Current Server Time: 23:06 (Central Europe)

#apertus IRC Channel Logs

2021/03/19

Timezone: UTC


06:09
BAndiT1983|away
changed nick to: BAndiT1983
06:20
Bertl_oO
off to bed now ... have a good one everyone!
06:20
Bertl_oO
changed nick to: Bertl_zZ
06:35
dcz_
joined the channel
07:55
se6ast1an
good day
07:56
markus_
joined the channel
07:58
rassu-123
joined the channel
08:07
rassu-12317
joined the channel
08:08
rassu-123
left the channel
08:28
rassu-12317
left the channel
08:30
mumptai
joined the channel
08:36
rassu-123
joined the channel
08:43
rassu-123
left the channel
09:15
Jitesh
joined the channel
09:17
Jitesh40
joined the channel
09:17
Jitesh40
left the channel
09:20
Jitesh51
joined the channel
09:21
Jitesh51
left the channel
09:26
Jitesh30
joined the channel
09:27
Jitesh
left the channel
09:30
Jitesh30
left the channel
09:40
RexOrCine1
joined the channel
09:40
RexOrCine
left the channel
09:40
syung
joined the channel
09:43
syung
left the channel
09:45
syung
joined the channel
09:45
syung
Hello everyone
09:45
syung
left the channel
09:46
syung
joined the channel
09:47
BAndiT1983
hi syung
09:48
syung
Please help me with some the facilities with freenode chat client
09:48
syung
Each time i login via my nickname it again shows that you joined and the chat diappears
09:48
syung
or it provides me another nickname kindly help me out
09:48
BAndiT1983
are you using a dedicated client or webchat?
09:49
syung
webchat web based client
09:49
syung
its like webchat.freenode.net
09:49
BAndiT1983
i would suggest to grab a dedicated IRC client as webchat is just a simple tool and fails often
09:50
BAndiT1983
pidgin or quassel for example
09:50
syung
I tried pidgin but unable to set it up.
09:51
BAndiT1983
quassel was easier to set up, as pidgin was originally developed for AIM/ICQ, at least that's my impression of using it
09:51
syung
Okay i ll try Thank You!!
10:07
syung69
joined the channel
10:18
syung96
joined the channel
10:19
syung96
left the channel
10:19
syung
left the channel
10:20
syung
joined the channel
10:23
markus_
Good morning! Here is my progress on the Draw2BitIcon() code this far: https://github.com/MarkusEngsner/AXIOM-Remote
10:24
syung69
left the channel
10:25
syung
left the channel
10:26
BAndiT1983
markus_: hi, do you have a screenshot of the result?
10:27
syung
joined the channel
10:28
markus_
BAndiT1983: There should be one in the GSOC_Draw2BitIcon.md file in the main directory
10:28
BAndiT1983
ah, sorry, haven't checked that yet
10:28
BAndiT1983
looks great!
10:38
syang123
joined the channel
10:38
syang123
left the channel
10:46
syang12345
joined the channel
10:46
syang12345
left the channel
10:47
syung
left the channel
10:48
syung
joined the channel
10:51
syung
left the channel
10:57
preetimenghwani[
joined the channel
11:00
JiteshHemji[m]
joined the channel
11:02
BAndiT1983
markus_: am just looking quickly over the code, will check after work more, but you can actually remove the case 0x0 in GetColor(), as default already covers it
11:04
markus_
BAndit1983: True, I left it this way because I wasn't sure if there should be some other handling when sending in data outside the allowed range
11:04
markus_
or if it should just fail silently
11:04
BAndiT1983
that's what i've suspected
11:05
BAndiT1983
can check what happens on the real board later, if the values are outside of the range, but the display driver chip is also doing some processing or checks
11:06
BAndiT1983
if the conversion ensures that the values are already correct, then we can use simpler code in the firmware, just one of possible ways
11:07
BAndiT1983
had no time to check the interpolation topic for days, was thinking about simple calculations with factors, which could prevent the bit shifting to extract individual ones
11:13
BAndiT1983
hm, interesting, something like that: https://stackoverflow.com/questions/18937701/combining-two-16-bits-rgb-colors-with-alpha-blending
12:03
dcz_
left the channel
12:20
markus_
that does look a lot more efficient
12:24
BAndiT1983
still have to find a profiler which doesn't need a lot of voodoo and strange builds, to be able to check execution times and also the number of calls to methods, we had a problem at some point in the firmware, where drawing was done 70 times, instead of 7
12:30
markus_
Profiling on-board or in the visualizer?
12:30
BAndiT1983
in visualiser, or at least on the PC, as on-board it's a bit of difficult topic, but we can think of something for that
12:32
BAndiT1983
number of calls is the main concern at first, to prevent performance hits, as drawing will be on-demand in the future so the drawing speed is less of concern, also dirty rectangles will help there, the framebuffer is held in display memory persistently, while powered on
12:46
BAndiT1983
by the way, while developing and optimizing the visualiser, this tool came very handy, is more interesting for people with interest in opengl, vulkan etc.: https://renderdoc.org/
13:06
markus_
cool! I've done some stuff with Qt on Embedded Linux recently, and renderdoc certainly seems like it could help with optimizations for that
13:10
se6ast1an
Setting up timer to get performance data on the remote itself would also be interesting: https://lab.apertus.org/T1192
13:26
Bertl_zZ
changed nick to: Bertl
13:26
Bertl
morning folks!
13:27
BAndiT1983
hi
13:27
satacker[m]
hi
13:28
satacker[m]
can i use sha1sum for checksum ?
13:33
satacker[m]
P.S. I will use popen
13:37
Bertl
sure, sha1sum is fine
13:40
illwieckz
left the channel
13:41
illwieckz
joined the channel
13:52
priyam
joined the channel
14:10
priyam
left the channel
14:31
abist
joined the channel
14:32
rassu_12[m]
joined the channel
14:33
abist
left the channel
14:52
priyamraj2000[m]
joined the channel
15:23
RizwanMemon[m]
joined the channel
15:41
markus_
BAndiT1983: You have probably already looked at this, but doesn't callgrind work for the profiling?
15:42
markus_
I tried it now on the Visualiser executable, and it gives number of calls and some info on time too
15:42
BAndiT1983
yep, it works too, would like to have some nice UI to get the infos instantly, like orbit does
15:42
BAndiT1983
https://github.com/google/orbit
15:43
BAndiT1983
nice, any strange peaks?
15:48
markus_
Well, the alpha blend method that you linked earlier requires only a third of the instructions that my LerpColor method did
15:49
BAndiT1983
cool, glad that it helped
15:57
BAndiT1983
seems like you know the stuff, this will help to sort out some stuff in the menu system, which causes crashes, from the debug sessions (also added exception handlers for the chip itself) it looked like stack was overwriting the heap, as pic32 is writing backwards there (this is by design)
18:07
BAndiT1983
changed nick to: BAndiT1983|away
19:46
BAndiT1983|away
changed nick to: BAndiT1983
20:16
intrac
left the channel
20:17
intrac
joined the channel
21:32
mumptai
left the channel
23:18
markus_
left the channel
23:34
BAndiT1983
changed nick to: BAndiT1983|away