<lkcl>
if anyone else would like to take over answering questions, please feel free ok?
<lkcl>
i've only ever once run into the "whoaaaa, cowboy, you've made too many comments" problem on slashdot - i don't think it's going to happen again, here, but it's just... 5am or so. time to sleep :)
<lkcl>
/peaceout
<CIA-14>
rhombus-tech: Brian master * r2c5ff96433e0 /allwinner_a10/orders/CapnJackD.mdwn:
hp_ has joined #arm-netbook
n6pfk has quit [Remote host closed the connection]
mikey_w has quit [Remote host closed the connection]
gsilvis has quit [Read error: Connection reset by peer]
gsilvis has joined #arm-netbook
gimli has quit [Quit: gimli]
oliv3r has joined #arm-netbook
<oliv3r>
Hi, I just read the page linked from the slashdot article, and notice that no AMLogic chips would be concidered. I wonder why that is? I know that pivos has published lots of sources for the AMLogic in github and that amlogic actually has some code out in the wild for their VPU, whereas Allwinner aren't putting anything out for their VPU.
Simon_G has joined #arm-netbook
Simon_G has left #arm-netbook [#arm-netbook]
vgrade has joined #arm-netbook
kamiquasi has joined #arm-netbook
<lkcl>
oliv3r: look at that source code more closely. you'll find it's just the standard kernel "shim" that interfaces some shared-memory to the userspace proprietary MALI400 opengl libraries.
<lkcl>
oliv3r: AMLogic cost us a lot of money and a client, thanks to their GPL-violating NDA.
<lkcl>
we gave them a chance to make amends... and they haven't taken it.
<lkcl>
we're not interested in fighting battles; we're interested in making money. as long as there's at least one SoC vendor willing to work with us and supply product, why should we fight with any others?
<lkcl>
this is a lesson that wookie taught me.
<Triffid_Hunter>
lkcl: is there one?
<lkcl>
so anyway, you're incorrect: *neither* allwinner *nor* amlogic are providing the source code of the userspace MALI400MP opengl library, because they are *not permitted* to, by ARM. it's proprietary. to ARM.
<lkcl>
Triffid_Hunter: sorry (meep!) - one what? lots of ambiguity here :)
<Triffid_Hunter>
lkcl: one SoC vendor who wants an open toolchain enough to actually do something about it
<lkcl>
Triffid_Hunter: yes - allwinner.
<lkcl>
ok: they all release the toolchains
<lkcl>
.... for the most part :)
<Triffid_Hunter>
I've been reading that even the binary blobs for the mali are fairly useless
<lkcl>
what they all *don't* do is require that you sign a GPL-violating NDA that entirely jeapordises the products built around them even before they're designed!
<lkcl>
Triffid_Hunter: yeah... that's a global problem not just isolated to one CPU, but to all SoCs that use MALI.
<lkcl>
yeah. i think libv was involved in nouveau at some point. or... maybe... yeah i can't recall.
<lkcl>
best to ask him or hang around on the #libv freenode channel
* Gumboot
shifts about anxiously
jlj has joined #arm-netbook
* lkcl
removes the pin on Gumboot's seat
madmalkav has joined #arm-netbook
penguin42 has joined #arm-netbook
naobsd has joined #arm-netbook
pawel5870 has quit [Quit: Leaving.]
madmalkav has quit [Quit: Leaving]
jlj has quit [Ping timeout: 252 seconds]
Hexxeh has joined #arm-netbook
<hno>
oliv3r, i would not say Allwinner haven't put anything out for their VPU. They have published everything they have and can publish, including fairly up to date mali binary libs.
<hno>
they have no other clients using X11 and something with the ARM MALI X11 driver is odd so X11 3D accel do not, but OpenGL ES in itself do work,
<hno>
Ah right, you said VPU. They have binary libs for the VPU as well, just that no one knows if they work under Linux or how to properly use them..
Quarx has quit [Ping timeout: 244 seconds]
Quarx has joined #arm-netbook
tuliom has joined #arm-netbook
tuliom has quit [Client Quit]
tuliom has joined #arm-netbook
mikey_w has joined #arm-netbook
n6pfk has joined #arm-netbook
chz_ has joined #arm-netbook
mSquare has left #arm-netbook [#arm-netbook]
chz_ has quit [Quit: Leaving]
bsdfox has joined #arm-netbook
bsdfox has quit [Changing host]
bsdfox has joined #arm-netbook
gimli has joined #arm-netbook
popolon has joined #arm-netbook
gimli has quit [Quit: gimli]
<CIA-14>
rhombus-tech: Mathijs master * r3440d2665014 /allwinner_a10.mdwn: small formatting fix
ssalevan has joined #arm-netbook
<penguin42>
I assume the Cubieboard is Tom ?
