tllim has quit [Read error: Connection reset by peer]
tkaiser has quit [Ping timeout: 246 seconds]
_whitelogger has joined #linux-sunxi
OnkelUlla has joined #linux-sunxi
Asara has quit [Ping timeout: 244 seconds]
Asara has joined #linux-sunxi
lurchi_ has joined #linux-sunxi
lurchi__ has quit [Ping timeout: 246 seconds]
OutBackDingo has joined #linux-sunxi
night199uk has quit [Ping timeout: 245 seconds]
dddddd has quit [Remote host closed the connection]
Mangy_Dog has joined #linux-sunxi
selfbg has joined #linux-sunxi
night199uk has joined #linux-sunxi
f0xx has joined #linux-sunxi
Mangy_Dog has quit [Ping timeout: 246 seconds]
wwilly_ has quit [Quit: Leaving]
wwilly has joined #linux-sunxi
embed-3d_ has joined #linux-sunxi
embed-3d has quit [Ping timeout: 276 seconds]
reinforce has joined #linux-sunxi
lvrp16 has quit [Write error: Connection reset by peer]
arnd has quit [Read error: Connection reset by peer]
scroll2_ has quit [Write error: Connection reset by peer]
arnd has joined #linux-sunxi
lvrp16 has joined #linux-sunxi
steev has quit [Read error: Connection reset by peer]
Tartarus has quit [Read error: Connection reset by peer]
scroll2_ has joined #linux-sunxi
steev has joined #linux-sunxi
Tartarus has joined #linux-sunxi
chewitt has joined #linux-sunxi
cnxsoft1 has joined #linux-sunxi
cnxsoft has quit [Ping timeout: 246 seconds]
yann has quit [Ping timeout: 255 seconds]
<wens>
montjoie: only possible for someone who knows the language and is alreay in China :p
<KotCzarny>
experience not required?
<wens>
their attempts at hiring foreigners didn't go so well AFAIK
AneoX_ has joined #linux-sunxi
AneoX has quit [Ping timeout: 246 seconds]
ldevulder has joined #linux-sunxi
IgorPec has quit [Quit: Leaving]
jbizcocho has quit [Ping timeout: 246 seconds]
tllim has joined #linux-sunxi
maz has joined #linux-sunxi
jbizcocho has joined #linux-sunxi
<libv>
there's leviathanch, who did like an intership there
<libv>
he was all lovey-dovey before he actually got there
<libv>
and came back pretty bitter and disappointed
<libv>
but allwinner pretty much had their own alexandru voica (who was at img talking big nonsense as well)
<libv>
iirc, that guy ran off to NextThingCo
<libv>
and we all know how that one went
<libv>
"Ben El-Baz"
<lvrp16>
what's the alexandru voica story?
<lvrp16>
pretty much all the good engineers left allwinner anyway during the downsizing
<lvrp16>
it's a skeleton crew that barely knows their own stuff
<libv>
have they ever known their stuff?
<libv>
what downsizing, when?
<KotCzarny>
a20 wasnt bad
<lvrp16>
libv: my remark is relative XD
<libv>
lvrp16: :)
<libv>
KotCzarny: well, they threw together things that were bought and popular, and cheap
<KotCzarny>
still, isnt it the story of most socs?
<libv>
the fact that it was cheap was #1, then it was badly managed so we got GPL adherence (for a bit), and we got kernel and uboot code
<lvrp16>
they had a downsizing where their entire staff was cut in half two years ago
<libv>
and then linux-sunxi and the cubieboard happened because of those two factors
<KotCzarny>
and it was good
<libv>
lvrp16: tied to the tablet market collapsing and the fact that people have had their fill of cheap android based tv boxes?
<KotCzarny>
and still is, given recent finding about sata fixage
<libv>
KotCzarny: there are way better chips out there
<libv>
but those socs have no community
<libv>
and limited enduring upstream support because of it
<KotCzarny>
slap 4 a53 cores into a20 and see it fly
<KotCzarny>
pity it wont happen though
<libv>
orange pi and lemaker happily push out amlogic, rockchip and realtek boards
<libv>
and actually, fosdem would love to use one particular orange pi rockchip board, or one particular lemaker realtek board
<libv>
as the hw side of things looks ideal (hdmi input, for instance)
<wens>
is realtek support happening?
