Current Server Time: 05:42 (Central Europe)

#apertus IRC Channel Logs

2020/03/04

Timezone: UTC


01:27
lexano
left the channel
01:29
lexano
joined the channel
01:43
lexano
left the channel
01:46
lexano
joined the channel
02:50
Spirit532
left the channel
02:54
Spirit532
joined the channel
03:02
Spirit532
left the channel
03:02
Spirit532
joined the channel
05:24
rajdeep
joined the channel
05:43
rajdeep
left the channel
06:07
Bertl
off to bed now ... have a good one everyone!
06:07
Bertl
changed nick to: Bertl_zZ
09:54
metal_dent[m]
BAndiT1983: Hi, for gsoc promotion you can post a tweet about apertus with proper #s and mentions :)
09:57
metal_dent[m]
(just a little suggestion ð )
11:06
se6ast1an
We did that already
13:06
ashok_singh[m]
left the channel
13:16
ashok_singh[m]
joined the channel
14:06
ashok_singh[m]
Hello everyone,
14:06
ashok_singh[m]
was the origin shifted from top left to bottom left in Drawpixel() only for developer convince or there are other reasons too?
14:07
preetimenghwani[
Hello everyone
14:07
preetimenghwani[
I am aiming towards T781 task 1.
14:07
preetimenghwani[
I am confused how to get serialised data should we use OSERDES2 for that?
14:08
preetimenghwani[
*T871
14:25
se6ast1an
<ashok_singh[m]> was the origin shifted from top left to bottom left in Drawpixel() only for developer convince or there are other reasons too? <- yes, basically that was the reason
14:33
ashok_singh[m]
I have been reading the logs and testing a few things out and concluded that the xbm imagemagick solution for the image storage in c array would be taking approximately 8~10 times more space than storing it raw in pbm. Hence I would like to purpose that we should opt for the raw storage.
14:35
se6ast1an
btw if you have some imagemagick experience could you take a look at https://lab.apertus.org/T1164
14:49
se6ast1an
would be much appreciated
14:52
ashok_singh[m]
Well I too am not an expert (just a script kiddie), but it was working fine in my imagemagick 7.0.9-27.
16:31
BAndiT1983|away
changed nick to: BAndiT1983
16:32
BAndiT1983
ashok_singh[m]: it was shifted, because of performance reasons, otherwise cache misses have degraded it a lot
16:32
BAndiT1983
the file maybe looking bigger, but the data there is smaller, it will be converted by compiler and stored in flash
16:37
ashok_singh[m]
> the file maybe looking bigger, but the data there is smaller, it will be converted by compiler and stored in flash
16:37
ashok_singh[m]
The size of the c array turns out to be same but the absolute size of the header file will get increased many foldes. So shall I avoid using raw?
16:38
BAndiT1983
have you tried to build the firmware with XC32 and to compare with radare21/cutter the output? then you would see how the data looks like when it is prepared for the pic32
16:38
BAndiT1983
*radare2/cutter
16:40
ashok_singh[m]
I have not yet used r2 or any another tool on it, but I know after building the project the size would end up same.
16:41
BAndiT1983
so what is the advantage of raw data here?
16:42
ashok_singh[m]
It will ONLY reduce the space being consumed by the repo.
16:43
Bertl_zZ
changed nick to: Bertl
16:43
Bertl
morning folks!
16:43
ashok_singh[m]
Good morning Bertl :)
16:43
BAndiT1983
few kb are not really making any difference nowadays, it's important that the data is handled easily
16:43
BAndiT1983
hi Bertl
16:44
BAndiT1983
and XBM is fitting the task, as it allows to modify things without many problems
16:44
BAndiT1983
it'S more important to save space of the pic32 flash, than to do unnecessary optimizations to input data
16:44
ashok_singh[m]
Ok, so I would be shifting to xbm.
16:44
ashok_singh[m]
Thank you
16:46
Bertl
preetimenghwani[: hey, what do you mean by OSERDES2?
16:50
preetimenghwani[
Its basically output SERDES library
16:54
Bertl
I thought you might be referring to 'a' SERDES library element, but where do you plan to use it for T871, Task 1?
17:00
preetimenghwani[
For deserialisation of bitstream i need a serialised bitstream so how to get that?
17:01
Bertl
so, you are taslking about testing your code, yes?
17:02
preetimenghwani[
Yes!
17:03
Bertl
okay, so what hardware do you plan to use for testing?
17:05
preetimenghwani[
I have virtex-5 hardware to test it
17:07
Bertl
so you can use the OSERDES there to generate test bitstream
17:07
Bertl
*bitstreams
17:08
Bertl
but you could also test your code with a testbench generating those bitstreams
17:14
preetimenghwani[
Okay i will try out and report it to you!
17:14
preetimenghwani[
Just for confirmation we aren't allowed to use ISERDES right?
17:15
Bertl
you see the bullet points on Task 1, what do they say?
17:17
sergio__[m]
joined the channel
17:20
preetimenghwani[
Yeah i read but i thought we need to write desearlisation logic
17:23
Bertl
it's up to you what hardened primitives you use or not
17:23
Bertl
you are not permitted to use IP (proprietary or open) though
17:24
preetimenghwani[
Okay thanks a lot!
17:25
Bertl
and just as note (because we had this misunderstanding once): the switching happens at runtime, not as a selection at compiletime :)
17:28
preetimenghwani[
Oh okay i get it!
17:28
preetimenghwani[
Thanks!!
17:28
Bertl
you're welcome!
17:35
ashok_singh[m]
Hi BAndiT1983 it would be really nice if you could help me with the test case which we have to implement, instead of making the window white shall I include some test image for it or make it draw a chess board using hard code?
17:38
BAndiT1983
ashok_singh[m]: first you have to fix your code before approaching tests, at the moment the base is not looking like the rest of the firmware to me
17:39
BAndiT1983
challenge task is not about results, which are also important, but to show that you know how to write readable code and replicable process, like how to convert the image
17:39
sergio__[m]
Hola ashok_singh,
17:39
sergio__[m]
Hi, are you also planning to apply for the firmware improvement project?
17:40
BAndiT1983
hope everyone is aware, that visual things will be less for the task, it's important to learn about pic32 hardware features, learn XC32 environment and understand how linker script has to be adjusted
17:51
ashok_singh[m]
<BAndiT1983 "challenge task is not about resu"> Sorry but the conversion tool was written in a hurry. I will be improvising it tomorrow after my exam.
17:52
ashok_singh[m]
> <@sergio__:matrix.org> Hola ashok_singh,
17:52
ashok_singh[m]
> Hi, are you also planning to apply for the firmware improvement project?
17:52
ashok_singh[m]
Hi, and yes I will be applying for the firmware improvement project.
17:53
BAndiT1983
XBM should already be enough at the moment, why not use IM, you would make Bertl very happy
17:54
ashok_singh[m]
<BAndiT1983 "XBM should already be enough at "> Sure will make it up and running tomorrow.
18:01
aombk
left the channel
18:03
aombk
joined the channel
19:39
niemand
joined the channel
19:39
niemand
left the channel
19:39
niemand
joined the channel
20:23
niemand
left the channel
22:00
HEGAZY
joined the channel
22:25
BAndiT1983
changed nick to: BAndiT1983|away
22:40
gsalazar
joined the channel
22:56
gsalazar
left the channel