Nekochan Net

Greets all,

So I installed a Dual Channel Display and a DMediaPro 2 board in my Octane2, and it won't boot:



I installed software for the DM2 but the DCD doesn't need anything, right? I mean, running 6.5.30 as I am, it should be autodetected, right? So I suppose my only option is to take the DCD and DM2 out and try it from the normal VPro output...?

The DCD when connected to a V12 is autodetected, yes. If you do not know for sure that the dcd is good, then bolt 'er up to a v12 and try it. Should work correctly.

Your graphics cable will have to come off one of the dcd ports : if you connect to the primary one on the graphics card it acts weird. That's on a Fuel but should be same on Octane.

The DCD is installed on a V12, with it's DVI outputs running to a DM5 on a VBOB like this (no genlock though):

It may be a strange idea, but have you tried a different monitor?
There were threads in past which described strange errors on
Odyssey and Voyager (SG2) graphics with DVI links. At first they
seemed to have nothing to do with monitors, but they disappeared
when a different monitor was connected...

As I said... a wild guess only... but maybe worth to try...

_________________
= -
+ | d | i | g | i | t | a | l | +apple +[...]

That's the one with a strange error cured by connecting a different monitor to a Prism.
I'm almost sure that I read something similar in context of V12+DCD. But maybe I'm wrong...
http://forums.nekochan.net/viewtopic.php?f=3&t=16724118#p7329471

This one shows a different cable setup with DM2/3, DM5 and V12/DCD. It's a Tezro, but
that shouldn't change much...
http://forums.nekochan.net/viewtopic.php?f=3&t=16724197&start=15#p7330421

Don't know if that helps, but maybe it's a source of some ideas...

_________________
= -
+ | d | i | g | i | t | a | l | +apple +[...]

Actually this is the same monitor I always use with my Octane, it's got both 13W3 and VGA inputs, and I'm using a DVI to VGA converter cable. But either way I don't think hooking an incompatible monitor to the DVI output could cause the kernel panic I'm getting in this case...

The "converter" thing always makes me kind of nervous but it can work ... I ran a GDM-900 (or whatever the letters are, it was the single-plane wide aspect Sony monitor) over a dcd -> dvi -> vga cable. Still, unless you know that your particular cable and monitor do work, you can't discount anything.

Normally the system will autodetect a dcd at boot time and display it in a hinv or the System Manager -> About This System. Obviously should not panic at startup. You should be able to bypass the DM2 entirely, plug directly into the monitor and it should work. Bad dcd perhaps ? Or the DM2 is speaking back through the edid (or not but should?), telling the dcd bad stuff causing the whole mess to fail ?

Bypass the dm2 and see what happens ?

Looks like your problem might be occurring before the kernel attempts to load:
The portion of the Octane mid-plane that receives the XIO carrier has four XIO connectors, designated as quadrants in Octane speak: The original Octane XIO module carrier was quad device capable, while VPro-equipped Octanes use a tri-module carrier. VPro graphics boards connect at Quadrant B/Device 11 (Quadrant A/Device 12 is blocked by the physical height of VPro boards). For the record, some other Octane XIO Devices (not accessed by the XIO quad or tri carrier) are:

the system module/heart chip - XIO Device 8;
the PCI card cage - XIO Device 13;
and the bridge chip that controls Octane IO functions - XIO Device 15.

So in the case of your VPro equipped Octane, the XIO Device (11) failing power-on diagnostics is your V12.

Makes the DCD seem like the obvious suspect, but FWIW, I'd return the Octane to its pre-dcd/dm2 condition to confirm your V12 will pass diagnostics and boot. If V12 minus the DCD doesn't pass, you might take a close look at the V12's compression connector, there have been reports of "XIO Slot Link Failure" errors being resolved by the replacement of the connector. If the freestanding V12 still doesn't pass power-on diagnostics, I'd suggest trying to reinitializing your Octane's POD logs.

If all does seems well with the free-standing V12, testing the DCD and DM2 individually could help rule out the (less likely) possibility of a combined hardware conflict.

