| 00:03.34 | mnemoc | ash should have a flag to prefer builtins when available |
| 00:11.52 | hno | brom extracted. |
| 00:37.08 | *** join/#arm-netbook mikey_w (~mike@c-71-63-115-202.hsd1.va.comcast.net) |
| 00:37.15 | *** join/#arm-netbook mike__ (~mike@c-71-63-115-202.hsd1.va.comcast.net) |
| 01:07.01 | CIA-91 | rhombus-tech: alpha master * r3be3748e209e /allwinner_a10/orders/alpha.mdwn: |
| 03:20.32 | destinal_ | are there any good guides to porting android to a new platform? |
| 08:39.35 | *** join/#arm-netbook gordan (~gordan@78-86-195-86.zone2.bethere.co.uk) |
| 11:26.58 | *** join/#arm-netbook penguin42 (~dg@tu006.demon.co.uk) |
| 12:23.01 | *** join/#arm-netbook QingPei (~qingpei@123.118.106.251) |
| 12:24.19 | CIA-91 | rhombus-tech: Luca master * rbc94451ff44f /allwinner_a10/orders/galo.mdwn: |
| 13:03.17 | CIA-91 | rhombus-tech: Omar master * r0fc5cead019b /allwinner_a10/orders/omar.mdwn: |
| 13:05.11 | CIA-91 | rhombus-tech: Omar master * r6b09b4cc2489 /allwinner_a10/orders/omar.mdwn: |
| 14:25.23 | *** join/#arm-netbook gordan (~gordan@78-86-195-86.zone2.bethere.co.uk) |
| 15:15.18 | *** join/#arm-netbook QingPei (~qingpei@123.118.106.251) |
| 15:20.54 | *** part/#arm-netbook QingPei (~qingpei@123.118.106.251) |
| 16:22.29 | CIA-91 | rhombus-tech: Christian master * r5956e9de5113 /allwinner_a10/orders/drscalet.mdwn: |
| 16:43.13 | penguin42 | wonders how much that androidtabletbay charge in shipping |
| 16:44.07 | CIA-91 | rhombus-tech: lkcl master * red18b7d2066e /allwinner_a10/orders/not_valid/galo.mdwn: no contact details, had to move to not_valid. will restore once noticed |
| 16:44.18 | CIA-91 | rhombus-tech: lkcl master * r4005cdf8f104 /allwinner_a10/orders/omar.mdwn: order processing cleanup |
| 16:44.18 | CIA-91 | rhombus-tech: lkcl master * r287afbe6c39d /allwinner_a10/orders/ (galo.mdwn not_valid/galo.mdwn): no contact details, had to move to not_valid. will restore once noticed |
| 17:43.24 | penguin42 | lkcl: I'm surprised at the number of preorders wanting so many cards |
| 18:07.11 | lkcl | penguin42: i know. alpha-grade, at $30 it's 391 cards. |
| 18:07.35 | lkcl | at $40 that drops to 255. |
| 18:07.44 | penguin42 | lkcl: Lots of people wanting like 5-10 cards as well |
| 18:08.27 | lkcl | beta-grade it's large numbers because bari needs 1,000 |
| 18:09.08 | lkcl | i think this is going to go nuts once we have actual boards. |
| 18:09.55 | penguin42 | yeh you probably need to use someone else to do the actual selling/shipping etc |
| 18:10.12 | lkcl | yep. my friend adam. |
| 18:10.35 | penguin42 | haha ok, I was going to suggest someone like Amazon |
| 18:10.39 | zenitraM | some with seriously non-ddoseable servers |
| 18:10.45 | zenitraM | s/some/someone |
| 18:10.46 | lkcl | well ... *lol* |
| 18:11.13 | lkcl | when i get to it i'll write something in pyjamas. |
| 18:11.15 | penguin42 | lkcl: And definitely not RS or Farnell |
| 18:11.43 | lkcl | penguin42: awwwwww :) |
| 18:12.29 | penguin42 | lkcl: So is it a matter of still just waiting for the connectors at the moment before manufacturing the alphas? |
| 18:12.37 | A2Sheds | yeah |
| 18:12.41 | A2Sheds | just connectors |
| 18:13.11 | A2Sheds | I'm just going to push ahead with what we have now |
| 18:13.30 | A2Sheds | make any changes if the connectors need mods |
| 18:13.32 | lkcl | A2Sheds: whew. good on. |
| 18:13.55 | A2Sheds | lkcl: computex is sold for floor space |
| 18:13.55 | lkcl | A2Sheds: btw, do contact that guy who has the HDMI one that's 1/3 the price of the MOLEX part |
| 18:13.59 | penguin42 | pity, you've missed my lond easter break for playing! |
| 18:14.04 | lkcl | A2Sheds: ah well |
| 18:14.07 | penguin42 | long |
| 18:14.07 | A2Sheds | know anyone that would give us some shared space? |
| 18:14.20 | specing | Hmm, is there a datasheet in english for the A10 SoC? |
