| 02:27 | randomnickname | hello |
| 02:27 | lubiana | hi |
| 02:28 | randomnickname | how warm does the hummingboard's SoC module get in typical use? |
| 12:29 | randomnickname | jnettlet: seems like you were the very first to bring up a kernel on a hummingboard prototype! |
| 12:41 | jnettlet | randomnickname, the upstream kernel will run on it as well. But most the community kernels are based from my work. |
| 12:46 | randomnickname | jnettlet: just out of interest, what was the version of the very first stable kernel for the hummingboard? |
| 12:46 | randomnickname | bringing up a new ARM platform for linux seems like a lot of work |
| 12:48 | jnettlet | randomnickname, when the project was started 3.4 was the latest but "stable" was 3.0.35 |
| 12:49 | jnettlet | we started our work on 3.10, and then I moved to 3.14 for stability reasons and some features were added that I wanted |
| 12:49 | randomnickname | if my research is correct, the hummingboard's general availability was around january of this year? |
| 12:51 | jnettlet | something like that. But it is based off the original dev platform called the carrier one. |
| 12:51 | randomnickname | with the hummingboard practically being a cut-down cubox, i am surprised it had to have its own kernel |
| 12:51 | jnettlet | it doesn't |
| 12:52 | jnettlet | they just have different device-tree entries |
| 12:52 | jnettlet | other than minor hardware differences they are exactly the same just a different form factor |
| 12:52 | randomnickname | oh, okay. i must have misunderstood something then |
| 12:55 | randomnickname | i see two official FW images for the hummingboard, both of jessie |
| 12:57 | randomnickname | i'll try both to find out how much improvement 7 months of development brings to a platform |
| 12:57 | randomnickname | and perhaps try building a wheezy userland for use with that kernel, as opposed to the original jessie-testing |
| 12:58 | jnettlet | well those aren't even very representative. We are currently working on a build server that will push out optimized packages for most the major distributions. |
| 12:59 | randomnickname | jnettlet: what needs optimizing in general? |
| 13:00 | randomnickname | it was my understanding that the raspberry pi needs its own derivative of debian because of armv6 and hardfloat |
| 13:01 | randomnickname | the hummingboard is a common armv7 platform, and what about the "standard" debian arm releases would need to undergo optimization? |
| 13:01 | jnettlet | mostly things regarding graphics and multimedia |
| 13:01 | randomnickname | i see. driver-related work? |
| 13:02 | jnettlet | and then some other things. In general most the major distributions provide a pretty poor ARM experience out of the box |
| 13:02 | randomnickname | right. that is sad to hear |
| 13:02 | jnettlet | well that is generally because they are trying too hard to keep things generic and only support what is upstream. |
| 13:03 | jnettlet | since ARM is still evolving so much that puts you easily a year or more behind what is optimal. |
| 13:03 | randomnickname | hmm. |
| 13:04 | randomnickname | take something such as windows 7 for example, it is pretty generic as far as hardware support goes, will run on any PC with a P2 and up |
| 13:04 | jnettlet | That is mainly why ChromeOs/ChromiumOs is built from Google's own portage tree |
| 13:04 | randomnickname | maybe it is different in the linux world |
| 13:04 | jnettlet | no it is about architecture. |
| 13:05 | jnettlet | Windows 7 won't just run on any ARM chip |
| 13:05 | randomnickname | right, I am aware of that |
| 13:06 | randomnickname | my point was, windows 7, even though it can run on a P2 (and thus doesn't make full use of newer x86 features like OS X, for example), can be pretty fast on modern hardware |
| 13:06 | randomnickname | does it work about the same for generic ARM linux builds? |
| 13:08 | randomnickname | sorry for the misclarification |