<libv>
but there's just no way that we are going to spend the endless manyears on getting those platforms useful and stable
AneoX has joined #linux-sunxi
<libv>
so a20 is our only option
<lvrp16>
lemaker? you mean banana pi?
<libv>
i am always confused about whic company actually owns the bpi
<libv>
yes, banana pi
<lvrp16>
lemaker is defunct after amd screwed them
<libv>
oh, i see
AneoX_ has quit [Ping timeout: 258 seconds]
<lvrp16>
sinovoip is the banana pi guys, they haven't a clue
ganbold has joined #linux-sunxi
<libv>
aha, so after allwinner and nextthingco, ben el baz now moved over to the blockchain hype
<libv>
my toolchain friends at suse also played with the arm based opteron for a bit
<libv>
i think we should be glad that amd chose to focus and now is executing zen well
tllim has quit [Read error: Connection reset by peer]
Mangy_Dog has joined #linux-sunxi
<wens>
haven't had a reason to swap out my PC yet
jbrown has quit [Ping timeout: 258 seconds]
<lvrp16>
too bad they cancelled threadripper 3xxx
<lvrp16>
would have been a beast compile machine
<libv>
maybe threadripper is no longer a viable product? with rome based desktop cpus having enough cores, and rom based epic cpus being affordable enough, to not require a product in between?
<libv>
also, how certain is it that threadripper will not exist going forward?
victhor has joined #linux-sunxi
<wens>
aren't they holding a press event real soon? hot chips? computex?
<lvrp16>
amd probably got massive amounts of orders for it's 7nm server chips that they didn't have any chips left to make hedt
<libv>
i actually would be happy with a 16core 32thread cpu that is not threadripper, as then i could fit it in a mini-itx case that i could throw in a backpack and drag over to customers
<libv>
the fact that there were no real mini-itx boards for threadripper, stopped me from getting threadripper a year or so ago, and to instead spend a lot less and go for nvme only
<wens>
you can't get a lot of RAM with mini-itx though
<libv>
now the 4core 8 thread skylake system still ran rings around my customers machines last year (mainly as they all insist on running windows with ubuntu vms)
<wens>
lol
<libv>
ram was never the bottleneck when bitbaking
wwilly_ has joined #linux-sunxi
tllim has joined #linux-sunxi
<libv>
single thread performance is, as there are bound to be a few places where the whole thing hits a focal point
wwilly has quit [Ping timeout: 268 seconds]
<wens>
depends on what you're baking I suppose
f0xx has quit [Ping timeout: 244 seconds]
<libv>
well, my experience is that adding more cores to the problem is also not the solution, but it is the only cheap solution
<libv>
in that fixing the focal point is not always workable
<libv>
yocto/bitbake benefits hugely from more cores, but not fully linearly
<wens>
linking is still single core
<libv>
i am not too unhappy that the arm based opteron did not come to fruition, as we now indeed have more cores than we could dream of in 2016
<libv>
arm servers never were going to deliver enough of a power usage advantage to warrant the disruption
<lvrp16>
socionext's 24 core a53 comes to mind...what a bizarre chip
<libv>
and unless numbers are high enough, they would never be cost effective
<lvrp16>
why would someone build such a thing
<lvrp16>
what market could possibly use it
<libv>
i was doing lima while being paid by nokia to mostly "be available" and fix bugs
<libv>
codethink was making quite some noise about lima
<libv>
and to some extent supported it
<libv>
but i was being sent off to paid customer projects
<libv>
an unpaid internal project was baserock
<libv>
an over-the-air update system, which morphed into a full blown from-source embedded distribution
<libv>
the megalomania went as far as one of the guys involved trying to tie up a bunch of 4 core armadas with... a15s? to form a arm server
<libv>
i am not sure if they ever got hw to boot and run on more than one module
<libv>
i never dared ask since
f0xx has joined #linux-sunxi
<libv>
this baserock thing, and their server had most of the company working on it
<libv>
it never really flew
<libv>
the original reason to go there was because they got their hands on an early calxeda machine
<libv>
and they could not get the info to get things like uboot running on it
<wens>
aalm: true, but other countries aren't nearly as welcoming as they used to be
<lvrp16>
huawei has secretly won the server chip race. just it's not advertised in the western media. if intel cuts of xeon tomorrow and amd cuts off rome, huawei will win all of China's server business
<aalm>
i don't think we're going to see the kind of protectionism crap usa came up with here in eu
<lvrp16>
they have been building all kind of random shit for the last two years. ssd controllers, network controllers, fpgas lol
<wens>
wonder which vendor Cloudflare uses
<lvrp16>
it's like the only Chinese business that has original designs lol
<wens>
are there any other ARM server vendors left? Qualcomm cut theirs
<lvrp16>
huawei vs intel vs amd pretty much, the other designs have significant flaws
<DonkeyHotei>
marvell?