| 18:14.38 | A2Sheds | room for only 2 booth babes and a few cards :) |
| 18:15.13 | A2Sheds | I guess we could have 1 booth babe, but it would effect our credibility :p |
| 18:15.41 | penguin42 | ah - so this is why booth babes are so slim? |
| 18:15.53 | A2Sheds | yes lol |
| 18:16.54 | lkcl | specing: under NDA, there's a 70 page document available.... copies of which can be found online :) |
| 18:17.03 | specing | 70page? |
| 18:17.07 | A2Sheds | I have to check with the Linux tabletop area as well |
| 18:17.08 | specing | that is little |
| 18:17.24 | specing | lkcl: do you have a link or a filename I should search for? |
| 18:17.31 | lkcl | specing: yeah something like that. yes, that's right. it's why we are working from the source code as well |
| 18:17.53 | lkcl | specing: look in the arm-netbook archives, people have posted links before. |
| 18:18.01 | A2Sheds | and of course the imx6 gets launched the same week as computex |
| 18:18.28 | lkcl | A2Sheds: jooooy. are they telling people its price? |
| 18:18.38 | A2Sheds | we could buy a dev board, and swap the imx6 from their card to ours |
| 18:19.01 | A2Sheds | no, it's the typical clustef**k |
| 18:19.47 | A2Sheds | nobody has any clear info other than the big release party and the strategic customers with dev boards |
| 18:19.57 | penguin42 | hmm I hadn't realised the iMX6 was such a big range - all the way from a '6SoloLite' with a single A9 upto a quad |
| 18:20.42 | A2Sheds | ikcl: any word back from Samsung on their 2 or 4 core cortex a-9?? |
| 18:20.58 | lkcl | A2Sheds: naah. |
| 18:21.33 | A2Sheds | http://am.renesas.com/applications/automotive/cis/cis_highend/rcar_h1/index.jsp |
| 18:21.37 | A2Sheds | was the other |
| 18:21.53 | A2Sheds | more vapor looking for a sucker |
| 18:21.56 | lkcl | penguin42: yeah, but it's kinda a bit late. the iMX6 should have been available 6-8 months ago |
| 18:22.02 | lkcl | A2Sheds: awwww, i kinda liked that one :) |
| 18:22.20 | A2Sheds | heh, at least freescale isn't 2 years behind :) |
| 18:22.27 | A2Sheds | so they are getting closer |
| 18:22.55 | lkcl | he he |
| 18:22.56 | Gumboot | I reckon you could probably fully document an SoC in 70 pages if you glossed over all the guessable details. |
| 18:23.12 | A2Sheds | http://www.xilinx.com/products/silicon-devices/epp/zynq-7000/index.htm waiting to try them |
| 18:23.44 | lkcl | Gumboot: it's more kinda a summary. big summary, of features on a per-section basis. then they just dove straight into the source code, sod the documentation |
| 18:23.59 | A2Sheds | too bad they are only 800mhz cores |
| 18:23.59 | lkcl | Gumboot: they just weren't expecting anyone to be.... interested in it! |
| 18:24.25 | lkcl | A2Sheds: aww come on, that's still screamingly good. and means only 0.5 watts too. no, less: it's 0.28 nm |
| 18:24.38 | A2Sheds | in 6 months well be using the a10+ or A12 or whatever they call it |
| 18:24.46 | A2Sheds | hey if they work, lets see |
| 18:25.17 | A2Sheds | I'm all for ARM + fpga at ~$10@ 2M pieces |
| 18:25.22 | Gumboot | It's going to get confusing, isn't it? Collisions between Allwinner A*, Apple A*, and Cortex A*... |
| 18:26.11 | A2Sheds | the problem with the xilinx is no GPU! |
| 18:26.29 | A2Sheds | same as the parts Marvell offered, the GPU-less desigins |
| 18:27.07 | Gumboot | Are there any worthwhile attempts at open-source GPUs? |
| 18:27.13 | A2Sheds | sort of |
| 18:27.14 | lkcl | A2Sheds: ohh, you remember somebody pointed out icubecorp? |
| 18:27.52 | A2Sheds | http://icubecorp.com/ |
| 18:27.57 | lkcl | yep. |
| 18:28.05 | lkcl | we got in touch with them (RHT). |
| 18:28.26 | A2Sheds | anything we can use? |
| 18:28.38 | lkcl | they've been doing some samples on the MVP programme with TSMC |