| 13:08 | jnettlet | no, it is all about the drivers. The ARM cores work like that but ARM cores are very generic, where ARM gets its performance with low power consumption is by having lots of small dedicated chips for different functions. the VPU, GPU, DMA engines |
| 13:09 | jnettlet | intel just has really elaborate high power cpus that can do lots of things really fast, but need a lot more mpower |
| 13:11 | randomnickname | ah! that's how it is |
| 13:12 | randomnickname | i have always wondered why there aren't new variants of windows for each new generation of intel processors, but every ARM SoC needing its own tailored kernel |
| 13:15 | randomnickname | I think U-Boot's bmp display machinery needs work as well |
| 13:16 | randomnickname | where the image is, it just displaces the text on the video console, instead of letting text start after the image |
| 13:18 | randomnickname | in arch/arm/lib/board.c of u-boot-imx6, it goes static int display_banner(void){ printf("\n\n%s\n\n", version_string); |
| 13:19 | randomnickname | something tells me i can just slap more \n's before the %s to clear the vertical area of the image, but that won't be an optimal solution |
| 13:54 | cbxbiker61 | http://www.thingiverse.com/thing:496637 a pretty cool hummingboard case, nice with the cut-out for the heat sink |
| 14:00 | Artox | [13:12] i have always wondered why there aren't new variants of windows for each new generation of intel processors, but every ARM SoC needing its own tailored kernel |
| 14:00 | Artox | thats because a) an SoC is more than a cpu |
| 14:00 | Artox | b) things are not wired up in a standard way |
| 14:01 | cbxbiker61 | things are much better with the device tree kernel's |
| 14:01 | Artox | c) drivers need to land in mainline |
| 14:01 | Artox | cbxbiker61: yes, but watching e.g. my phone (omap3530), its still not fully in mainline and its beint integrated for 2 years now |
| 14:03 | cbxbiker61 | phone chips are less likely to get the "generic" work, since the phone providers like proprietary lockin |
| 14:03 | Artox | yes. you need to carefully choose your phone |
| 14:04 | randomnickname | thank you, Artox |
| 14:07 | randomnickname | any word on the i4 pro SoM having a heatsink included? |
| 14:27 | randomnickname | cbxbiker61: not bad color matching for the case |
| 14:27 | cbxbiker61 | yep |
| 14:27 | randomnickname | but, i'd be mindful of the heatsink getting dusty... |
| 14:28 | cbxbiker61 | always compromized, i'd just blow it off with compressed air |
| 14:29 | randomnickname | hopefully, chromium can run well on the cubox/hummingboard |
| 14:40 | Artox | randomnickname: it can |
| 14:40 | Artox | its already been done |
| 14:40 | Artox | plays youtube fine without touching the gpu |
| 14:40 | Artox | and, suing the vpu things are way better |
| 14:40 | Artox | using* |
| 14:43 | randomnickname | great, maybe it can be my main computer |
| 14:43 | randomnickname | i listen to music and surf the net, that's about it |
| 14:43 | randomnickname | and i know gnome has a great fancy text editor, gedit |
| 14:44 | Artox | thats been my vision for a while |
| 14:44 | randomnickname | so that's taken care of |
| 14:44 | Artox | add email and some gaming though |
| 14:44 | randomnickname | for multimedia (movies), i have a normal ultrabook |
| 14:45 | randomnickname | vlc for linux pretty much hates everybody's guts |
| 14:46 | Artox | it works real good on x86 though |
| 14:46 | Artox | has all teh codecs working simpler than gstreamer |
| 14:48 | randomnickname | what is gstreamer? |
| 14:49 | Artox | I am not explaining |
| 14:49 | Artox | if you dont know, you are a lucky one |
| 14:49 | randomnickname | voodoo black magic? |
| 14:50 | randomnickname | i have been intrigued by ARM as a usable desktop platform for a few years now |
| 14:50 | randomnickname | i was not very happy with the raspberry pi, but keeping its price in mind, i have no complaints |
