The collected works of recondas - Page 10

I download a copy of 4.8.0 from the ftp and expanded it. That much looks ok, but I can't go any farther at the moment - the install requires that a tablet be connected <during installation>, and mine is currently on loan.

A visual comparison of the expanded driver file probably won't be too helpful, but just in case here's shot of the expanded 4.8.0 tarball:

Hello arhiman57 - welcome to nekochan.

That's a pretty nice collection you've got there.

So Thomas W. - is your Onyx2 still floating belly up <or did you figure out what was wrong>?

This really an excellent opportunity for the nekochan membership. Along with the Tezro, the O350 seems to have been one of the hardest SGI systems to acquire.

The O350 can also easily be converted into the functional equivalent of a rack-mount Tezro Workstation with the addition of a Fuel VPro board: the original discussion is here ; a photo how-to of the O350/VPro installation process is here ; and instructions for changing the PROM splash-screen to display as a Tezro are here . It'd be a pretty nice way to end up with a quad processor - VPro equipped workstation.

ramq wrote: The Base Compute Module works standalone just like any standard system. They are fitted with an IO9 adapter card which provides gigabit Ethernet, SCSI for internal disks, IDE for system CD/DVD-ROM and provides NVRAM to keep system enviroments. There's also correct IDE+SCSI cabling for the drives and ofcourse disk sleds. Without the IO9 card the machine cannot work standalone. (These cards are a rare sight and often cost lots of money on eBay)
The MPX Module lacks the IO9 daughtercard, the drive cables and sleds. (Without this it won't work standalone and must be connected to either a Base Compute Module or a NUMAlink router.) Other than that, it's an ordinary Origin brick.

If anyone ends up needing one, I have a spare IO9 .

SAQ wrote: I've not understood the drive to "convert" to a Tezro. The Onyx350 InfinitePerformance will do anything a Tezro will do and maintains the possibility of further expansion later. Is the InfinitePerformance splash screen just really bad looking?

Drive? For whatever reason you're making way too much out of Tezro PROM screen switch. It's there as an option for whoever owns a capable system, to use or not as they see fit. Nothing more, nothing less.

FWIW, I happen to prefer the IP PROM screen - but since I hardly ever reboot, I don't see it enough to get that excited about it.


If any one feels compelled to make the Tezro PROM molehill any larger, please do so in the thread already started on the subject , so we don't drag ramq's FS post too far off topic.

yetanother**ixuser wrote: i would keep it in the building during demolition an let it R.I.P....way to old crap which promises headache only

Feel better now?

As an additional note for anyone looking for a SCSI DVD-ROM that can also be used to boot IRIX install CDs - I recently acquired an HP StorageWorks Model C7499B DVD-ROM. The drive mechanism in the HP StorageWorks Model C7499B is OEM'd by Pioneer - the HP label and firmware identify the mechanism as a <Pioneer> model 305S. The HP7499/Pioneer 305S are rated to read DVDs at 10x, and CDs at 40x.

Unlike the HP C4315 listed above, this one has a faceplate attached. The face plate one this one is black, they appear to also have been released with a beige faceplate <with a slightly different model number - C7499 A >. Either version of the C7499 comes with a proprietary HP mounting bracket/SCA converter that's easily removed <as a bonus you'll end up with a nice 50-to-68 pin internal SCSI converter>.

The C7499B was tested with an Onyx 300 and an Origin 350. When attached to either system, the drive successfully booted an IRIX 6.5 Installation Tools CD-ROM <be sure to install a jumper on the header marked 2048/512>. Once either system was booted, the C7499 also successfully provided audio-over-SCSI. Because of the similarities with the other Pioneer DVD-ROMs listed in this thread, I suspect it also has the ability to boot a user-made DVD installation disk <see jan-jaap's post above for details>.

....and as a nice fringe benefit, this C7499B arrived with an hp-ux 11i version 2 for Itanium 2-based systems DVD inside the drive.

theinonen wrote: Take Russian ejection seat as an example, it works and has actually saved many lives.

A fact that's always nice to know. Is Aeroflot equipping the entire fleet?

You might try booting into single user mode:
stop at the PROM, select option 5, type "single" <no quotes> and press enter.

Once booted, I'd suggest following the IRIX Installation and Customization nekowiki as a template to see if the errors are correctable.

I'd also turn off ESP - "chkconfg esp off" and turn on verbose boot messages with "chkconfig verbose on"<again without the quotes>.

Nihilus wrote: Get a cheap TFT with Sync-On-Green and voilà. It does exist dual SGI/SUN-adapaters, in fact I own one (which actually works ).

I'm glad it worked for you, but your post might inadvertently add confusion to a situation that seems to have resulted in more than enough confusion already.

If yours is one of the all too common generic 13W3 to VGA adapters that some sellers wrongly market as being compatible with either SGI or Sun 13W3 graphics connectors, it probably works because your TFT supports sync-on-green, not because it supports both SGI and Sun 13W3 pin-outs. To be fully SGI/Sun compliant it would need to provide some method of changing the pin-out, either by means configuration jumpers or some other method. If the adapter is hardwired it is probaly not fully compliant with both the SGI and Sun versions of Sync-on-green without the addition of some circuitry.

