Current Server Time: 14:43 (Central Europe)

#apertus IRC Channel Logs

2019/04/13

Timezone: UTC


00:28
comradekingu
left the channel
00:42
futarisIRCcloud
left the channel
02:21
aombk
left the channel
03:04
futarisIRCcloud
joined the channel
05:30
Bertl_oO
off to bed now ... have a good one everyone!
05:31
Bertl_oO
changed nick to: Bertl_zZ
05:40
BAndiT1983|away
changed nick to: BAndiT1983
07:43
siddhantsahu
left the channel
07:44
siddhantsahu
joined the channel
07:45
mithro
left the channel
07:45
Guest27507
left the channel
07:45
Guest27507
joined the channel
07:47
mithro
joined the channel
10:54
rohan_
joined the channel
10:55
rohan_
hey BAndiT11983
10:58
rohan_
why are the function in Adapter/IAdapter.h defined as pure virtual functions?
10:59
rohan_
would we not be declaring them at all in the future?
11:02
se6astian|away
changed nick to: se6astian
11:04
rohan_
hey se6astian
11:10
BAndiT1983
changed nick to: BAndiT1983|away
11:11
rohan_
left the channel
11:54
se6astian
changed nick to: se6astian|away
12:42
se6astian|away
changed nick to: se6astian
13:22
Bertl_zZ
changed nick to: Bertl
13:22
Bertl
morning folks!
13:56
anuejn
morning :)
14:11
se6astian
great to see progress on firmware usb front!
14:11
se6astian
well done
14:13
se6astian
beside gabe the community effort wasn't that great after the call to action via email unfortunately...
14:16
anuejn
well there was someone contributing a patch
14:16
Bertl
probably related to the fact that there are not many developers among the Early Beta owners
14:16
anuejn
yup
14:16
vup2
the guys contributing fixing the path's are gsoc's studendts
14:16
vup2
*people
14:16
se6astian
gsoc helped a bit definitely, great to see activity from students!
14:17
anuejn
but i can see a trend of people slowly starting to contribute small things
14:17
anuejn
which is nice
14:19
vup2
yes i think putting the issues and the current development stuff on github help a bit
14:23
anuejn
we might also want to host the builds of the other branches somewhere
14:24
anuejn
so that people could easily test these for us
14:25
Bertl
btw, any clues about the slow SD card yet?
14:25
anuejn
something like https://builds.magiclantern.fm/experiments.html
14:26
anuejn
i havent looked into it yet... vup?
14:26
vup2
not yet
14:26
vup2
i'll try to take a look today
14:26
Bertl
great! thanks!
14:26
vup2
however i don't know for certain if i have a fast enough sdcard available to properly testit
14:28
comradekingu
joined the channel
14:31
Bertl
any UHS 1 or class 10 SD card should have fast enough read speeds to go over the current limit
14:32
vup2
well at the moment i am using a class 6 SD card ;)
14:32
vup2
but i have some more lying around
14:44
futarisIRCcloud
left the channel
14:59
vup2
Bertl: what read speeds were you getting?
15:00
vup2
and expecting?
15:01
vup2
btw irc log are broken for https
15:02
Bertl
consistently half of the expected maximum ... and yes, IRC logs are from a time when https was an emerging technology :)
15:03
vup2
ok found it in the logs
15:03
vup2
so i am getting about 19-20 MB/s
15:04
Bertl
that would be an improvement
15:04
vup2
which is a bit lower then maximum (and what i am getting with the same SD card on my laptop: 24 MB/s)
15:04
vup2
but the thing is i didn't change anything
15:04
Bertl
very strange
15:04
Bertl
how do you test?
15:04
vup2
also at that speed dd is taking about 50% cpu time
15:05
vup2
just dd if=/dev/mmcblk0 of=/dev/null status=progress
15:05
Bertl
did you flush the caches before?
15:05
Bertl
alternatively use direct I/O
15:06
vup2
yes did flush caches before
15:07
Bertl
so that is with the latest SD image?
15:08
vup2
somewhat latest
15:08
vup2
ok so with direct io i am getting a consistent 17.3 MB/s
15:09
vup2
not sure why that is lower then with dropping caches and normal io
15:09
vup2
also that matches what a simple ` pv /dev/mmcblk0 > /dev/null` gets
15:26
vup2
Bertl: ok so i definitely found the may to change the frequency and it is definitely is at 50MHz
15:27
vup2
that gets about 16.5 to 17 MB/s
15:29
vup2
(not sure if that would be better with a better SD card, but this SD card gets >23MB/s with my laptop)
15:30
Bertl
well, it looks definitely better than my tests
15:30
vup2
and not using direct io (but dropping caches before a test) gets about 19 to 20MB/s which matches what xilinx found in their benchmarks
15:30
vup2
on which kernel version are you on?
15:31
vup2
some time ago we updated to 4.20 (that was together with fixing QSPI)
15:31
Bertl
ah, that was probably after my testing
15:32
vup2
ok maybe try a newer version then
16:01
Bertl
will do, care to paste a link to the latest version?
16:23
vup2
you can find them here https://github.com/apertus-open-source-cinema/axiom-beta-firmware/releases
16:23
vup2
the newest one being https://github.com/apertus-open-source-cinema/axiom-beta-firmware/releases/download/nightly%2F1d7e001d/axiom-beta-1d7e001d.img.xz
16:48
lexano
left the channel
16:48
Bertl
left the channel
16:48
anuejn
left the channel
17:06
anuejn
joined the channel
17:06
Bertl
joined the channel
17:07
lexano
joined the channel
17:19
comradekingu
left the channel
18:16
Spirit532
left the channel
18:17
Spirit532
joined the channel
19:46
BAndiT1983|away
changed nick to: BAndiT1983
19:47
comradekingu
joined the channel
19:48
BAndiT1983
rohan_, this class is an interface and pure virtual functions are used, as there is no keyword in C++ for, like in C# or Java
19:49
BAndiT1983
every interface class has I as prefix for the name
21:33
BAndiT1983
changed nick to: BAndiT1983|away
21:46
se6astian
changed nick to: se6astian|away
21:52
Bertl
off to bed now ... have a good one everyone!
21:52
Bertl
changed nick to: Bertl_zZ