#apertus IRC Channel Logs

good day

morning folks!

MEETING TIME!

the usual suspects :)

ah, hi _bluez, great!

your internships start today

metal_dent[m]: \o/

thank you!

very good

bluez/metaldent do you want to briefly tell the community here about your internships?

yes

sure

we are trying to reduce the fpga resource usage by getting rid of axi lite memory mapped registers and instead update those register values via Dynamic Partial Reconfiguration

Hi all!

and try to understand stuff in more detail

so thats it i guess... i will be working on this with Bertl for the next 5 months at least

ah, yes you just said 5 months, :)

great thanks!

my internship period is also the same, and i'll be working on my gsoc project

in the other part of the internship, me and BAndiT1983 had planned to find any security issues or vulnerabilities with remote or firmware

so i'll try to replicate this problem and will try to fix it, so that's it from my side

so you will be regulars to the monday meetings again I assume, great!

indeed :)

unfortunately the ospark order we placed recently has been rendered inoperational, bertl discovered an issue with one of the innerr layers design file, I am sure he will share more about this in his report

at the same time jlcpcb has shown to not be able to manufacture the hardware in the way we need it

at least they seem to take things seriously and so I have high hopes for the capabilities and quality

BAndiT1983 anything you want to share?

okay, BAndiT1983 ?

anyway, as se6ast1an mentioned, we screwed up the OSHpark panel

to simplify things for Tele assembly (pick and place)

this is quite hard to spot on the panel assembly, because it almost looks like the connection is there, but it isn't

I only found it after the boards have been sent to the fab, so nothing we could do there

so we can use those as is and with a little rework, we probably can at least make the main boards work for testing or development

that's it from my side for this week

ah, yes, right

and it turned out to be a tricky issue with firmware 2.x where the gpio settings disable the I2C mux on his Beta

once we figured that out, it was only a matter of selecting the correct HDMI mode and we got everything working as expected

Great!

not sure if he got to it yet

BTW, I heared from Maxim that the other 'Belgian' beta is also with you atm?

*heard

there is none I am afraid

like the OLPF we received some time ago

fingers crossed

Yes, a vaccine and beta production... what a good start of a new year :)

we can restart the AXEÖHM giveaway campaign with ampules of antibody blood!

late news :)

I worked mostly on the recorder this week

so now we have a gstreamer like architecture where a pipeline of nodes can each process images

the rewrite can make use of multiple cores by pipelining the processing steps

it now uses vulkan compute shaders instead of opengl

seems like it is somewhat faster (~60fps with the beta 4k images on my laptop)

it still lacks some way to render the result but thats what I am currently working on

however that is still very slow (~10fps for hardware h264 encoding)

(or ffmpeg)

that could be solved by using the dmabuf infrastructure vup found out but is still some work

for example one could directly read a file into a gpu buffer (not implemented yet)

many thanks!

the actual goal was originally to write a gui application and for that it seemed a bit off path to do everything in gstreamer

if I rewrote everything once again maybe I wouldintegrate it in gstreamer

however gstreamer still is really complex and that comes with a cost during development ;)

otherwise I would conclude todays meeting

sounds exciting :)

meeting concluded!

good evening!

sorry, was not at my computer, was called away