| 18:29.24 | lkcl | they _should_ have something in about 6 months. it'll be 65nm and 700mhz, dual-core, but each core will have 8 threads for video processing and 6 for graphics processing. |
| 18:29.35 | lkcl | _and_ it's *entirely* software-programmed. |
| 18:30.02 | lkcl | they have a compiler expert who used to work for SGI. he's chosen the http://open64.net route |
| 18:30.53 | lkcl | it's just that they're a leeetle tiny bit delayed. like... they should have come out with this UPU (CPU+GPU hybrid) ooo.... 3 maybe 4 years ago |
| 18:31.00 | A2Sheds | oh great |
| 18:31.07 | lkcl | but.... it's entirely software-programmable! |
| 18:31.15 | A2Sheds | well AMD might have ARM + radeon by then |
| 18:31.26 | lkcl | which means that when ramped up to 28nm it becomes _seriously_ competitive |
| 18:31.46 | lkcl | A2Sheds: that'll be something... but nowhere near as price-competitive as this thing will be. |
| 18:31.55 | A2Sheds | we can get PCIe radeon parts as well |
| 18:32.22 | A2Sheds | we just need to port ATOM BIOS to arm and work out the unified memory stuff |
| 18:32.25 | lkcl | i've not been able to establish what ARM charges for licenses but it's making a big dent in all the SoC pricing |
| 18:32.46 | lkcl | ohh bloody hell. bloody ATOM BIOS. what _is_ it with that crap, why is it still proprietary?? |
| 18:32.53 | A2Sheds | it's a floating price |
| 18:33.20 | lkcl | i wonder if AMD will rip ATOM BIOS out of the ARM+radeon jobbie? |
| 18:33.24 | lkcl | probably not... |
| 18:33.29 | A2Sheds | heh, ask Bridgman |
| 18:33.51 | A2Sheds | ask libv about him, :) |
| 18:34.20 | lkcl | mm? :) |
| 18:34.29 | A2Sheds | he must have some files, pictures or something similar tucked away in a safe deposit box |
| 18:34.48 | A2Sheds | he's indestructible |
| 18:35.13 | A2Sheds | he's always the one behind the hold up of open source anything from AMD |
| 18:35.32 | lkcl | ahh deep joy. |
| 18:35.39 | A2Sheds | coreboot, drivers, why atom bios is closed |
| 18:35.45 | lkcl | is going to light a fire |
| 18:35.54 | A2Sheds | he came from ATI |
| 18:35.56 | lkcl | safely. |
| 18:36.20 | A2Sheds | it always goes back to bridgman |
| 18:36.28 | A2Sheds | he's in #radeon |
| 18:36.31 | penguin42 | I thought they had Radeon going on ppc |
| 18:37.03 | A2Sheds | not right now, but he's usually in there during the week |
| 18:37.30 | A2Sheds | I talked to the radeon devs about ARM 2 years ago |
| 18:37.40 | A2Sheds | they were all interested in helping |
| 18:37.50 | A2Sheds | we just never pulled the trigger |
| 18:38.29 | A2Sheds | ask arlied or libv |
| 18:39.18 | A2Sheds | ATOM Bios is reversible, even some tools out there that automate much |
| 18:39.27 | A2Sheds | that was going to be rather simple |
| 18:39.41 | penguin42 | A2Sheds: Any idea how benh does it on ppc ? |
| 18:40.02 | A2Sheds | but UMA was going to need some rework since ARM does not have UMA setup like x86 |
| 18:40.18 | A2Sheds | penguin42, yeah, they just reworked ATOM bios |
| 18:40.37 | A2Sheds | that why libv and arlied said it wouldn't be too much work |
| 18:41.09 | penguin42 | A2Sheds: Ah ok; if you can do it on ppc it shouldn't be too bad on ARM, at least you're not going to be fighting endianness issues |
| 18:41.35 | A2Sheds | DrJakob also works with Tungsten graphics / now VMware on all this |
| 18:42.02 | A2Sheds | they all said it was very workable |
| 18:42.17 | A2Sheds | all those guys know each other |
| 18:42.51 | A2Sheds | I'm all for an ARM soc with PCIe + Radeon |
| 18:43.13 | penguin42 | nod |
| 18:43.19 | A2Sheds | you can slow down the GPU engines to lower the poser consumption |
| 18:43.24 | A2Sheds | poser/power |
| 18:44.04 | ManoftheSea | light a fire? |
| 18:44.08 | A2Sheds | heck, 2 x86 cores + GPU is under 5w right now for the AMD Fusion parts |
| 18:44.37 | A2Sheds | and those are 1ghz dual cores for the cpu and at least 500MHz for gpu stream processors |