| 14:50 | Artox | well, I have a complaint: its broken |
| 14:51 | Artox | but I agree |
| 14:51 | randomnickname | the other day, i got gimp on it to convert png logos |
| 14:51 | randomnickname | for linux kernel build |
| 14:51 | randomnickname | oh boy, 10 seconds to create a new blank canvas |
| 14:51 | Artox | anyone in here with ipv6 who could perform a ping6 for me? |
| 14:52 | Humpelstilzchen | Artox: target? |
| 14:53 | cbxbiker61 | pm me the ipv6 address |
| 14:53 | randomnickname | i have a beaglebone black on the way |
| 14:54 | Artox | alrady done cbxbiker61, thanks |
| 14:54 | Artox | as in, issue resolved |
| 14:55 | randomnickname | after i am finished with the usual tinkering (compiling kernel and/or bootloader from source), beaglebone will sit in a corner as a console server |
| 14:56 | randomnickname | i have a few embedded devices with serial consoles I like to poke around |
| 14:59 | Artox | so. someone jsut got an ipv6 ticket; hopefully it will work out |
| 15:00 | randomnickname | Artox: how is ARM serving your desktop computing needs? |
| 15:00 | Artox | I had been using the cubox-i with thunderbird for testing; but the opengl* was rather unstable |
| 15:00 | Artox | so right now, not at all |
| 15:00 | Artox | but soon I'll be abroad without x86 |
| 15:01 | Artox | (who wants to carry a pc in a plane) |
| 15:01 | randomnickname | i hope apple switches to ARM somewhat soon |
| 15:01 | Artox | but cubox is just fine |
| 15:01 | Artox | what |
| 15:01 | Artox | why would they do that? |
| 15:02 | Artox | they'd just render their systems incompatible to all commercial games and office programs |
| 15:02 | randomnickname | they probably won't, but I feel that they are best poised to push ARM as a platform for general computing, whether they decide to do it or not notwithstanding |
| 15:05 | randomnickname | there was 68k -> PPC and PPC -> x86 before already |
| 15:06 | randomnickname | i wonder if boot camp users make up a significant percentage of mac purchasers? |
| 15:13 | Artox | ther was ppc before |
| 15:13 | Artox | but now that apple is on x86 |
| 15:13 | Artox | no company will bnother with any bianries otehr than x86 |
| 15:13 | Artox | and its not about bootcamp, its about x86 |
| 15:13 | Artox | the instruction set |
| 15:13 | Artox | and possibly its about OpenGL too. on arm you usually dont get desktop OpenGL |
| 15:14 | Artox | thus the new incompatibilities |
| 15:14 | Artox | I would say though that microsoft did an interesting move in making directx11 the same for arm and x86 |
| 17:24 | Exaga | LOL! Hey... just been looking at the specs of the new raspberry Pi model A+ and seen that they've incorporated the composite video output into the 3.5mm audio jack socket - just like the HummingBoard! \o/ :D |
| 17:26 | Exaga | and using microSD cards now instead of full size SD cards - just like the HummingBoard! \o/ :D |
| 17:26 | Artox | well |
| 17:26 | Artox | they know whats good |
| 17:26 | Exaga | exactly ;) |
| 17:26 | Exaga | and hi Artox o/ |
| 17:27 | Artox | my phone has had that for years now |
| 17:27 | Artox | the tvout that is |
| 17:27 | Exaga | i'm still using a Nokia N8 |
| 17:28 | Exaga | old but reliable |
| 17:28 | Artox | lucky you |
| 17:28 | Exaga | lol why lucky? |
| 17:28 | Artox | I got this super-awesome phone for geeks |
| 17:28 | Artox | and its lying around its case open |
| 17:28 | Artox | unusable |
| 17:28 | Exaga | :/ |
| 17:28 | Artox | for a bug aound battery charging |
| 17:29 | Artox | I am just now making a bootabel sdcard with 3.18-rc4 kernel |
| 17:29 | Exaga | for which OS? |
| 17:29 | Artox | opensuse |
| 17:29 | Exaga | cool |
| 17:29 | Artox | I jsut dont get android/replicant, so I use suse for debugging |
| 17:29 | Exaga | i haven't worked on anything for about 5 days |
| 17:29 | Exaga | i need a break |
| 20:56 | Coolg33k | close |