<tkaiser>
PaddyF: Thx for the feedback. In case you run into poblems please report in the Armbian forum (most if not all Armbian devs use their devices headless so feedback is highly appreaciated)
<PaddyF>
okay, great
<PaddyF>
even the xorg.conf suits my needs
<tkaiser>
PaddyF: Since this is a pre-release could you please do an 'sudo armbianmonitor -u' and provide the support URL? Just to double check if board auto detection works
<PaddyF>
amazing how some people can read sense in such text
ninolein has quit [Ping timeout: 240 seconds]
ninolein has joined #linux-sunxi
egbert has quit [Disconnected by services]
egbert has joined #linux-sunxi
mpmctoo has quit [Ping timeout: 250 seconds]
cnxsoft has joined #linux-sunxi
Andy-D has quit [Ping timeout: 276 seconds]
p1u3sch1 has quit [Ping timeout: 252 seconds]
p1u3sch1 has joined #linux-sunxi
hipboi has joined #linux-sunxi
Andy-D has joined #linux-sunxi
hipboi has quit [Client Quit]
Andy-D has quit [Ping timeout: 248 seconds]
hulu1522_ has joined #linux-sunxi
hulu1522 has quit [Ping timeout: 244 seconds]
Andy-D has joined #linux-sunxi
atsampso1 has quit [Ping timeout: 264 seconds]
atsampson has joined #linux-sunxi
vagrantc has quit [Ping timeout: 240 seconds]
p1u3sch1 has quit [Ping timeout: 252 seconds]
p1u3sch1 has joined #linux-sunxi
huawei has quit [Read error: Connection reset by peer]
Shirasaka-Hazumi has joined #linux-sunxi
tomboy64 has quit [Ping timeout: 248 seconds]
tomboy65 has joined #linux-sunxi
hulu1522_ has quit []
robogoat has quit [Ping timeout: 276 seconds]
tomboy65 has quit [Ping timeout: 246 seconds]
tomboy65 has joined #linux-sunxi
tomboy65 has quit [Ping timeout: 250 seconds]
tomboy65 has joined #linux-sunxi
tomboy65 has quit [Ping timeout: 244 seconds]
tomboy65 has joined #linux-sunxi
AA78 has joined #linux-sunxi
<AA78>
hey longsleep
JohnDoe_71Rus has joined #linux-sunxi
<AA78>
whenever you got a moment mate I would like to discuss helping out with the pine64
IgorPec7 has joined #linux-sunxi
phil42 has quit [Ping timeout: 276 seconds]
<AA78>
msg when you are around
tomboy64 has joined #linux-sunxi
tomboy65 has quit [Ping timeout: 244 seconds]
cosm has joined #linux-sunxi
cnxsoft has quit [Quit: cnxsoft]
tomboy64 has quit [Ping timeout: 252 seconds]
AA78 is now known as help
help is now known as Guest40997
Guest40997 is now known as AA78
cnxsoft has joined #linux-sunxi
tomboy64 has joined #linux-sunxi
AA78 has quit [Quit: Leaving.]
AA781 has joined #linux-sunxi
tomboy64 has quit [Ping timeout: 244 seconds]
tomboy64 has joined #linux-sunxi
AA781 has quit [Quit: Leaving.]
IgorPec7 has quit [Ping timeout: 244 seconds]
AA781 has joined #linux-sunxi
cnxsoft has quit [Quit: cnxsoft]
p1u3sch1 has quit [Ping timeout: 260 seconds]
p1u3sch1 has joined #linux-sunxi
phil42 has joined #linux-sunxi
<steev>
PaddyF: well there's usually more than just that
cnxsoft has joined #linux-sunxi
<ordex>
wigyori: any clue why I can't find r8188eu.ko anywhere although I have selected it in the menuconfig of openwrt ?