| 18:44.38 | penguin42 | A2Sheds: It'sa bad sad, I think you can probably attach as much RAM to the Radeon chips these days as the ARM Socs |
| 18:45.03 | penguin42 | ^bad^bit |
| 18:45.08 | A2Sheds | heh 1GB for the A10, the radeons 4GB |
| 18:46.05 | A2Sheds | I'm am working on the Fusion APU eoma68 card as well |
| 18:46.14 | A2Sheds | just to do it |
| 18:46.31 | A2Sheds | that will give us some market feedback |
| 18:46.42 | penguin42 | what's the price comparison to the ARM ones? |
| 18:47.23 | A2Sheds | well a fusion mini-itx board is ~$100 |
| 18:48.05 | A2Sheds | so it's going to be in that range since we are also including 1-4GB of DDR3 on card :) |
| 18:49.07 | penguin42 | is the HP Microserver based on the same chips? They work out about £130 including disk, case, PSU etc |
| 18:49.11 | A2Sheds | they will also have coreboot support so 1-2 sec boo times from power-on to boot prompt |
| 18:49.29 | A2Sheds | Intel has no competitive solution |
| 18:50.14 | A2Sheds | they have turned their back on coreboot, so they are 6 months + out from having a fast power-on to boot prompt solution firmware |
| 18:50.17 | Gumboot | Has anybody, recently, tried to slip through a 2D accelerator with an OpenGL interface on the hope of dealing successfully with all of the GUI composition without really bothering too much about 3D games? |
| 18:50.39 | A2Sheds | Gumboot, in an ARM soc?? |
| 18:50.50 | Gumboot | Anything, really. ARM SoC seems most likely. |
| 18:51.15 | A2Sheds | not that i am aware but it would be perfect, low power and low cost for 99% of the actual apps |
| 18:51.21 | Gumboot | I just boggle at the way GPU performance has become so critical whilst being used for such mundane tasks. |
| 18:51.29 | Gumboot | Yeah, that. |
| 18:51.30 | penguin42 | Gumboot: The KDE set doesn't need GL, it can switch back to something else (XRender?) |
| 18:51.44 | A2Sheds | yeah, who plays 3d games or renders 3d cad models on their tablet?? |
| 18:51.46 | lkcl | Gumboot: staggering, isn't it. that's why i like what icubecorp and ignenic did |
| 18:51.52 | A2Sheds | only a few for games |
| 18:52.08 | A2Sheds | nobody uses ARM for CAD |
| 18:52.15 | penguin42 | A2Sheds: I can see AR stuff would be cute on tablet |
| 18:52.25 | ManoftheSea | not just tablet. |
| 18:52.35 | ManoftheSea | When you've got a compute rthat small, you wear it. |
| 18:52.39 | ManoftheSea | Get a HUD. |
| 18:52.57 | Gumboot | The worst of it is watching everybody move things like web browsers over to GL composition. In principle it shouldn't become less efficient, but when everything runs through a texture unit for what should generally be trivial memcpy() operations, it's an enormous leap backwards. |
| 18:53.47 | penguin42 | Gumboot: Nod |
| 18:54.01 | A2Sheds | http://www.amd.com/us/products/embedded/graphics-processors/Pages/ati-radeon-e4690.aspx |
| 18:54.07 | Gumboot | I know I was speaking to one vendor who said that they do try to detect (heuristically) the behaviour and back it with a sensible implementation, but I can't remember who. |
| 18:54.25 | Gumboot | Was it in here? |
| 18:54.35 | ManoftheSea | Gumboot: you mean you don't want the interface from Jurassic park on your web browser? |
| 18:54.36 | A2Sheds | AMD has some similar discrete GPU with PCIe we could buy if we have the end customers |
| 18:54.50 | Gumboot | ManoftheSea: That was unix, right? |
| 18:55.01 | Gumboot | I barely remember the movie, or the interface. Just the jokes. |
| 18:55.29 | penguin42 | A2Sheds: What's the difference between that and say the chips on the HD4xxx PCIe cards in PCs ? |
| 18:55.30 | A2Sheds | what was the GPU in the marvell dev box? |
| 18:55.46 | penguin42 | Gumboot: That was a program called FSN on SGI boxes |
| 18:55.50 | A2Sheds | penguin42, sometimes just the part numbers |