<lkcl>
penguin42: i don't know! :)
<lkcl>
it would be hilarious if it was
<penguin42>
lkcl: It seems unlikely to be a coincidence
<penguin42>
lkcl: Looks like a nice little board
popolon has quit [Quit: Quitte]
ssalevan has quit [Quit: Leaving.]
<furan>
hey gimli, so is a10 xbmc dead?
<furan>
oops
<furan>
scrolled up
hp_ has joined #arm-netbook
<L84Supper>
any word yet on the AC timing specs for the A10 GPIO's and IRQ's? Has anyone done any testing?
tuliom has quit [Read error: Connection reset by peer]
<hno>
L84Supper, no, and my DSO Nano probe cable have gone missing..
hp_|2 is now known as hp_
MMlosh has quit [Ping timeout: 260 seconds]
pawel5870 has quit [Quit: Leaving.]
MMlosh has joined #arm-netbook
popolon has joined #arm-netbook
jlj has quit [Ping timeout: 268 seconds]
jlj has joined #arm-netbook
eebrah has joined #arm-netbook
<libv>
lkcl: i was not involved with nouveau, btw
<libv>
i was _very_ heavily involved with kicking ATI ass though, and getting a free driver and free docs out, with a lot of help from AMD beating on them too
<libv>
(and two other suse guys of course :))
xxiao has left #arm-netbook [#arm-netbook]
<penguin42>
libv: In that case thank you
* penguin42
pats his HD4350 running the open driver
<libv>
well, some guys at redhat joined up with ATI
ssalevan has joined #arm-netbook
<libv>
and because ATI wanted us at SuSE dead
<libv>
and the redhat guys were miffed because they were not involved in one of the biggest free software stories on graphics...
xxiao has joined #arm-netbook
<libv>
we now have an AtomBIOS based mess, with a driver that is mostly perl code, that works just a bit, enough to get the flack off of AMDs back, but definitely not enough to encroach on fglrx
<penguin42>
libv: Yeh I'm OK because my card is fanless; the power management has big problems for people - as a desktop card though it works fine
<penguin42>
(possibly not if you're a serious gamer)
<libv>
and the SuSE^WEvil Novell employees were deemed microsoft conspirators anyway and we were trying to subvert free software with microsoft marketing by freeing ATI and writing actual C code :p
<libv>
(this statement is quite absurd, but it is, honestly, only a tiny step above of the shit we really got thrown at us)
<lkcl>
libv: allo. i thought i remembered you saying you weren't involved with nouveau - couldn't remember for sure
<lkcl>
yaay on the ATI stuff :)
<penguin42>
it'll be interesting what ARM stuff ATI comes up with; they've been making noises I think but no detail
<lkcl>
libv: bizarre
<libv>
lkcl: we both know how unbelievably ugly free software can be ;)
<lkcl>
penguin42: yeah. i'm just amazed that AMD/ATI aren't "getting it" and creating their own ARM SoC. i mean... they've got their own 3D GPU core for god's sake, they should fucking well leverage it
<libv>
they sold their low power core
<lkcl>
libv: i'd say "tell me about it", but... really - please don't
<libv>
it's now called the adreno
<lkcl>
libv: what? argh....
<lkcl>
idiots
<libv>
nothing new for amd
<libv>
geode was dropped completely a year or so before the atom happened
<lkcl>
oh, was that the stuff that went into the HTC Universal? there was a low-power GPU...
<libv>
adreno
<penguin42>
lkcl: Maybe it's because it costs money and they're not exactly cash rich at the moment; it's also a very competivie world in the tablet/mobile world already
<libv>
in the qualcomm snapdragons
<lkcl>
geode was National Semi originally.
<lkcl>
libv: ahhh
<lkcl>
ohh yeah, i recognise the name, now.
<libv>
amd desperately needed cash in those days
<libv>
they also sold their fabs
<libv>
i believe that by early 2008, the entire worth of AMD was below of what they had paid for ATI
<lkcl>
all right, i gotta go. conference call for the next 3 hours
<lkcl>
sorry i just noticed it's 8pm...
