17:42 |
Megaf |
Thinking about going to the swimming pool |
| 17:42 |
IhrFussel |
The digging of protection nodes was just an example which kinda confirms that the server doesn't send mapblocks in between steps ... mod lag really shouldn't affect the map loading, that's horrible...imagine you fall down a hole and suddenly stay MID-AIR because the server is lagging |
| 17:44 |
Megaf |
IhrFussel: eh, that happens all the time |
| 17:44 |
IhrFussel |
Because a mod is lagging the server* .. Megaf yes I know and it's time to eliminate that issue |
| 17:44 |
Megaf |
good luck |
| 17:45 |
Megaf |
I would if I knew how |
| 17:45 |
Megaf |
but I can't code a single line of Java |
| 17:45 |
* Megaf |
hides |
| 17:46 |
IhrFussel |
Luckily it's c++ not Java but same here^^ |
| 17:47 |
Megaf |
interesting |
| 17:48 |
Megaf |
I just set dedicated_server_step to 0.001 now I can no longer overtake the mapload |
| 17:48 |
Megaf |
map loads faster than I can run |
| 17:48 |
Megaf |
that confirms IhrFussel issue |
| 17:48 |
Megaf |
mapload is locked to server thread |
| 17:51 |
benrob0329 |
Its not necessarily bad that things happen on globalstep, its when you get too many long-running things on globalstep |
| 17:51 |
benrob0329 |
And/or too slow of globalstep |
| 17:52 |
benrob0329 |
Actually now that I think about it mapgen/map loading needs to be separate from globalstep, and we need custom async functions |
| 17:54 |
IhrFussel |
The highest globalstep I have is this one and the measurements are completely fake since that code only runs every 15 secs not every step globalstep[1] ..................................... | 1 | 1603962 | 344 |
| 17:55 |
IhrFussel |
0.3 ms average, 1.6 seconds max value ... those values don't help if the profiler counts every single step |
| 17:56 |
IhrFussel |
So I added my own get_us_time() measurement and will see how much lag it really causes on average after restart |
| 17:57 |
benrob0329 |
Idea: 3 threaded system |
| 17:58 |
benrob0329 |
1 thread map loading, 1 map generation, 1 for everything else |
| 17:59 |
sfan5 |
congrats you just described the current design |
| 18:00 |
Megaf |
15:22 <Megaf> again, bad design |
| 18:00 |
Megaf |
15:04 <Megaf> That's bad design |
| 18:00 |
Megaf |
15:01 <Megaf> IhrFussel: Minetest is bad by design. The whole block/chunk think has to be redone from sratch |
| 18:00 |
Megaf |
Should I repeat that again? |
| 18:00 |
benrob0329 |
sfan5: then things need to be done asynchronously, and on_generate needs to be passed to the mapgen thread |
| 18:01 |
IhrFussel |
Same about ABMs I guess...does the profiler also count the times when ABMs don't mach the chances? ABM 'mobs_animal:rat spawning' .................... | 1 | 172761 | 19521 |
| 18:01 |
benrob0329 |
In theory, anyways |
| 18:01 |
sfan5 |
Lua is single-threaded, you can't do this |
| 18:01 |
Megaf |
I will try very hard to not mention Python. But I don't promisse anything |
| 18:01 |
Megaf |
o crap |
| 18:01 |
* Megaf |
runs to the swimming pool |
| 18:02 |
Megaf |
Minetest is an amazing piece of software |
| 18:02 |
IhrFussel |
sfan5, so what you say is the engine waits for Lua on EACH mapblock loaded before it sends it to clients? |
| 18:02 |
benrob0329 |
Lua is single threaded by default |
| 18:02 |
Megaf |
