Current Server Time: 22:55 (Central Europe)

#apertus IRC Channel Logs

2023/02/20

Timezone: UTC


01:45
aombk
joined the channel
02:01
aombk
left the channel
02:02
aombk
joined the channel
02:07
aombk2
joined the channel
02:08
aombk2
left the channel
02:08
aombk2
joined the channel
02:09
aombk
left the channel
02:09
aombk2
left the channel
02:10
aombk2
joined the channel
02:12
aombk2
left the channel
02:12
aombk2
joined the channel
02:13
aombk3
joined the channel
02:17
aombk2
left the channel
02:21
aombk2
joined the channel
02:25
aombk3
left the channel
09:09
Vansh_Panchal
joined the channel
09:10
Vansh_Panchal
left the channel
12:12
illwieckz
left the channel
12:25
illwieckz
joined the channel
16:49
se6astian_
Still on the road so no meeting today from my side
16:59
Bertl
okidoki
17:07
anuejn
okay, I can report on some progress anyways:
17:08
anuejn
I worked a bit on the recorder (mostly because I was not home and thus couldn't do the hardware tests promised)
17:08
anuejn
you can see the work here: https://github.com/apertus-open-source-cinema/axiom-recorder/pull/41
17:08
anuejn
in summary it changes how gpu based nodes are written
17:09
anuejn
they are now kind of special cased and have a wrapper that makes them normal nodes
17:09
anuejn
they describe what they want to do in a more or less static / analytical(?) way
17:10
anuejn
in a next step this would allow us to compose them in a way that does not have a huge performance penalty
17:11
anuejn
by avoiding writing the output of one node to a buffer and reading said buffer back in the next node.
17:11
anuejn
instead we directly merge the shader code of the nodes into one and avoid the memory bandwidth cost
17:12
anuejn
this way we can write 'smaller' nodes with good conciouse and then compose them to do actually helpful stuff
17:13
anuejn
(for example one rgb to hsl node and one multiplication node that can then increase only the saturation)
17:13
anuejn
I also started testing the rockpi - beta combination but that didn't go anywhere yet
17:14
anuejn
I think I will change strategies and moving parts and first try to get a more boring pipeline to work (current ci images and no gui recorder)
17:14
anuejn
thats it from my side :)
17:15
Bertl
I did some work on the RPI2040 side and tested with two Pico Pis that one can easily program the other
17:16
Bertl
I also designed a minimal breakout board for the RP2040 which will hopefully soon allow to do some more tests regarding booting one RP2040 from the other
17:17
Bertl
the RP2040 definitely looks quite nice from the programming PoV and I think it is currently the best solution for the future Axiom Remote
17:17
anuejn
Bertl: sounds great :D
17:17
Bertl
we still need to do some testing regarding LCD/TFT but I think that should be doable with reasonable framerates
17:18
Bertl
button handling and debouncing should also be no problem, this can likely be done with the PIO code or some simple interrupt routine
17:19
Bertl
I also worked on the button board for the Axiom Beta Compact, so we should get that sorted soon as well
17:19
Bertl
that's it from my side.
17:20
anuejn
Bertl: Button card?
17:20
anuejn
does that only contain a power button or anything else?
17:21
Bertl
basically a power button and a header to reduce the wiring for now
17:22
Bertl
in the future it will also contain some logic and a status LED
17:24
anuejn
Okay
17:40
se6astian_
great progress guys, thanks for the insights!
17:40
se6astian_
arrived now
17:40
se6astian_
just quick update from me: ordering of parts for CP enclosure and around it is commencing, only a few parts left
17:40
illwieckz
left the channel
17:41
se6astian_
to clarify: not the enclosure itself, parts around it like screws, connectors, springs, etc.
17:41
se6astian_
the enclosure itself is already long ordered and being milled currently I must assume
18:49
illwieckz
joined the channel
22:05
anuejn
cool!