| 18:56.15 | A2Sheds | Open RD |
| 18:58.15 | A2Sheds | the opneRD has a 2d discrete part by Volari Z11 |
| 18:58.49 | A2Sheds | http://en.wikipedia.org/wiki/XGI_Technology |
| 18:59.43 | A2Sheds | http://www.seco.com/en/item/secomgm-z11 |
| 19:00.31 | A2Sheds | http://www.xgitech.com/ is down |
| 19:01.01 | penguin42 | I guess less than 2W is decent |
| 19:01.02 | A2Sheds | so I don't know if any parts are left anywhere or if the masks are even around |
| 19:01.32 | A2Sheds | http://www.tridentmicro.com/ is still alive |
| 19:01.43 | penguin42 | anyway, time to disappear |
| 19:03.27 | A2Sheds | http://www.phoronix.com/scan.php?page=news_item&px=ODQ5Nw |
| 19:11.18 | A2Sheds | http://www.eetimes.com/electronics-news/4304118/Imagination-outstrips-all-other-GPU-IP-suppliers |
| 19:11.23 | A2Sheds | just FYI |
| 19:13.57 | A2Sheds | http://www.eetimes.com/electronics-news/4211563/Vivante-s-1-GHz-multicore-GPU-sampling- |
| 19:14.14 | A2Sheds | http://www.eetimes.com/electronics-products/advanced-technology/4217401/Curtiss-Wright-rolls-GPU-based-on-Nvidia-Fermi-architecture |
| 19:15.45 | A2Sheds | but still no real new choice for a 2D discrete GPU, everybody just integrates a core into the SOC |
| 19:30.59 | *** join/#arm-netbook lkcl (~lkcl@host86-131-170-29.range86-131.btcentralplus.com) |
| 20:19.46 | *** join/#arm-netbook L84Supper (~ly@unaffiliated/l84supper) |
| 20:38.57 | *** join/#arm-netbook toxicpsion (~toxicpsio@out-ab-210.wireless.telus.com) |
| 20:48.18 | toxicpsion | mnemoc: did you ever get your livesuit running in a VM? |
| 20:52.19 | mnemoc | toxicpsion: I didn't try again |
| 20:54.12 | toxicpsion | thouht i'd check, browsing ibot to catch up.. that's how i flash here - XP on Virtualbox on Ubuntu. |
| 20:57.23 | hno | toxicpsion or mnemoc, would you be able to measure voltage on a pin of the recovery button while pressed? Need to know if it's an active-high or active-low pin. |
| 21:02.33 | toxicpsion | hno: only got my tablet... and can't get probes at it easily... |
| 21:07.47 | hno | toxicpsion, that's ok. I'll figure it out. Should be connecting to GND I think. |
| 21:08.22 | hno | toxicpsion, something you can answer however, do the recovery button also reset the device? |
| 21:08.45 | hno | or do one need to hold down the recovery button and then power on? |
| 21:12.33 | toxicpsion | it does on mine. |
| 21:12.43 | toxicpsion | no need to hold. |
| 21:18.34 | hno | ok, thanks. that simplifies hunting down the signal greatly. |
| 21:24.39 | hno | ans wonder what the 2 exra pins in USB4 is on the is for. A guess is that one is 3.3V power but not sure. |
| 21:27.19 | hno | miscounted the USB interfaces before. There is two internal USB headers in the mele (USB4 & USB6), where USB6 is the OTG port. |
| 22:01.40 | zenitraM | hno, toxicpsion: on mine the recv button doesn't reset the device |
| 22:02.12 | mnemoc | which is the recovery button? |
| 22:02.33 | zenitraM | it is a pinhole button |
| 22:02.56 | zenitraM | pressing it while pressing the ON button with the tablet powered off will make it boot directly to FEL |
| 22:03.17 | mnemoc | ah, a tablet |
| 22:03.31 | mnemoc | I was scratching my head while searching for a pinhole on the mele |
| 22:04.32 | zenitraM | hno is trying to find where the button that triggers FEL is on the mele. probably on the extra usb pin. |
| 22:09.51 | hno | not entirely sure it's the extra USB pin. It's an OTG port so it actually needs 5 pins. (VCC, d+, d-, GND, id) |
| 22:12.20 | hno | and do not make much sense to put it there if FEL also requires power button to operate. |
| 22:42.25 | *** join/#arm-netbook steev (u1383@gateway/web/irccloud.com/x-terkzahnkxyzjcbq) |
| 22:42.54 | *** join/#arm-netbook steev (u1383@gentoo/developer/steev) |
| 23:46.02 | *** join/#arm-netbook CaCtus491 (~Kent@123-243-197-152.static.tpgi.com.au) |