<lkcl>
/peace
popolon_ has joined #arm-netbook
popolon has quit [Ping timeout: 246 seconds]
popolon_ has quit [Client Quit]
popolon has joined #arm-netbook
merbzt has joined #arm-netbook
<libv>
lkcl: instruction set is the last of the worries, arm and others have evolved in low power situations, whereas intel and amd were in a race for pure performance, over many many years. i doubt that the size of instruction translation hw in intel and amd compared to the instruction translators under arm are what amount to the massive performance/watt difference
hp_ has quit [Ping timeout: 246 seconds]
<penguin42>
libv: Indeed in Intel's knights corner presentation recently they highlighted the x86 specific logic was only 2%
<penguin42>
libv: And there's something to be said for CISC keeping your icache utilisation high
<oliv3r>
lkcl: sorr for the late response, had to leave for the day; but that sucks that they have a GPL violating NDA. I don't know the current license of their code, it probably is 'just as bad' but its on githup, meaning atleast reverse engineering it is 'a' possibility. :)
<oliv3r>
But how is their VPU tech just a shim to the mali400 gpu? I was under the strong impression that they had an actual VPU. I know the A10 claims to have one? not to mention that ARM has their own VPU IP, seperate from the mali400. I didn't know mali, in fact, could encode/decode video.
<specing>
libv: I have a Geode, works good as a router
<oliv3r>
lkcl; as for the GPU bit, yes, the gpu is mali400 stuff, no question about it; they all use it. Lima is trying to fix that. And for fighting battles; I thought it was somewhat the mission goal of rhombus, get GPL software supported hardware. A10 is working with you, and that's great, That they can't help with the mali400 I get, but what about their 'home developped' VPU?
<oliv3r>
I have 5 'wise' thinclients, based on geode's
<oliv3r>
doing nothing. If someone working on coreboot/video drivers on those suckers wants a free unit, i'm happy to ship one. I even have pimped their ram and added a 1gb DoC pata thing
<specing>
I bought it specifically for a router
Quarx has quit []
destinal is now known as destinal-away
<specing>
Im probably going put some ldap onto it and probably nagios and some traffic graphers
<rm>
I need to order an USB gigabit adapter for my MK802 router :)
<specing>
uhm
<rm>
well, it's only a secondary one, handling IPv6 and VPN
<rm>
my primary router is a TL-WR703N
<RaYmAn>
you must have very slow internet =P
<rm>
the 703n actually handles close to 100 mbit in full duplex
<RaYmAn>
doing NAT?
<rm>
yes
<rm>
and PPPoE
<RaYmAn>
impressive
<oliv3r>
I got fiber running at sync. 100mbit; got a cheap N36L lifting that
<RaYmAn>
specing: It's fiber to the condo building, then spread out over ethernet to individual condos - and there's only one usable plug in each condo
<rm>
sooo, that's what routers are for, right
* RaYmAn
use a core2duo mac mini (running linux) as a router
<Turl>
FTTB :P
<RaYmAn>
yeh - we're changing provider soon-ish though. They'll provide gigabit fiber to each condo - probably can't get more than 90mbit then though, but at least it'll be prepared for higher (and IP-TV won't affect download speeds :P)
<oliv3r>
they actually run dual fiber here, 1 is for internet (and internet related services, like voip and IPTV). the second line is purely for analog and dvb-c television
<RaYmAn>
interesting
<oliv3r>
they have some really small optical -> analog/dvb-c converter thingie
<RaYmAn>
I guess their boxes aren't capable of gigabit
<oliv3r>
what makes it nice is, they are 'ready' for dual optical i guess
<oliv3r>
no the modems and infrastructure in this part of town does not handle gigabit
<oliv3r>
but it will, soon-ish
<oliv3r>
the new stuff in our city is now being rolled out, capable of gbit
<oliv3r>
and our fiber is 'open'
<oliv3r>
meaning several ISP's can offer you their service
<RaYmAn>
that's cool
<oliv3r>
there is one ISP (not available to me yet) that offers gbit for 99E a montttth on said line
<RaYmAn>
wow
<penguin42>
oliv3r: Is that GPON optical ?
<oliv3r>
GPON? the two fibers entering our home is optical yes as is the modem's input
<rm>
GPON is this super-tree topology where there's over 9000 people on the same shared 3-Gbit-something line
<rm>
with fully passive splitters along the way
<oliv3r>
oh, no idea how it all technically works beyond our home. I do know there's opticla switches at the other end, and that's it :p
<specing>
I see what you did there rm
<rm>
and they promote this as FIBER OPTICS TO THE HOME and GIGABIT
<specing>
"over 9000"
<oliv3r>
ah, nah; this is Eindhoven here, you know, smartest region in the world :D
<oliv3r>
they'd loose a lot of face i guess
<RaYmAn>
I suspect all ISP's over-subscribe to some level ;)