<ordex>
wigyori: it looks like the source files gets compiled (I can see the .o files in the module foler) bu no .ko is produced
<steev>
that's pretty impressive that you were able to overheat yours; i run gentoo on 2 of them without issue compiling everything but chromium and llvm
<ordex>
wigyori: mh maybe it's built in :P
<KotCzarny>
ordex, or included in some other .ko
<KotCzarny>
one module to drive them all
<ordex>
KotCzarny: nope, it is supposed to come as standalone module
<KotCzarny>
try running make modules again then
IgorPec7 has joined #linux-sunxi
<KotCzarny>
maybe there was some error?
<KotCzarny>
you can also try touching the .c file before
<KotCzarny>
or even make V=1 modules
<ordex>
KotCzarny: I am crosscompiling inside the openwrt environment - doing a make V=1 modules is not the easiest thing :D
<ordex>
but the .o is created, so it gets compiled
<ordex>
btw i just checked the .config and it was really =y
<KotCzarny>
Planned for 4.6 : Initial bringup , timer, watchdog and reboot
<poffs>
so what does this mean? it'll boot, but no kernel upgrades?
<KotCzarny>
it means the only support you will get from manufacturer, ie. no support
<KotCzarny>
and the kernel is ancient 3.4.something
<poffs>
i have never expected any support with any linux device lol
<poffs>
so no kernel updates
tomboy64 has quit [Ping timeout: 240 seconds]
<KotCzarny>
kernel updates is part of the support
tomboy64 has joined #linux-sunxi
<poffs>
so... if a device is considered supported, you can just apt-get update; apt-get upgrade ? without breaking anything, I mean
<KotCzarny>
it's all about uboot and kernel support (regarding linux-sunxi)
<poffs>
ok
<KotCzarny>
once you have uboot and kernel you can just use any linux distro compiled for arm
Netlynx has joined #linux-sunxi
<KotCzarny>
Without a heatsink you won't exceed 1.2GHz when running CPU intensive tasks, when choosing a good one and enough airflow you might get up to 1.6 GHz and everything above needs an annoying fan. The good news: This is still no overvolting/overclocking since it's just about better heat dissipation to avoid throttling. The bad news: You need to solder a sane way to provide more than 5V/2A to the board
<KotCzarny>
and that's regarding cpu speed
<poffs>
more than 5v/2a?
<poffs>
really?
<KotCzarny>
apparently yes
<KotCzarny>
octacore at high speeds can eat some power
<KotCzarny>
these conditions CPU activity is responsible for the board consuming between 5.5W-6W
<KotCzarny>
Due to the dvfs settings Vcore will be increased to 1000mV or even 1080mV when reaching 1800 MHz. And then CPU activity alone is responsible for a whopping 8W consumption
cajg has joined #linux-sunxi
<PaddyF>
is the 8W consumption typical for the M3 or for the M1 aswell?
bonbons has joined #linux-sunxi
<KotCzarny>
m1 eats 1-2W
<KotCzarny>
they are different socs, so you cant compare
<PaddyF>
i wonder why i get shitty google results since i changed my DNS to 8.8.8.8
<KotCzarny>
paddyf: google often has caches at local isps
<GeneralStupid>
dont know exactly what nomachine nx is, we used it around 2008 for X forwarding. If i remember correctly it just adds x forwarding modern technologies
<KotCzarny>
GeneralStupid: x forwarding isnt efficient on anything slower than 100mbit
<KotCzarny>
that's why vnc was born (among other things)
<KotCzarny>
and keep in mind that bandwidth eats cpu on allwinner chips
wingrime has joined #linux-sunxi
IgorPec7 has quit [Ping timeout: 252 seconds]
<wingrime>
nice to remeber me
<GeneralStupid>
KotCzarny: take a look at nx, it is efficient but it is x forwarding (kind of)
<GeneralStupid>
i know. i have big problems with wa - and i dont know exactly why. Bought a new sd card but it does not really speed things up
<KotCzarny>
From 2013, with the release of version 4.0, NX technology became closed source.