<libv>
wens: linus is not wrong
<lvrp16>
idk, i don't agree with linus on his point
<lvrp16>
most cloud workloads are not written in C anymore
<wens>
I'm not seeing ARM servers win on pricing though
<lvrp16>
thunderx2 is pretty weak compared to huawei, intel, and amd's offerings
<libv>
wens: as that's directly tied to volume
<libv>
and you won't get volume until you have "ecosystems", as linus describes
<lvrp16>
marvell just doesn't have the resources or the team to do high performance large core designs
<libv>
but you won't get to volume unless the proposition is a lot more interesting than the existing setup
<wens>
libv: unless the volume discount is really steep, I'm not seeing it. The listing prices for "comparable performance" boxes seem to be comparable as well lol
<libv>
either on price, performance per watt, and easy perfect migration
<libv>
the focus on servers singularily is pretty wrong
<wens>
PPW seems more likely, but the powerful cores aren't exactly power savers either
<libv>
a direct conclusion from linus' email is: provide atx style motherboards for people to work off
<libv>
and bring up discrete graphics boards with drivers (and thanks to atis strategy, bioses) to work on said arm64 machines
<lvrp16>
linus doesn't build large web systems, he is slightly out of touch with large system designs
<fALSO>
Linux orangepipc 5.2.0-rc1 #31 SMP Mon May 20 10:25:05 WEST 2019 armv7l GNU/Linux
<fALSO>
:)
<lvrp16>
R/SAS/Python/SQL
<wens>
I have to agree. large web systems don't even care , or can't care , about cache locality and low level stuff anymore
<lvrp16>
HDFS block size is 128MB lol
<wens>
lvrp16: that sounds more like big data than web systems
<lvrp16>
web analytics*
<lvrp16>
basically to "read" all the crappy features they build into phone apps
<lvrp16>
and determine which is better
<wens>
lol
<lvrp16>
why you have to update the god damn app on your phone on what seems like a daily basis
<lvrp16>
they push new features and test those new features and generate boatloads of useless data
<lvrp16>
then run analytics to see if square or round corner is better in the settings section
<wens>
A/B testing at Facebook lol
<wens>
sounds like a soul-depriving job :(
<lvrp16>
any company > 1B
<KotCzarny>
there's no soul in the machine, only algoirthms
<lvrp16>
my friends do this and they say that they have incredible talent on the team. masters and phd kids that are really smart...building android UI changes 365 days a year
<wens>
lvrp16: when you say it like that it sounds more like a waste of talent
<aalm>
what it is
<lvrp16>
wens: imho, they could be curing cancer
<KotCzarny>
there's no money in curing cancer
<wens>
lvrp16: making self-driving cars is closer wrt discipline
<lvrp16>
not when FLAG is paying them 250K a year with 200K signing bonus and 250K of stock over 4 years
DonkeyHotei has quit [Ping timeout: 255 seconds]
<wens>
only in the US
<lvrp16>
fresh grads only get a smaller salary of $150K
<lvrp16>
and stock could be worth $1M by the time it vests
<karlp>
maybe we're not looking the right way. sure, there's lots of money in "treating cancer" but we're also losing a massive time suck of people who die, people who spend countless hours in "treatment" and then all of the people working around it. curing it could/would actualyl be profitable, but not for the current financial year...
<lvrp16>
once you have the cure, it's not worth any money lol, the money is in the treatment
<KotCzarny>
karlp, it doesnt work that way in capitalism
<wens>
karlp: it is likely a plus for society, but not for the people footing the bill
<lvrp16>
treatment costs 50K-100K a year...cure costs $10K one time
<lvrp16>
what interests does a drug company have in finding a cure
<KotCzarny>
which means, the guys behind 'treating cancer' are the societal cancer
<lvrp16>
when they can make bank peddling the same shit
<lvrp16>
0 dollars invested, many dollars coming in...vs many dollars invested, few dollars coming in
<aalm>
here, take a opioid
<lvrp16>
fentanyl from China?
