The collected works of jan-jaap - Page 23

For the record: a Ghostery update fixed this.

dexter1 wrote: Oh, and steer clear of R12K 270MHz, these are particularly problematic.

Care to elaborate? I have one of those but it never gave me any grief. (Well, the noise level could be called problematic, but I only use it as an install server).

diegel wrote: Many bus errors are results of MipsPro aggressive optimization

A bus error in code which "works" on x86 is usually caused by unaligned memory access . Compiler optimization may expose it but the root of the problem is a flawed assumption by the coder. x86 CPUs will usually fix up unaligned access (at the cost of performance), but RISC CPUs will abort with a SIGBUS.

A stack trace or a SIGBUS handler should tell you where the offending access is happening. But since the cause of the problem is in the poor design of the code, the fix is usually not a oneliner.

hamei wrote: I have an Octane Compression in the lower left slot (looking from the rear.) If I put it in the lower right slot, it gets noisy as heck. Otherwise, it runs at Normal Octane Level.

Correct.

Now, see where a VPro card has it's XIO compression connector (on the PCB side where the dual DVI option goes):

A single VPro uses the space of both upper XIO slots, but attaches only to the top-right (looking from the back side).

And see how on the dual carrier the bottom card is turned 180 degrees:

Looks like it attaches to the bottom-left XIO site (looking from the back).

So, technically you should be safe from the infernal fastfan mode, unless the firmware is clever enough to know that the dual VPro setup is utilizing the bottom right slot space and engages fastfan anyway. Maybe ask 'mopar5150', he had one for sale recently.

robespierre wrote: Techpubs seems to have already removed its IRIX 6.2 and 5.3 section. Are those backed up somewhere?

Removed from the index page, but content is still there: IRIX 5.3 , IRIX 6.2 , IRIX 6.3 and IRIX 6.4 .

I have backups of the manual PDFs and tar.gz files, but not the manpages. But if you wanted to ressurrect that, you would use the actual manpages from an IRIX dist + some man2html gateway.

ClassicHasClass wrote: I'm here, but I have no life.

I have a life and it comes at the expensive of time spent on a computer hobby

Also most of my hobby time is currently spent on migrating my home server to a new system.

I've got a 013-2478-001 rev A here and that's a full installation set ( ~ 15 discs?) of IRIX 6.5. It may have come with 6.5.1 overlays originally as well, I have/had quite a few of those.

Hmm, well, my box has the sticker on the box. Otherwise it would have been gone indeed

It's companion box is (in my case at least) 013-2522-002 rev A, the Varsity update for 6.5, with the 7.2.1 compilers and Performer 2.2

The problem with the adapters is that almost all of them are for SUN systems. But SGI adapters do exist, and so do 13W3 - VGA cables.

The more subtle problem is that there's no such thing as the SGI 13W3 pinout, there are at least four of them. You found the Indy/Indigo2 version and the Octane version, but did you know a Reality Engine put out +10V on pin 5? If you wire that to the Vsync of the monitor like your adapter does you may very well release some magic smoke here or there.

Crimson Reality Engine, Onyx RE2:


Onyx1 Infinite Reality:


Onyx1 IR owners better have a fully SOG compatible monitor because nothing else is wired

Seems the end is near this time:

dexter1 wrote: Jan-Jaap, how big is the current techpubs IRIX stuff?

I have:


If manuals are available as PDF and .tar.gz, I have both. Some manuals show up under more than one category (and are hard linked on my server) so total disk space is a fraction less.

Maybe bitsavers can play a role here? I guess I could handle the data volume but I have some doubts about making some 3800+ PDFs with ominous copyright statements available on my FTP server. They have more experience with that.

NB: newest additions, last half year or so:

MooglyGuy wrote: Sounds good! Here's an updated PROM list:

Code: Select all

BOOT PROMS
Personal Iris 4D/20:    Version 4D1-3.1 PROM IP6 OPT Fri Nov 18 14:56:48 PST 1988 SGI
Personal Iris 4D/35:    SGI Version 3.3.2 Rev A IP12,  Mar 13, 1991
Indigo R3K:             SGI Version 4.0.1 Rev C LG1/GR2,  Jul  9, 1992
Indigo R3K:             SGI Version 4.0.1 Rev D LG1/GR2,  Mar 24, 1992
Indigo R4K:             SGI Version 4.0.5D Rev A IP20,  Aug 19, 1992
Indigo 2:               SGI Version 5.1 Rev B IP22 Sep 16, 1993
Indigo 2:               SGI Version 5.3 Rev C IP22 Oct 20, 1994
Indigo 2:               SGI Version 5.3 Rev E IP22 Sep 28, 1995
Indigo 2:               SGI Version 5.3 Rev E IP22 Jan 29, 1996
Indy R4K:               SGI Version 5.0 Rev B6 IP24 Sep 28, 1994
Indy R4K:               SGI Version 5.3 Rev B7 R4X00 IP24 Feb 16, 1995
Indy R5K:               SGI Version 5.3 Rev B10 R4X00/R5000 IP24 Feb 12, 1996
Power Indigo 2:         SGI Version 6.0.1 Rev A IP26 Dec  9, 1994
Power Indigo 2:         SGI Version 6.0.1 Rev A IP26 Dec 23, 1994
Indigo 2 Impact:        SGI Version 6.2 Rev A IP28 Jun 10, 1996
Indigo 2 Impact:        SGI Version 6.2 Rev A IP28 Aug 26, 1996
O2:                     Tue Oct 22 10:58:00 PDT 2002 VERSION 4.18 O2 R5K/R7K/R10K/R12K IRIX 6.5.x

IP32prom IP32PROM-v4
Octane A ID1:           SGI Version 6.5 Rev 4.9 IP30 May 22, 2003
Octane A ID2:           SGI Version 6.5 Rev 4.16 IP30 May 22, 2003
Octane B ID1:           SGI Version 6.5 Rev 4.9 IP30 May 22, 2003
Octane B ID2:           SGI Version 6.5 Rev 4.16 IP30 May 22, 2003
Octane A & B are weird in that they have the same IDs, yet a binary compare yields many differences.
Origin 2000:            SGI Version 6.156  built 11:27:56 AM Nov 18, 2003

CPU PROMS
IP22 R4400 150MHz (256 bytes)
IP22 R4400 250MHz (128 bytes)


Additionally, I've got an R10k Onyx that I'm dumping the PROM of by traditional means. If any of you have an SGI machine whose PROM revision differs from the above listed, let me know, and I can tell you how to dump the PROM over a serial connection.

By the way, krafty, did you ever happen to dump the PROM off of that Crimson of yours? If not, perhaps the next time I'm down to West Palm Beach we could uncover the beauty and I could borrow one of the PROMs off of it for a friend to dump. Of course, that'll have to wait until you're done fixing everything up.

Does anybody have a copy of these? I'm specifically looking for a PROM chip or dump for an R4400 IMPACT Indigo2. I've got this nice 250MHz R4400 CPU staring at me but I don't have the required PROM chip (September '95 or newer).

I've got a burner.

I work for a company which specializes in military and aerospace applications of IEEE-1394 (FireWire). IRIX is one of our supported platforms.

ivelegacy wrote: Can I ask why

Yeah, but then I'd have to kill you

Good to hear it's only a loving wife. For a moment it sounded like a terminal disease...

Nice Octane and a beautiful photo as well. Tripod & long exposure?

kokoboi wrote: I'm making a complete mirror of techpubs website, so far I downloaded 22GB

Did you simply point 'wget' at it and are you downloading endless amounts of cgi generated html?
Best I can tell there's about 6.1GB worth of unique PDFs and .tar.gz files on Techpubs (some files are listed in multiple sections).

NB: latest and only addition since late November was this one:

on December 22nd. In the category 'Linux'.

I've got 2553 unique pdf files and 1339 tgz files.
My scripts marked 12 pdfs and 10 tgz files as orphaned, which means I have a local copy but they are no longer on TechPubs.

jeffe23 wrote: This is a circa 1997 Octane prototype

Cool. Is it transparent or green?

First of all: congrats with a very nice catch, MrBill!

miod wrote: I don't have room for such a machine and this wouldn't be a reasonable decision.

Buying a Crimson has nothing to do with reasonable decisions, it's a matter falling victim to incurable red fever .

Been there, done that , didn't regret it

miod wrote:

jan-jaap wrote: Buying a Crimson has nothing to do with reasonable decisions, it's a matter falling victim to incurable red fever .

Yes, but I already have the unique Octane Crimson , remember?