tomboy65 has joined #linux-sunxi
<GeneralStupid>
KotCzarny: my system administrator career is over since 2009 ;) so i cant know that
<KotCzarny>
just pasted from the wiki you linked ;)
tomboy64 has quit [Ping timeout: 240 seconds]
<KotCzarny>
i prefer command line anyway
<KotCzarny>
i even wrote me an audio player because i wasnt able to find any with the requirements i needed
<GeneralStupid>
yes, me too. But i like the concept of x forwarding (the concept, not the 60th moon landing science in it)
<GeneralStupid>
ssh -X i guess and then it works :D
leio has joined #linux-sunxi
<KotCzarny>
yeah, for a quickie, but its not efficient
<KotCzarny>
even if you turn on the ssh compression
<GeneralStupid>
no thats what i meant with "moonlanding science"
<GeneralStupid>
the concept is nice
<GeneralStupid>
not its technical implementation
tomboy64 has joined #linux-sunxi
tomboy65 has quit [Ping timeout: 244 seconds]
tomboy65 has joined #linux-sunxi
tomboy64 has quit [Ping timeout: 244 seconds]
tomboy65 has quit [Ping timeout: 240 seconds]
tomboy65 has joined #linux-sunxi
tkaiser has joined #linux-sunxi
<tkaiser>
poffs: The next 10k batch of Banana Pi M3 comes with a barrel design for DC-IN so if you improve heat dissipation it clocks up to 1.8GHz
p1u3sch1 has quit [Ping timeout: 276 seconds]
p1u3sch1 has joined #linux-sunxi
sdschulze has joined #linux-sunxi
<GeneralStupid>
my wlan link (usb) from my orangepi pc is SOOO bad
<KotCzarny>
buy something better?
<GeneralStupid>
the "router" is just 3 meters away, my laptop gets rates around 200 Mbit/s and the orangepi is around 20 Mbit/s ...
<GeneralStupid>
KotCzarny: i tried two wlan sticks, with different chips
<KotCzarny>
move usb irq to another core?
<GeneralStupid>
they worked both fine in my laptop
<premoboss>
GeneralStupid, both stick usb are good on laptop and bad on orangepi?
<GeneralStupid>
i already use irqbalace
<GeneralStupid>
premoboss: yes
<premoboss>
can you monitor the cpo load during wifi connection? i gues id one core is 100% busy or not.
avph has quit [Ping timeout: 246 seconds]
<GeneralStupid>
premoboss: yes it is
<GeneralStupid>
premoboss: mpv gets a load of 100% and wa raises, if wa lowers the load goes to normal
<premoboss>
so it seems cpu is doing pulling, no irq, no dma
<GeneralStupid>
around 20 or 30 %
<premoboss>
it seems to be a bad software configuration.
<GeneralStupid>
but dma should really be used...
<premoboss>
yes, "should".
<GeneralStupid>
do i need to build mpv again?
<GeneralStupid>
how to debug this kind of 'errors'?
<KotCzarny>
watch /proc/interrupts ?
<KotCzarny>
compare with pc?
<premoboss>
GeneralStupid, i dont know your particoular software setup, i can only suggest you to dig inside the souce code and/or compilation parameters.
<GeneralStupid>
it looks like that only cpu0 is really doing interrupts
<tkaiser>
GeneralStupid: yes, we use systemd on jessie. In case you want to remove it I'm not interested in feedback. But you can try out the image anyway, just needs 2 GB SD card and a few minutes patience on the first boot (reboots one time)
<KotCzarny>
but essentially you can just take bits of startup scripts from armbian
<tkaiser>
KotCzarny: There's a bit more than startup scripts, eg. optimised mpv in this desktop build...
<GeneralStupid>
tkaiser: but thats not "DMA" related or is it?
zuikis has left #linux-sunxi [#linux-sunxi]
<GeneralStupid>
i dont get why wa consumes so many cpu when there is a dma controller
<tkaiser>
No, 'just' libvdpau (jemk) since you asked about how to build mpv
<GeneralStupid>
iam running an 0.13 built of mpv with libvdpau
<GeneralStupid>
i upgraded libvdpau a few days ago
<GeneralStupid>
maybe i should also upgrade mpv
<GeneralStupid>
tkaiser: i just need to add, it works a lot better with the new irq setting...
