#emc-devel | Logs for 2007-07-20

Back
[04:13:59] <steve_stallings> steve_stallings is now known as steves_logging
[19:11:10] <skunkworks> Hey - where have you been?
[19:11:33] <SWPadnos> working, mostly
[20:24:30] <SWPadnos> so - what's the state of SMP+RTAI at the moment?
[20:24:37] <SWPadnos> is it still "works for some people" ?
[20:25:29] <cradek> it seems to work fine
[20:25:49] <cradek> but, it's not possible to build a smp+rtai kernel that runs on all (reasonable) machines
[20:26:00] <SWPadnos> heh - that's my problem then ;)
[20:26:08] <cradek> what is?
[20:26:34] <SWPadnos> I have a project that I'll be using HAL for, and many of the higher end industrial computers have dual-core CPUs
[20:26:44] <SWPadnos> I don't have a reasonable machine, I thinl
[20:26:45] <SWPadnos> think
[20:27:05] <alex_joni> hi guys
[20:27:06] <SWPadnos> the SMP kernel you had built a long time ago didn't boot at all on my dual dual-core
[20:27:08] <SWPadnos> hi Alex
[20:27:12] <cradek> the specific problem I had was the smp+rtai kernel requires a local apic, even if on a mono cpu machine
[20:27:24] <cradek> oh, well that may be another issue
[20:27:27] <SWPadnos> the SMP kernels I built had to have module versioning enabled to boot
[20:27:36] <SWPadnos> which of course doesn't work with RTAI
[20:27:46] <cradek> I'm not surprised if you have to build a different kernel for each kind of smp machine
[20:28:04] <cradek> making a kernel that runs on many machines does not seem to be an important goal for the rtai folks
[20:28:08] <SWPadnos> indeed
[20:28:23] <SWPadnos> as long as I can get it working on whatever board I end up with, I'm OK
[20:28:41] <SWPadnos> but then again, I'd like to have some idea if it'll work before buying the board(s)
[20:28:42] <alex_joni> SWPadnos: it might take a couple iterations
[20:28:49] <SWPadnos> it already has ;)
[20:28:59] <alex_joni> of kernels I mean
[20:29:21] <SWPadnos> heh - I meant kernels too
[20:29:30] <alex_joni> ah, ok :)
[20:29:37] <SWPadnos> though I don't have any industrial boards at teh moment - I'm trying to get an order together now
[20:30:00] <alex_joni> I have an older one.. no SMP and rtai was working just fine on it
[20:30:32] <SWPadnos> I had tried to get SMP/RTAI working on my dual dual-core, but it never did
[20:31:00] <SWPadnos> Hmmm - my wife's away - maybe I should try the liveCD on her machine, to see if that chipset is any good
[20:31:21] <SWPadnos> I tried on an Nvidia 6100 - it was absolute crap
[20:31:34] <SWPadnos> I think I even tried without X running, and it was still crap
[20:52:29] <SWPadnos> well, it looks like my wife's PC is reasonable for RTAI
[20:52:38] <SWPadnos> max latency of ~13 uS
[20:53:19] <SWPadnos> maybe I'll get myself a similar machine and do some SMP testing there
[20:57:11] <alex_joni> hmm.. there's a hasselblad at B&H for under 11k$
[20:58:23] <SWPadnos> an H1/2/3, or a film camera?
[20:58:50] <alex_joni> http://www.bhphotovideo.com/c/product/437386-REG/Hasselblad_70320525_503CWD_SLR_Digital_Camera.html
[20:59:00] <alex_joni> 503CWD.. seems new
[20:59:24] <SWPadnos> the 503 is an old design, but the back is pretty reasonable looking
[20:59:56] <alex_joni> doubtfull that it's an interchangeable back
[21:00:12] <SWPadnos> it is
[21:00:25] <SWPadnos> it's also interchangeable with the film holders, I bet
[21:00:58] <SWPadnos> note the winder crank on the side - I think the body is intended for film use
[21:01:17] <SWPadnos> "The digital part of the Hasselblad 503CWD can be fitted onto all Hasselblad V System cameras as well as view cameras using the Hasselblad film magazine interface plate and flash sync connection to trigger the digital capture."
[21:07:18] <alex_joni> cool :)
[21:07:21] <alex_joni> still mmtb
[23:05:48] <Guest241> Guest241 is now known as skunkworks
[23:24:06] <alex_joni> night all