I think Dr. Dave's .sig applies here:

Dr. Dave wrote: Once you step up to the big iron, you learn all about physics, electrical standards, and first aid - usually all in the same day

Devil Master wrote: Can't find any picture of the rear of an Onyx2, though...

http://www.sgistuff.net/hardware/system ... k-2200.jpg

look at the rightmost of the "big" cards, it is the DG-5/2 display generator. The DB9 in the bottom half between the BNC connectors is the stereo output connector.

ivelegacy wrote: my girl friend said
- chaos is not allowed, things must be in order!
- things here (especially things in the living room) must have a purpose

She is right, though I cannot imagine life with you is ever without chaos, even with all objects in their proper place

ivelegacy wrote: so, can I use Lord Crimson as "wine refrigerator" ?

Yes .

Be warned that if you should choose to do this, I will come over, quarter you , and kill you. In that order

ivelegacy wrote:
DB9M RJ45 kit

I bought a couple of dozen of these (both the male and female gender) a couple of years ago. I have a ton of equipment with serial management ports, and a Cyclades ACS 48 console server (48x serial port on RJ45). Whenever a new weirdo piece of kit shows up at my place I make an adapter. You just have to make sure you label these things properly or you will forget what system they go with.

Posting this from my 13" MacBook Pro, model "early 2011". I upgraded the RAM and replaced the disk with an SSD and it still works just fine for me.

recondas wrote: .....the safe bet would be to obtain the 500MHz PIMM with the original nodeboard still attached.

Absolutely. Every time you mess with these things you risk damaging it, and 500MHz nodes are hard to find.

That said, I have a 4x 500MHz R14K Onyx2 deskside and one of the nodes failed at some point. Fortunately for me, it was the nodeboard and not the HIMM so it used an IP31 with 250 or 300MHz CPUs as a donor. Worked just fine.

Beware that some late model option(s) (I think HD video IO) require a nodeboard with a fairly late rev. XBow on the IP31, which in practical terms means CPUs >= 400MHz. So to be on the safe side, use a 400MHz as a donor if you must buy only the HIMM.

It used to be on the SGI "cool software" pages (can't find it on archive.org ). Which didn't even require a login. I faithfully downloaded all that stuff:

You can even run another version of IRIX in your chroot. There are some limitations (IRIX 6.2 pthreads userland doesn't work on IRIX 6.5 kernel for example), but it's quite useful for building software etc.:


Pretty sure IRIX 5.3 never ran on IP35

ivelegacy wrote: the real problem is: what do you run on them ?

Duh. FSN of course

ivelegacy wrote: where have you fount that middle SCSI UW tower ? I have been locking for something similar since a while

It found me Sometimes when you buy a system you get the software and peripherals as well and that's how I got this thing.

necron2600 wrote: Also, a hinv in the POST command line only shows 1 graphics card.
And the ODY32 boards have the same part numbers.

Anyone have any ideas? I am single CPU.. maybe SGI really meant it when they said it requires dual cpu for multi-head? Is a revision of one of my components too old? (Xbow is at 1.3?)

It's the fact that only one ODY board shows up which is a little strange. I suspect a dirty or damaged compression connector, or the PCB surface pads the compression connector mates to are dirty. If you change the VPro boards around, does it still see the same VPro board? If so, then you've got one VPro board with a dirty/broken compression connector. If not, then I'd suspect the frontplane. The frontplane PCB is often covered with a cake of fine black dust, built up over the years, and unless there was something in that slot before, it's also on the connector pads where your new card goes. I've seen this a couple of times when fitting a PCI cage to an Octane. To fix this I removed the frontplane and cleaned the connector pads with a cotton swab with some 97% alcohol.

To understand the requirement for dual CPUs, you have to understand the product placement back in the day: the dual head option was part of a "dual seat" option where two people would share a single Octane, but each have their own graphics head, keyboard and mouse (this requires a CADduo card). You can put up to 3 MGRAS graphics boards in an Octane, but I have never heard of the triple CPU option

FWIW: If needed, you can use a compression connector from any useless XIO option (FC, HVD SCSI, ...) to salvage another board, like a VPro graphics card.

I never measured the consumption of this one, but I remember that my other Onyx2 (4x 500MHz R14K, full RAM, IR3 w./ dual RM10's, couple of XIO and PCI options) consumes close to 1000W when running a demo.

That's probably roughly the thermal budget of a system that size without aggressive 'wind tunnel' cooling which would disqualify it from use in an office environment. A 4D PowerSeries with dual CPU cards and VGX with dual RMs uses roughly the same, an R10000 Onyx1 IR a little more, but if that single, large blower in the bottom of the thing kicks into fast mode you want to be far, far away ...