<KotCzarny>
told'ya
tomboy65 has quit [Ping timeout: 244 seconds]
<GeneralStupid>
KotCzarny: thanks
<GeneralStupid>
tkaiser: thanks
tomboy65 has joined #linux-sunxi
IgorPec has joined #linux-sunxi
avph has quit [Ping timeout: 246 seconds]
avph has joined #linux-sunxi
<tkaiser>
GeneralStupid: In case you still have irqbalanced installed you should drop it. It has a memory leak and does nothing unless you use Russell King's fixed version on ARM
zuikis has joined #linux-sunxi
ricardocrudo has joined #linux-sunxi
avph has quit [Ping timeout: 246 seconds]
vishnup has quit [Quit: Leaving]
<KotCzarny>
shouldnt irqbalance be done by kernel alone?
<longsleep>
tkaiser: did you add the new dvfs_table to u-boot or only load the changed files by Kernel?
avph has joined #linux-sunxi
<tkaiser>
longsleep: I replaced both u-boot-with-dtb.bin and /boot/pine64/sun50i-a64-pine64-plus.dtb
<longsleep>
tkaiser: ok, should work then. When i tried to play aroud with the dvfs_table i got some errors in kernel log that the freqency could not be set
<longsleep>
currently comparing Pine64 and Odroid-C2 - interesting stuff :)
<tkaiser>
longsleep: I have to repeat anyway since this Linpack thing leads to throttling at 1200MHz
<tkaiser>
longsleep: Do you use real benchmarks or Phoronix stuff?
<longsleep>
tkaiser: not comparing performance, i compare usability
<longsleep>
eg glxgears, pine64 is at 57fps odroidc 102fps in the default settings both Ubuntu mate
<longsleep>
both is not really nice to use if you ask me
<longsleep>
interesting is that firefox works on odroidc and just segfaults on pine, must be the kernel then
<longsleep>
and the most important thing, virtualization is still broken on C2, works fine on Pine64
<longsleep>
pine64 without heatsink is just below 80°C, sometimes triggering the trip point when running glxgears or scrolling around on the desktop, might be a good idea to use 85°C as the thirs trottling point and increase all other trip points by 5 as well
<KotCzarny>
80C? wow
cosm has quit [Ping timeout: 268 seconds]
cosm has joined #linux-sunxi
<longsleep>
yeah, just idling in the desktop is about 60C (all without heatsink)
<KotCzarny>
madness
<KotCzarny>
and it's winter
<AA781>
interesting indeed
vagrantc has joined #linux-sunxi
<longsleep>
yes, this is with 23C ambient temperature, no enclosure
<KotCzarny>
now imagine what will happen when ambient temp. reach 28-32C
<KotCzarny>
or more
<longsleep>
I did not do much testing with Odroid-C2 but can already tell that the fing does not get much warmer than 65C
<longsleep>
s/fing/thing :)
<longsleep>
but it has a rather large heatsink
<AA781>
so these things run little hotter than expected (pine)
<longsleep>
depends what you expect
<KotCzarny>
h3 wasnt ice cream neither
<longsleep>
but yes, A64 gets hot
<AA781>
well the max they say is 105 C if I remember correctly right?