<karlp>
KotCzarny: I know :)
<karlp>
lvrp16: I never said drug company would do it :)
yann has joined #linux-sunxi
chewitt has quit [Quit: Adios!]
tl_lim has joined #linux-sunxi
ldevulder_ has joined #linux-sunxi
ldevulder has quit [Ping timeout: 244 seconds]
ldevulder_ is now known as ldevulder
victhor has quit [Ping timeout: 258 seconds]
Net147 has quit [Read error: Connection reset by peer]
Net147 has joined #linux-sunxi
lurchi_ is now known as lurchi__
tllim has quit [Quit: Leaving]
kaspter has quit [Quit: kaspter]
Net147 has quit [Quit: Quit]
Net147 has joined #linux-sunxi
dddddd has joined #linux-sunxi
wwilly_ has quit [Quit: Leaving]
imcsk8 has quit [Ping timeout: 244 seconds]
tl_lim has quit [Read error: Connection reset by peer]
airwind has quit [Quit: airwind]
msimpson has joined #linux-sunxi
msimpson has quit [Read error: Connection reset by peer]
tllim has joined #linux-sunxi
msimpson has joined #linux-sunxi
tl_lim has joined #linux-sunxi
kaspter has joined #linux-sunxi
tnovotny has joined #linux-sunxi
selfbg has quit [Remote host closed the connection]
tl_lim has quit [Quit: Leaving]
kaspter has quit [Ping timeout: 248 seconds]
f0xx has quit [Ping timeout: 246 seconds]
afaerber has quit [Quit: Leaving]
jbrown has joined #linux-sunxi
jstefanop has joined #linux-sunxi
vagrantc has joined #linux-sunxi
jstefanop has quit [Remote host closed the connection]
jstefanop has joined #linux-sunxi
dlan has quit [Remote host closed the connection]
jstefanop has quit [Ping timeout: 252 seconds]
leviathanch has joined #linux-sunxi
leviathanch has quit [Client Quit]
leviathanch has joined #linux-sunxi
leviathanch has quit [Client Quit]
msimpson has quit [Remote host closed the connection]
msimpson has joined #linux-sunxi
return0e_ has quit [Ping timeout: 268 seconds]
return0e has joined #linux-sunxi
msimpson has quit [Read error: Connection reset by peer]
random_yanek has quit [Ping timeout: 258 seconds]
reinforce has quit [Quit: Leaving.]
victhor has joined #linux-sunxi
yangxuan has joined #linux-sunxi
random_yanek has joined #linux-sunxi
cnxsoft1 has quit [Quit: cnxsoft1]
random_yanek has quit [Ping timeout: 246 seconds]
f0xx has joined #linux-sunxi
yangxuan has quit [Read error: Connection reset by peer]
random_yanek has joined #linux-sunxi
yann has quit [Ping timeout: 245 seconds]
msimpson has joined #linux-sunxi
jbrown has quit [Ping timeout: 258 seconds]
megi has joined #linux-sunxi
<vagrantc>
hrm. was looking forward to testing ethernet on cubieboard4 with 5.1.x ... but i don't get any output on the serial console
<vagrantc>
linux 4.19.x works fine ... haven't tried 5.0.x yet.
jbrown has joined #linux-sunxi
VargaD has joined #linux-sunxi
<buZz>
vagrantc: made sure to enable serial console in kernel parameters aswell?
<buZz>
(and in dtb i guess)
reinforce has joined #linux-sunxi
<vagrantc>
buZz: there some new configuration between 4.19.x and 5.1.x ?
<buZz>
plausible, yes
<vagrantc>
console=ttyS0,115200
<vagrantc>
also tried without so it would fallback to chosen property
<vagrantc>
and also tried kernel builds from two different OSes
<vagrantc>
i don't need something other than CONFIG_SERIAL_8250_CONSOLE=y enabled, no?
<vagrantc>
chosen {
<vagrantc>
stdout-path = "serial0:115200n8";
<vagrantc>
};
<buZz>
8250 is the x86 uart, are you sure thats the right driver?