the stuff it can do is trully magnificent |
| 18:02 |
sfan5 |
wat |
| 18:02 |
Calinou |
Megaf: that's why I decreased server step on my server a lot |
| 18:02 |
Calinou |
(3× faster than default) |
| 18:02 |
IhrFussel |
We are talking about making mapblock loading independent from Lua |
| 18:03 |
sfan5 |
IhrFussel: when a MapBlock is generated it (obviously) has to wait for Lua to finish doing its stuff |
| 18:03 |
Megaf |
It's a shame we are running into limitaions from our dependencies and past decisions. AKA Irrlicht and Lua |
| 18:03 |
IhrFussel |
Not generate.....load |
| 18:03 |
sfan5 |
that's why I said "wat" |
| 18:03 |
Jordach |
Megaf: Lua is used in Garry's Mod and has no damn issues |
| 18:03 |
Jordach |
same for Roblox |
| 18:03 |
sfan5 |
lua is not involved in map loading |
| 18:03 |
IhrFussel |
We are talking about loading mapblocks from disk/RAM |
| 18:03 |
Megaf |
Jordach: so it's just bad design? :) |
| 18:03 |
IhrFussel |
Then how can mods lag the map? |
| 18:03 |
Jordach |
Megaf: it's called stop using 2008 potatoes |
| 18:03 |
Megaf |
Jordach: Though I think Minetest is a lot more complex |
| 18:04 |
sfan5 |
IhrFussel: block sending happens in the server step, which Lua can delay |
| 18:04 |
IhrFussel |
Then remove it from the step, it's bad by design |
| 18:04 |
IhrFussel |
Send blocks whenever a client requests them |
| 18:04 |
sfan5 |
<sfan5> Lua is single-threaded, you can't do this |
| 18:05 |
benrob0329 |
https://github.com/torch/threads/blob/master/README.md |
| 18:05 |
benrob0329 |
http://lua-users.org/wiki/ThreadsTutorial |
| 18:05 |
benrob0329 |
Lua can be threaded, its just not by default |
| 18:05 |
sfan5 |
I'm not saying it's good design, but you can "just move it somewhere else" |
| 18:05 |
benrob0329 |
I think that can be partially blamed by C99 not having threads |
| 18:05 |
Megaf |
55* |
| 18:05 |
sfan5 |
benrob0329: good to know that all of Minetest's problems can be solved by googling 2 minutes and pasting the links here |
| 18:05 |
Jordach |
C99 didn't expect multicore |
| 18:06 |
Megaf |
really? |
| 18:06 |
* Megaf |
googles |
| 18:06 |
Megaf |
!g multicore c99 |
| 18:06 |
Megaf |
=/ |
| 18:06 |
sfan5 |
C99 has no threading capabilties built-in, but instead each OS had their own API |
| 18:06 |
sfan5 |
I don't see how this matters though |
| 18:07 |
Jordach |
sfan5: http://media.giphy.com/media/DL4LEfi4PXos0/giphy.gif |
| 18:07 |
IhrFussel |
Sending mapblocks as part of each step increases the overall load I'd imagine since the engine has to poll requests each step..unless I'm misunderstanding something |
| 18:07 |
sfan5 |
nah it doesn't increase load |
| 18:07 |
sfan5 |
it's just bad because all of the other stuff that happens can delay this |
| 18:07 |
benrob0329 |
sfan5: good to know that my "two minutes of googling" still produced results that shows Lua can be threaded, not that it will solve MTs problems. |
| 18:08 |
IhrFussel |
Well then it's still ba cause it can destroy server gameplay to an extent |
| 18:08 |
IhrFussel |
bad* |
| 18:08 |
benrob0329 |
Those tend to be more complicated than async functions, although that would probably help. |
| 18:09 |
IhrFussel |
I'm pretty sure most players wouldn't mind lagging mods if they could still explore normally |
| 18:12 |
benrob0329 |