<longsleep>
125
<longsleep>
absolute max
<longsleep>
no idea what pine folks said
<KotCzarny>
but silicon gets dirty in high temps
<longsleep>
yes
<AA781>
need to find out from a friend regarding the pi's a53 implmentation
<KotCzarny>
so it will definitely affect chip life
<AA781>
how hot it runs
<longsleep>
right now our settings pretty much prevent the Pine64 getting hotter than 95C
<longsleep>
but i am willing to bump that to 100 if it proves stable
<AA781>
I have been keeping up with the progress on the forums
<AA781>
wonder how the thermal is behaving with droid
<longsleep>
but essentially everyone who wanst do do something serious with it should get a heatsink
<KotCzarny>
sink it in oil tank
<longsleep>
Android throttles much earlier and even loose cores if you manage to get it doing multithreaded stuff at all
<longsleep>
its not getting that hot if 3 cores are down :)
<AA781>
my concern is if it will run stabel enough without losing power
<longsleep>
yes, if you power it through euler
<AA781>
this will be my first venture into the arm world
<tkaiser>
The ODROID folks said already that they also experience throttling without the heatsink with CPU utilisation alone. And there a much stronger GPU is used
<longsleep>
AA781: depends on your PMU and your USB cable other wise
<longsleep>
tkaiser: yesh, the heatsink gets quite warm and does a great job
<longsleep>
without it would be insane at their speeds
<AA781>
was thinking of utilizing a usb charging hub
<AA781>
2.4 amps per port (8 port)
<KotCzarny>
longsleep, i bet getting off some of that resin on the chip would make heatsink much more efficient
<tkaiser>
longsleep: They had an advantage. The C1 and C1+ before. C1 with optional heatsink caused complaints.
<AA781>
should provide enough power
<longsleep>
AA781: get a short cable, like 15cm and pray
<AA781>
tat seems to be the operative word for the pine lol
<tkaiser>
AA781: The tiny contacts are also a problem (1.8A max by specs)
<AA781>
pray
<longsleep>
tkaiser: well i used the C1 without heatsink first and had problems too :)
<AA781>
wish april was here and I cna give more than hand
<AA781>
but at this point thought I would get caught up with things on this end of kernel development so I can jump right in
<longsleep>
tkaiser: i still would prefer that they would not ship with the premounted heatsink thoug, removing it can damage the chip
<AA781>
when they arrive
<AA781>
the s805s?
<longsleep>
yes
<AA781>
ya they certainly gum up those things
<AA781>
you should see the mxq tv boxes
<AA781>
lol
<AA781>
not any better
<longsleep>
we will see how things go when most kickstarter backers have received their pine
<AA781>
but thermal I noticed on the s805 (the mxq android boxes) seems to hover at around 70C on idle
<AA781>
wonder what it would be like without
<longsleep>
but as far as i am concerned Pine64 is pretty solid now and can be used for repeatable clean arm compiling clusters just fine
<longsleep>
compiling on pine64 is so much faster than in qemu :)
<AA781>
the emulation vs baremetal differnce
avph has quit [Ping timeout: 246 seconds]
<longsleep>
yes, i use a 20 node ODROID-C1 cluster at the moment, but that one can only do distcc and has no virtualization. That will imrove a lot with Pine64 and also can do arm64 jobs
<AA781>
I am trying my hand at hosting cluster
<AA781>
with the pine
<longsleep>
arg, i am missing so many keys today - i probably need a break
<AA781>
take one mate
* longsleep
is afk for a while
<AA781>
from everythign I have seen you all deserve one
doppo has quit [Ping timeout: 248 seconds]
avph has joined #linux-sunxi
doppo has joined #linux-sunxi
<AA781>
back to work for me
AA781 is now known as AA78-working
leio has joined #linux-sunxi
Nacho has quit [Remote host closed the connection]
Nacho has joined #linux-sunxi
<GeneralStupid>
tkaiser: i removed irqbalance
<GeneralStupid>
are there any upgrades for osd (g2d?) for h3 ?
JohnDoe_71Rus has joined #linux-sunxi
<rellla>
GeneralStupid: is done within software via pixman. G2D is not available on H3
<GeneralStupid>
rellla: do i have to enable this?
<rellla>
export VDPAU_OSD=1 and have libpixman installed.
premoboss has quit [Ping timeout: 250 seconds]
premoboss has joined #linux-sunxi
avph has quit [Ping timeout: 246 seconds]
avph has joined #linux-sunxi
<vagrantc>
wens: i got usb working on the cubieboard4 by applying a the .dts patch from your cubieboard4 branch ... is there any reason not to push that upstream?
<vagrantc>
wens: worked for me with mainline 4.5-rc7