All of the VPro boards (V6, 8, 10 and 12) have the DCD connection header, so if testing points at the DCD as the cause of your Octane's power-on diagnostics failure, before you write off a fairly rare/expensive piece of hardware you could attempt a retest of the DCD with any VPro equipped system.

_________________
***********************************************************************
Welcome to ARMLand - 0/0x0d00
running...(sherwood-root 0607201829)
* InfiniteReality/Reality Software, IRIX 6.5 Release *
***********************************************************************

So okay, thanks for the replies guys I bit the bullet and removed both the DCD and the DM2, and it booted just fine. Then I put the DM2 back in, and it booted, and hinv showed that Irix sees the DM2, which is exactly what I'd expect since I installed the DM2 software prior to putting the board in the first time, just as the instructions say you have to do. So now I'm like, is there any point to try putting this DCD back in? It's gotta be smoked, right? And I hate hate hate racking the V12 in and out of my system because I'm so paranoid "it might never work again!"

Well alright, I'm not looking for any replies I'm just venting. I'll buck up my courage and put the DCD back in and edit this post with the results good bad or indifferent...

Get out your 10x scope first and go Sherlock on the connector. Isn't that the most likely failure point for electronic gizmos ?

(That train of thought never worked for me. In my case it was always the most expensive part that failed. I could do my troubleshooting from the price book, but you may be luckier.)

Think again: viewtopic.php?f=3&t=16724589&p=7334060

_________________
Now this is a deep dark secret, so everybody keep it quiet
It turns out that when reset, the WD33C93 defaults to a SCSI ID of 0, and it was simpler to leave it that way... -- Dave Olson, in comp.sys.sgi
Currently in commercial service: (2x)
In the museum: almost every MIPS/IRIX system.
Wanted: GM1 board for Professional Series GT graphics (030-0076-003, 030-0076-004)

You mean I actually have to go out and buy a DVI monitor to use this DCD?

This deal is getting worse all the time...

UPDATE: So I got a DVI monitor, and I hooked it up as in the picture above, and it booted okay but nothing's getting through to the monitor. I found these seemingly relevant tidbits in the syslog file:



It looks like it's seeing the DCD and not liking the vfo file I've been using up to now with the 13W3 output, so it tries to default sensibly, then presumably it's saying "no devices found" because it's not seeing my new DVI monitor at the end of the cable? Anyone know the difference between what SGI thinks a DVI cable should be and what the rest of the world thinks? Or, better yet, where I can get an SGI DVI cable that will work going from a VBOB to a monitor?

^^^^^^^^^^^^^^^^^^^
Edited my last post instead of posting anew, so bumping...

My experience with dcd was on a fool so not exactly the same but ....

It won't use the primary connector. You can get some sort of flashing backy-forthy eye-strain gobbledygook out of the singleton connector but otherwise, no worky.

iirc, you must use one of the 2@ resolutions. With a single monitor, if my oldtimer's hasn't kicked in, I just did everything on the screen that was connected. (Only had one screen.) Either connector would work. When I tried connecting to the original connector I got this flashing stuff that just gave you a headache. This was with a crt.

About dvi cables, I have one that looks great. Really nice, heavy wire, good looking heavy duty connectors. Even says SONY on it. Doesn't work. I have no idea why (except that here is China, for all I know the wires extend an inch into the cable then stop.) Anyway, the point is, in this modern hi-tech Pareto world of profits here, labor there, buy a brand name and screw it up, you can't fucking depend on anything. Every single piece you by is potentially a pile of shit. Try a different cable that you know works elsewhere ?

Just keep plugging away at it, you'll get there. Someday

I haven't actually tried it by hooking one of the DCD outputs directly to the monitor; it's hooked up as shown in the picture above with the DCD outputs feeding the VBOB and the third VBOB DVI connected to the monitor. Since putting the DCD in I haven't tried using the 13W3 output because, as you and others have mentioned, that doesn't work, and it says as much in the DCD user's guide, despite the fact that Autodesk has had this diagram published on their website for years:



So, possibly this indicates that the 13W3 will still work if the only DCD output being used is the "right" one, but it seems more likely (to me anyway) that the diagram is wrong and they meant to show the left DCD output hooked up to the "graphics monitor," not the 13W3...