**** BEGIN LOGGING AT Tue Apr 27 02:59:57 2010
Apr 27 03:01:13 <tinker-f595> rektide: real time clock worked for me
Apr 27 03:01:58 <tinker-f595> I remember your comments from when you first received your openRD
Apr 27 03:02:29 <tinker-f595> I thought you had gotten your real time clock working
Apr 27 03:05:28 <tinker-f595> I remember I had to set the hardware clock using hwclock when I first received my openRD Client
Apr 27 03:07:41 <rektide> no i never did.  the self test on boot always FAILED.
Apr 27 03:08:04 <rektide> given that it was a problem since day 1 i'm less concerned about the RTC
Apr 27 03:08:09 <rektide> but my ethernet getting flakier and flakier
Apr 27 03:08:22 <rektide> thats deeply troubling to me.
Apr 27 03:08:52 <tinker-f595> rektide so how come you did not RMA it shortly after you received it?
Apr 27 03:09:14 <rektide> too busy using it to play music. :(
Apr 27 03:09:37 <tinker-f595> hmmmm
Apr 27 03:09:40 <rektide> wasnt a huge deal.  linux works fine without a rtc, as soon as it has a net connection to run ntpdate on
Apr 27 03:10:05 <rektide> i may try writing them and begging for a partially sponsored upgrade
Apr 27 03:10:12 <tinker-f595> yes, but having an RTC not working on a new device indicates some bad mojo
Apr 27 03:10:19 <rektide> fair enough
Apr 27 03:10:34 <rektide> i wrote a test suite for outgoing ARM handhelds at one company
Apr 27 03:10:53 <rektide> we didnt really do any longitudinal studies to see if one error would bubble into another
Apr 27 03:11:19 <rektide> never checked to see whether localized peripherial errors stayed localized or indicated "shit is really bad on this unit"
Apr 27 03:11:24 <tinker-f595> rektide I have not been impressed with einfochips responsiveness to requests for kernel, driver, and uboot updates
Apr 27 03:11:36 <rektide> yes, indeed. :(
Apr 27 03:12:03 <rektide> einfochips is part to blame here.  marvell also bears significant culpability.
Apr 27 03:12:30 <rektide> its ridiculous that they spent so very much effort getting the damned thing going
Apr 27 03:12:39 <rektide> building that one off 2.6.22 kernel just for the platform
Apr 27 03:12:47 <tinker-f595> considering that they have eSata I would of expected them to make sure uBoot worked with it from day one and allowed you to boot a kernel on a sata drive connected via eSata
Apr 27 03:13:03 <rektide> with really no vision or will or understanding of how that was going to have to be broken down into actual integration with upstream
Apr 27 03:13:11 <tinker-f595> yes
Apr 27 03:13:35 <rektide> is uboot still in a decrepit state?
Apr 27 03:13:42 <rektide> that was the one i thought we'd get sooner or latter
Apr 27 03:13:51 <rektide> sorry but i never had fantasies -client was going to get any real support
Apr 27 03:13:52 <tinker-f595> and the openRD Client needs the XGI video drivers updated and mainlined
Apr 27 03:14:47 <rektide> XGI are the people that need to update and integrate XGI
Apr 27 03:15:02 <tinker-f595> well I asked on the openrd list the other day and someone claimed the uBoot at denx will allow eSata to be used.  I need to build the source and install it and see
Apr 27 03:15:29 <tinker-f595> well supposedly einfochips was trying to work with them or at least get the spec
Apr 27 03:16:34 <rektide> past summer / fall someone seemed to be on the case for getting openrd support mainlined into uBoot
Apr 27 03:16:52 <rektide> at which point, i abnegate any further responsibility from einfochips w/r/t openrd-{base,client}
Apr 27 03:17:12 <rektide> eSata is just gonna be an upstream capability that'll happen to work on openrd, or we can file a bug that their impl doesnt work on openrd
Apr 27 03:18:33 <rektide> "state of XGI" circa mid 2007 article from phoronix: http://www.phoronix.com/scan.php?page=article&item=717&num=2
Apr 27 03:18:35 <tinker-f595> well I looked through the the uboot tree at denx and could not see the marvell sata driver code.  I assume that they must be using some generic code that works by luck
Apr 27 03:18:51 <rektide> haha, its amazing how much AHCI gives us
Apr 27 03:19:05 <rektide> this laptop has an intel ata_piix chipset
Apr 27 03:19:22 <rektide> but compiling a kernel with that causes crazy long kernel panic at boot. :(
Apr 27 03:19:25 <rektide> i'm running without dma
Apr 27 03:19:26 <tinker-f595> ah I remember the other thing I noticed is that the upstream uboot code only refers to one drive.
Apr 27 03:19:26 <rektide> life sucks
Apr 27 03:19:26 <rektide> bad
Apr 27 03:19:41 <rektide> heh, yeah, i run a 4 bay enclosure on my openrd
Apr 27 03:19:57 <rektide> i just, uh, dont uboot off it.  cause i'm still on stock uboot.
Apr 27 03:19:59 <tinker-f595> I guess I could fix that by looking at the kernel code and working out what to put in the uboot code
Apr 27 03:20:07 <rektide> all attempts to upgrade have ended in despair
Apr 27 03:20:12 <tinker-f595> hmmm
Apr 27 03:20:22 <rektide> (mind you i havent tried in >8 months)
Apr 27 03:20:43 <rektide> so uh
Apr 27 03:20:46 <rektide> on a totally different topic
Apr 27 03:20:51 <rektide> AMD's bulldozer is finally getting here
Apr 27 03:21:00 <rektide> DX11 gpu, dual core, sub 1w
Apr 27 03:21:07 <tinker-f595> ???