Megaf: but seriously, Gmod has freaking world portals that track the player's POV implimented in lua |
| 18:13 |
Megaf |
link? |
| 18:14 |
|
Darcidride_ joined #minetest-hub |
| 18:14 |
benrob0329 |
Megaf: https://github.com/MattJeanes/Doors |
| 18:14 |
benrob0329 |
Take a wild guess what its used for :P |
| 18:16 |
benrob0329 |
But gmod also has render to texture and proper client prediction for mods, so it'd be impossible in Minetest |
| 18:16 |
|
Darcidride_ joined #minetest-hub |
| 18:18 |
paramat |
sfan5 if lua is not involved in block loading (as opposed to chunk generation) why can't block loading be in a separate thread? is it only because of LBMs? |
| 18:19 |
benrob0329 |
sfan5 said that block loading is on a separate thread |
| 18:19 |
sfan5 |
paramat: block loading is in a separate thread |
| 18:20 |
sfan5 |
but sending them to the client is in the serverthread, so affected by lua |
| 18:21 |
benrob0329 |
So, if the networking was separate from the lua execution, thing might load faster? |
| 18:25 |
IhrFussel |
Isn't ASIO doing exactly that? Making network operations async? |
| 18:30 |
IhrFussel |
benrob0329, https://github.com/minetest/minetest/pull/6485 |
| 18:33 |
paramat |
(yes sorry, i meant 'why not in parallel') |
| 18:35 |
IhrFussel |
The whole point of using another thread for map loading is defeated when that thread depends on the server thread anyway IMO, it's stupid |
| 18:40 |
|
Darcidride_ joined #minetest-hub |
| 18:46 |
* Jordach |
thought ASIO was this http://asio4all.com/ |
| 18:46 |
paramat |
well, it may be in another thread for other reasons |
| 18:47 |
|
Darcidride_ joined #minetest-hub |
| 18:50 |
IhrFussel |
The map loading thread is the exact same as the map generating thread... I only see 2 threads in top "Server" and "Emerge-0" |
| 18:52 |
IhrFussel |
Emerge-0 seems to do the generating, loading and saving of mapblocks |
| 18:54 |
|
Darcidride_ joined #minetest-hub |
| 18:55 |
benrob0329 |
Generating should probably be separate from loading |
| 18:55 |
benrob0329 |
on_generate being in a separate thread (mandate that its contents be thread safe? |
| 19:11 |
IhrFussel |
Generating has the exact same problem as loading: If a mod lags the server no more mapblocks are sent ... I hope someone is able to fix that |
| 19:16 |
Sokomine |
IhrFussel: so does it work reliable with no mods installed at all? |
| 19:18 |
|
Darcidride joined #minetest-hub |
| 19:19 |
IhrFussel |
Sokomine, it looked like the server still lagged when I tried it without any mods (except areas) but probably less? Mods definitely seem to make it worse cause the engine gets constantly "interrupted" when it tries to send blocks |
| 19:21 |
Sokomine |
so it still lagged even if there was nothing to do? the entire world was unkown blocks? |
| 19:22 |
IhrFussel |
No there was MTG still |
| 19:22 |
Sokomine |
try without mtg. mtg as such has some lbms |
| 19:22 |
Sokomine |
let it be a world really without any mods |
| 19:30 |
Sokomine |
or at least eliminate all lbms, abms and such. try with a modified mtg |
| 19:39 |
IhrFussel |
I'm trying to start a 2n server instance on my dedi via minetestserver --gameid minimal --worldname tw and get "connection error: aborted" |
| 19:41 |
IhrFussel |
main.cpp:e2: int main(int, char**): A fatal error occured: Connection error |