GL1zdA wrote: Indigo Family Product - Technical Report (incomplete)

Also to be found here: https://web.archive.org/web/20060107172 ... lanTR.html

And I happened to mirror that page many years ago, so here are the missing pictures:

ivelegacy wrote: I wonder what, in the Irix stuff, one has to run on a Onyx2 of 700 watt: Ansys, Maya ?

Scientific workloads, visualization, VR, ... And don't forget the various Discreet titles, but you'll need a lot of additional hardware for that.

Basically, if it's too big for a workstation, you move up the food chain.

ivelegacy wrote: oh, and I also wonder how the CD-drive in the picture can work when its feets are standing up vertically

No problem, there are little catches in the CDROM tray which prevent the disc from falling out. Also quite handy if you want to place an Indigo2 on it's feet.

ivelegacy wrote: Ethernet over fiber: is that possible ?

Of course .

ivelegacy wrote: I was asked to swap a few development boards for a pair (qty=2) of QLogic QLA 2342
they are PCI/64 2Gbps@133Mhz Fibre Channel cards, I sill need to find cables

QLA2342 is not an ethernet board, it's a Fibre Channel host adapter. It is good for attaching storage to an IRIX system, either directly (FC-AL) or using a SAN fabric.

It's possible to run TCP/IP over transport layers other than ethernet (Firewire comes to mind), but the software stack has to implement this. TCP/IP over FC is not possible with the QLogic FC adapters, at least not with IRIX.

That's a neat display!

jodys wrote: Then I bought a 3Com CoreBuilder 3500 which had fast ethernet (fiber and copper), gigabit fiber, and FDDI. That was a pretty neat convergence.

The downside of core switches is they're big, loud and power hungry animals.

That's why I'm using a relatively modest Cisco WS-C1400 FDDI concentrator, with SAS (single attach) connections to the SGI workstations, and a DAS (dual attach) uplink to the "main ring". In my case, the only other station on the "main ring" is a little Linux router box which routes between ethernet and FDDI network segments (FDDI uses a different, larger MTU than ethernet so bridging isn't trivial). In the past my Linux server used to do this, but my current server only has PCIe slots so I turned the FDDI router into a dedicated appliance using and old mini-ITX atom board.

skywriter wrote: At the time I built out my network, everyone was dumping FDDI. I could get FDDI adapters for everything {Sun and SGI}. [...] I could connect everything demo 4D's Indy's, Indigo's, O2, Octanes and origin & Onyx's to at least 100baseT next works.

I think a lot of that (and the guts of a lot of 4D hardware) ended up at my place at some point

robespierre wrote: Did you ever see equipment for FDDI-II?

Nope, I only ever saw one or two whitepapers.

What I have is fairly standard. 4D series with VME FDDI cards (which are modified Interphase V/FDDI 4211 design), GIO32 and GIO64 FDDIXpress cards for Indigo, Indy and Indigo2, and a HIO64 dual FDDI card which mounts on the IO4 of a Challenge/Onyx. Basically everything before 100baseT was standard. There's a PCI FDDIXpress card in my Origin 200, and I have a couple of SysKonnect PCI FDDI cards for Linux systems. There's one in that router box in that last photo.

Everything I have is optical. No CDDI here.

Disc images appear to be straight 'dd' style images.

If they manage to keep this online I've got a couple of hundred more for them

OK, I downloaded a couple of disc images. Most appear to be good, but it appears some of them are truncated.

For example, "O2 Demos 1.1.1 for IRIX 6.3 Including R10000" is a 225MB disc image. It has an SGI disk label (volume header) and an EFS partition. In SGI disk labels, partition #10 is always the entire disc. You can use this to sanity check the disc image length (on Linux):

will return the content length of the disc image. The actual disc image file should be at least this many bytes (most are slightly bigger, this seems to be a side effect of how they made these discs back in the day). For the O2 demos, that's some 615MB so this image is well short. An attempt to loop mount it and copy the files will result in I/O errors.

The same applies to:
Cosmo Software May 1996 for IRIX 5.3, 6.2, 6.3 and 6.4
IRIX 6.4 RECOMMENDED REQUIRED PATCHES 4/98
O2 INSERT FIRST

I downloaded only a handful of images -- I had most already. It makes me want to sanity check the contents of the images for CRC errors before I add whatever I got here to my collection ...