Apr 27 03:21:09 <rektide> only like 8-16 months
Apr 27 03:21:14 <rektide> x86 land
Apr 27 03:21:21 <tinker-f595> ah
Apr 27 03:21:39 <rektide> i look forward to SoC with badass DX11 and dual HDMI ports
Apr 27 03:21:48 <tinker-f595> hmm
Apr 27 03:21:52 <rektide> best altoids tin computer evar
Apr 27 03:21:59 <tinker-f595> brb
Apr 27 03:23:54 <tinker-f595> The Armada 510 has something called a "GC600 3D Gfx" on the SOC
Apr 27 03:24:18 <tinker-f595> http://www.marvell.com/platforms/Marvell_Armada510_Platform.pdf
Apr 27 03:27:01 <tinker-f595> actually this is a better description of that SOC http://www.marvell.com/products/processors/applications/armada_500/Marvell_Armada510_SoC.pdf
Apr 27 03:27:19 <rektide> 3d is just freaking hard
Apr 27 03:27:25 <tinker-f595> yes
Apr 27 03:27:56 <rektide> frankly SoC these days doesnt have nearly the ram bandwidth to even pretend to play in the 3d world
Apr 27 03:28:17 <rektide> bulldozer should be rocking a stupid fast 128 bit DDR5 interface or something obscene
Apr 27 03:28:29 <rektide> which will still be the limit
Apr 27 03:28:38 <tinker-f595> hmmm
Apr 27 03:28:56 <rektide> wii and gamecube were smart
Apr 27 03:29:07 <rektide> they have like 24MB of really fast onchip ram
Apr 27 03:29:15 <rektide> used for framebuffer / some small texture cache
Apr 27 03:29:33 <rektide> lets em get away with not needing dedicated ram for the gpu
Apr 27 03:29:50 <rektide> i think it was an ati + ibm collaboration, iirc
Apr 27 03:29:57 <rektide> old memory, not 100%
Apr 27 03:30:02 <rektide> anyways, its niche as fuck
Apr 27 03:30:10 <rektide> just so happens to be a niche that tickles my fancy
Apr 27 03:30:20 <rektide> GPU SoC, that is
Apr 27 03:30:38 <rektide> 1T, the on chip memory, ids just a cute annecdote
Apr 27 03:30:45 <rektide> *is just
Apr 27 03:31:23 <tinker-f595> rektide: makes me think of the Amiga and the chip ram for graphics
Apr 27 03:31:52 <rektide> heh, did it come on DIMM's?
Apr 27 03:32:01 <rektide> i'm trying to get an amiga emulator up
Apr 27 03:32:03 <rektide> so i can run k240
Apr 27 03:32:10 <rektide> i dont know jack all about amiga
Apr 27 03:32:23 <tinker-f595> rektide: I would love to find an old amiga in working condition
Apr 27 03:32:32 <rektide> aside from someone trying to sell a desktop PPC off FreeScale hw for some obscene price
Apr 27 03:32:47 <rektide> why not just emulate?
Apr 27 03:32:57 <rektide> i've revised my priorities; the less systems the better
Apr 27 03:33:07 <tinker-f595> well not quite the same as having the hardware
Apr 27 03:33:50 <rektide> i'm not unsympathetic, but this was always a sticking point between me and two of my geek friends in HS.  they were both insistent that video gaming without an arcade cabinet was folly, not the real deal.
Apr 27 03:34:05 <rektide> i could understand their affection, but really i didnt see what i was misisng usually
Apr 27 03:36:22 <tinker-f595> the cramps in the legs from standing hour on end ;)
Apr 27 03:36:43 <rektide> heh apparently
Apr 27 03:36:47 <tinker-f595> http://oldcomputers.net/amiga1000.html
Apr 27 03:36:49 <rektide> grainy ass screens that sometimes blew up
Apr 27 03:37:06 <rektide> now a days we pay extra for that, its called AA
Apr 27 03:37:56 <rektide> freescale is still kicking out coldfire and 68k
Apr 27 03:39:01 <tinker-f595> hmmm
Apr 27 03:42:48 <rektide> i was so ready to go PPC for a while
Apr 27 03:42:50 <rektide> PA Semi was a powerhouse
Apr 27 03:43:05 <rektide> Freescale was announcing tons of cool multicore SoC
Apr 27 03:43:40 <rektide> then apple bought PA Semi, the engineers quit, formed Agnilux, and then got bought by google
Apr 27 03:44:01 <rektide> and Freescale just hasnt been delivering.  they have some pretty sweet dual cores that've been out a while, but they're just not advancing like they'd need to
Apr 27 03:44:24 <rektide> and Freescale is contracting out all the SDK support to other people, which is even worse than the OpenRD state of affairs
Apr 27 04:05:54 <tinker-f595> hmmmm
Apr 27 04:06:35 <tinker-f595> well Marvell has some dual cores at 2 GHZ and TI may be working on some
Apr 27 04:06:45 <tinker-f595> or have some
Apr 27 06:02:52 <rektide> ti is indeed going "balls out" with A9 dual cores i understand
**** BEGIN LOGGING AT Tue Apr 27 20:12:00 2010
**** ENDING LOGGING AT Wed Apr 28 02:59:56 2010