| 19:43 |
sfan5 |
trying to bind to same port? |
| 19:46 |
IhrFussel |
sfan5, fixed it ... I forgot I had redis as DB in my test world.mt |
| 19:52 |
Jordach |
why does using minetest.add_entity return userdata instead of an object ref |
| 19:53 |
sfan5 |
that's an objectref |
| 19:54 |
Jordach |
sfan5: and using ref.custom_field doesn't work |
| 19:54 |
sfan5 |
ref:get_luaentity().custom_field |
| 19:58 |
Jordach |
even though inside on_step and on_rightclick for entities get access to self.fiel |
| 19:58 |
Jordach |
and self:objref_func |
| 19:58 |
sfan5 |
don't ask me |
| 20:12 |
|
paramat joined #minetest-hub |
| 20:33 |
Fixer |
Jordach: "giphy.gif", is not direct link, requires js crap, tl;dw |
| 20:33 |
Jordach |
¯\_(ツ)_/¯ |
| 20:34 |
Fixer |
can this -----ers provide god damn direct link? they need to stop being a ----- and make internet great again |
| 20:34 |
Fixer |
those* |
| 20:40 |
|
sofar joined #minetest-hub |
| 20:41 |
|
compunerd joined #minetest-hub |
| 20:47 |
IhrFussel |
max_lag 0.5 with JUST minimal, no other heavy processes...load average is 0,20-0,30 |
| 20:52 |
|
sofar joined #minetest-hub |
| 20:56 |
IhrFussel |
As soon as I dig something or place something the CPU of minetestserver goes high and causes some lag |
| 20:56 |
IhrFussel |
What is the actual **** is wrong here |
| 20:56 |
IhrFussel |
in* |
| 20:59 |
sofar |
what are you actually running as minetestserver? |
| 20:59 |
sofar |
what OS? |
| 20:59 |
sofar |
what build? |
| 20:59 |
sofar |
what kernel? |
| 20:59 |
IhrFussel |
Stock Ubuntu 16.04 |
| 21:01 |
sofar |
is this a cloud host or a VPS or something? |
| 21:01 |
sofar |
it sounds like you're running on a rPI lol |
| 21:02 |
sfan5 |
it's a dedi actually |
| 21:02 |
Calinou |
Fixer: videos are rarely provided as direct links |
| 21:02 |
sofar |
is it just running the minetestserver? or tons of other stuff? |
| 21:02 |
Fixer |
Calinou: it is a simple gif |
| 21:02 |
Calinou |
gfycat and giphy only use the GIF format as a fallback, pretty much all users actually view videos :) |
| 21:03 |
Calinou |
because friends don't let friends use GIF in 2017 |
| 21:03 |
sofar |
stock minetest from ppa ? self-built? |
| 21:03 |
Fixer |
it is not like I would watch your ads, come on ;) |
| 21:03 |
sofar |
Debug build? |
| 21:03 |
Calinou |
uBlock Origin is going to be your friend for a long time though :) |
| 21:03 |
Fixer |
just give me the f-ing direct link, i'm tired of waiting loading your slow js-shit |
| 21:04 |
CWz |
ublock origin vs ABP |
| 21:06 |
sofar |
IhrFussel: are you still on sqlite for map backend? |
| 21:08 |
Calinou |
CWz: ABP is pretty much dead today, use uBlock Origin |
| 21:09 |
Calinou |
praise gorhill :) |
| 21:12 |
IhrFussel |
sofar, of course it's sqlite I cannot migrate to another easily...but I'm trying to find out the exact source of the lags...I'm watching top while I'm doing different things in the world |
| 21:14 |
sofar |
how large is the map.sqlite? |
| 21:15 |
sofar |
perhaps `perf top` or latencytop can help to point to the problem |
| 21:15 |
sofar |
is your drive near full? |
| 21:15 |
sofar |
what's the storage? |
| 21:20 |
ThomasMonroe |
shivajiva, you on? |
| 21:21 |
ThomasMonroe |
there is a player on ##SC-Economy wondering why you shutdown your skyblock server |
| 21:23 |
IhrFussel |
sofar, perf top lists 36% ABMHandler::apply |
| 21:23 |
IhrFussel |
Now 55% o.o |
| 21:24 |
sofar |
that's expected |
| 21:24 |
sofar |
ABM's are the worst thing ever |
| 21:24 |
IhrFussel |
Everything below is < 3 % |
| 21:24 |
IhrFussel |
Does minimal use ABMs? |
| 21:26 |
sofar |
probably not, but I'm not sure |
| 21:27 |
sofar |
it doesn't actually mean much unless those ABM's are causing mapblock changes |
| 21:27 |
IhrFussel |
But my world is just running minimal right now, how can ABM be so high?# |
| 21:28 |
sofar |
on my ITB server, the CPU is 90%+ ABMs |
| 21:28 |
Fixer |
Calinou: iirc, you can watch video via direct link, no need to load that js junk |
| 21:28 |
Calinou |
only as GIF fallback, I guess |
| 21:28 |
Calinou |
(looks worse and is much less efficient) |
| 21:29 |
sofar |
there's like 7 ABM's in minimal, lol |
| 21:29 |
IhrFussel |
No I flew a bit around with fast and see 13% triLinearInterpolationNoEase |
| 21:29 |
sofar |
that's movement code, I think |
| 21:29 |
sofar |
that should just be CPU cycles without mapblock changes |
| 21:30 |
IhrFussel |
11% Noise::gradientMap3D |
| 21:33 |
IhrFussel |
But that's only when I fly around the map which is to be expected I guess |
| 21:33 |
IhrFussel |
As soon as I stop moving it only takes a few secs before ABM Handler is at the top again... so that seems to be the culprit? |
| 21:35 |
sofar |
ABM's will always be at the top |
| 21:36 |
sofar |
as a matter of fact, if they top out over 90% then that's good |
| 21:36 |
sofar |
in a way |
| 21:36 |
sofar |
it means all your server is doing is stuff that can't be avoided even if the server is idle |
| 21:39 |
IhrFussel |
sofar, I changed all minimal ABM intervals to 300 and now they don't appear in perf top anymore |
| 21:40 |
sofar |
excellent, try placing blocks and seeing what perf top now shows as top consumers |
| 21:41 |
IhrFussel |
5% CNodeDefManager::get, 3% __memcmp_sse2 |
| 21:43 |
sofar |
so reading of map data, no biggie |
| 21:44 |
IhrFussel |
It's shocking how even just ONE ABM run every 5 minutes let's the % spike to 15 |
| 21:45 |
IhrFussel |
Should I run my real world now and watch again? |
| 21:45 |
sofar |
better ABM's are run often but with low chance |
| 21:46 |
sofar |
to reduce and spread the workload |
| 21:46 |
sofar |
maybe look at latencytop |
| 21:46 |
paramat |
triLinearInterpolationNoEase and gradientMap3D is 3D noise, it's known to be intensive |
| 21:46 |
IhrFussel |
My test world max_lag is now 0.28 |
| 21:47 |
IhrFussel |
sofar, there is no latencytop... or do I need to specify it in perf? |
| 21:48 |
sofar |
latency is a separate tool, written by a colleague of mine |
| 21:48 |
sofar |
http://manpages.ubuntu.com/manpages/wily/man8/latencytop.8.html |
| 21:48 |
IhrFussel |
Which package has it? |
| 21:49 |
sofar |
I don't use ubuntu, sorry |
| 21:49 |
sofar |
it's probably its own package? |
| 21:55 |
IhrFussel |
I can only find this PPA but not sure what to type into apt-add-repository https://launchpad.net/ubuntu/+source/latencytop |
| 22:05 |
IhrFussel |
sofar, 16% copy_page ... something from the kernel |
| 22:35 |
Jordach |
mfw https://i.imgur.com/wYdOeI9.png |
| 22:36 |
|
srifqi joined #minetest-hub |
| 22:48 |
|
Sokomine joined #minetest-hub |
| 22:51 |
Calinou |
Jordach: jetpack theft is a serious crime! |
| 22:54 |
Jordach |
Calinou: Sony Vegas is now just called VEGAS |
| 22:55 |
benrob0329 |
Blender is now just called...oh wait |
| 22:55 |
benrob0329 |
:^) |
| 22:55 |
Jordach |
Blender 2.79 |
| 22:55 |
Jordach |
:^) |
| 22:56 |
Calinou |
I tried exporting glTF from Blender, to no avail |
| 22:56 |
Calinou |
with both the Krhonos and Kuboman exporters :( |
| 22:56 |
Calinou |
guess it's still beta-quality for now |
| 22:56 |
Calinou |
(glTF 2.0, that is) |
| 22:57 |
Jordach |
glTeam Fortress? |
| 22:58 |
Calinou |
glRandomCrits |
| 22:58 |
Calinou |
nah, it's a model format |
| 22:58 |
Calinou |
(made by the Khronos group, same as OpenGL) |
| 22:59 |
Jordach |
explains why it doesn't work |
| 23:02 |
benrob0329 |
Jordach: because blender exporters in but a few formats suck? :^) |
| 23:03 |
sofar |
IhrFussel: 16% copy_page? in perf top I assume? |
| 23:03 |
|
EvergreenTree joined #minetest-hub |
| 23:03 |
|
CBugDCoder joined #minetest-hub |
| 23:05 |
IhrFussel |
sofar, yes...now ~ 5% is that ok? |
| 23:06 |
sofar |
not sure why copy_page even shows up that high, even at all |
| 23:06 |
Fixer |
doom 1 finished (via brutal doom) |
| 23:17 |
benrob0329 |
how would I use textures that are in a subdir of the main textures directory? |
| 23:20 |
sofar |
just by their base name |
| 23:27 |
Sokomine |
benrob0329: as long as mt is still beeing called mt ...:_) |
| 23:29 |
IhrFussel |
sofar, copy_page is lower now ... 7% ServerEnvironment::ServerEnvironment |
| 23:29 |
benrob0329 |
sofar: could not load image |
| 23:30 |
sofar |
IhrFussel: copy_page is a kernel function - it points to slow memory |
| 23:30 |
benrob0329 |
it'd be nice if I was told if it was an error or a filename problem |
| 23:30 |
sofar |
5% doesn't sound much but it is, I think |
| 23:30 |
sofar |
benrob0329: don't you need 0.5-dev for that? |
| 23:30 |
sofar |
is it even merged? |
| 23:31 |
benrob0329 |
for texture subdirs? |
| 23:31 |
benrob0329 |
I think I've seen it used, if not then this is going to be a cluttered texture dir |
| 23:32 |
IhrFussel |
What is ServerEWnvironment? Anything particular that could cause lag there? |
| 23:33 |
IhrFussel |
Although lag is very low right now (< 0.5 sec) ... it's higher when ABMHandler::apply appears in the list |
| 23:34 |
benrob0329 |
forums seem to be down :-/ |
| 23:38 |
benrob0329 |
yup, can't use a subdir on stable |
| 23:40 |
sofar |
IhrFussel: ServerEnv stuff is the main server thread |
| 23:42 |
IhrFussel |
Is it normal that it's that "high"? copy_page is now at 2% by the way |
| 23:43 |
sofar |
it shouldn't ever appear that high |
| 23:43 |
sofar |
either it's doing way too many copy_page() calls, or they're taking way too long |
| 23:43 |
sofar |
what kind of "dedicated" host is it? |
| 23:45 |
IhrFussel |
Lots of kernel functions appear right below ServerEnvironment like clear_page ... it's a real dedi though just not the best |
| 23:45 |
sofar |
what's `grep model.name /proc/cpuinfo |uniq` say? |
| 23:46 |
sofar |
you're not running out of RAM, are you? |
| 23:47 |
IhrFussel |
Now __memcpy_sse2_unaligned is at the top ... will post the results from that command in a few minutes |
| 23:48 |
IhrFussel |
No, RAM usage is low (less than 50%) |