For an explanation of why it works when connected to your sync-on-green TFT, see dc_v01's posts from earlier in this thread <emphasis added>:

dc_v01 wrote: There is no such thing as an SGI/SUN adapter - it is an SGI adapter or a SUN adapter. The 10 small center pins route sync signals differently between the two. If it is a SUN adapter, it will not work with your monitor using standard sync signals. There is another potential method of syncing, however. The three large coaxial contacts in the 13W3 connector carry R,G,B (red, green, blue) signals. Those are standard between the two connectors, and all SGIs also support sending the sync signals on the green connection, in a format called Sync-On-Green (SOG). If your Sony monitor supports SOG signaling, you should be able to make it work even with a SUN adapter as the RGB lines are the same. However, in some cases the small center pins screw up the signaling anyway - it's possible to then make the adapter function by cutting out the offending pins (IIRC it's pin 9, but if you're doing this (using SOG) it's safe to snip them all and difficult to cut just one anyway). Check monitor support first, or get a for-certain SGI adapter that supports H/V sync. Note that some SGI models cannot disable sending sync signals on green (they're always there), which causes some monitors to have a green tint even when the proper H/V sync signals are routed correctly, so the only guarrantee is to use a SOG monitor. I have had good luck with non-SOG monitors but others aren't so lucky.

Those eBay SUN/SGI adapters are SUN adapters. They will only work with SOG monitors on SGI as I described above . If you wish to confirm, use a multimeter and compare the pinout to those for your workstation in the owner's manual on techpubs.sgi.com . Pay particular attention to pins 3, 4, 5 if IIRC on the 13W3. You will find they don't line up with the pinout of a standard VGA 15 pin connector.

edefault wrote: As mentioned above, this is true in general, but: SGI Indigo2 machines *do* put out Hsync/Vsync as well as RGB/SOG,
so it is possible to connect them to non-SOG monitors without any additional hardware, solely by correct wiring.These worked fine on Eizo FlexScan L985EX monitors, which are - AFAIK - not capable to synchronize on green.

I'm not disputing your assertion - I haven't used an Indigo2 in over 5 years, and then only with SGI-branded 13W3 CRTs.

As a moderator I guess I sometimes tend to see things from a different perspective. The nekochan archives have become a knowledge base for those seeking SGI/IRIX info. I've received more than a few PMs from members with newly acquired SGI hardware, who purchased the wrong 13W3 adapter based on an "Sun/SGI adapter worked for me" post that didn't include an explanation of the circumstances under which it worked <there are numerous "why doesn't this 13w3 adapter work?" posts on nekochan>.

I have no doubt you and Nihilus both have success with 13W3 adapters that are not SGI specific. My intent was to make clear to anyone who didn't already know that the use of an adapter with a Sun 13W3 pin-out requires specific underlying circumstances such as a SoG compliant monitor <or as you pointed out, specific SGI graphics hardware>.

Our collective knowledge on the issue grows every time some one posts how they were able to successfully resolve the 13W3 to VGA issue. I'm not suggesting that either you or Nihilus have, but when that post consists only of a brief statement that implies that any Sun/SGI adapters will work <without some accompanying detail of the specific circumstances>, it doesn't don't do the SGI/IRIX user community any favors in the long run.

Congratulations jj!

nekonoko wrote: Minor update; got a DM5 in recently:

A nice addition to an already very nice system.

The DMediaPro VBoBs will output diagnostic information on the serial port <#10 in your included diagram - it's the only male DB9>. If they don't pass power-on diagnostics <or don't out put anything at all>, you at least have a better idea if there's a hardware fault. If in working condition, they should appear in a gfxinfo -v. See nekonoko's post for an example: viewtopic.php?f=14&t=16718812&p=7329913&#p7329913

Use a serial terminal set at 9600/8/N/1, a null modem cable, then power up the VBoB with the serial terminal connected and running. Here's the output of one with a DM5:

and one for a DMediaPro VBOB with the TOI option board in place of the DM5. The TOI option board has two TMDS ports that connect to the same number of TMDS ports on the DG5-2/TVO (a.k.a. The DG5-2/HD-GVO) used with InfiniteReality3 or InfiniteReality4 graphics systems:

Like I mentioned in another post, more and more sellers are buying bulk surplus for resale. Some of that surplus is likely to be defective or dead hardware that got stuck back into storage by the original owner.

Congratulations - I know ramq hated to part with them, but they couldn't have gone to a better home.

Almost the same set up as my daily desktop <except that I couldn't resist the temptation to have a wannabee Onyx >.

nekonoko wrote: Great, let us know how it goes!

Yes, please do. There isn't a lot available in TechPubs beyond the operational manual, so anything we can add here will probably prove helpful to others as DMediaPro more hardware becomes available/affordable.

jwhat wrote: one of them fails to complete initial boot

If it gets far enough in the boot process to display the "Press ESC to enter the RAM diagnostic mode OR load alternative image" message, I wonder if attempting to load an 'alternative image' might be a possible method for resurrection.

jwhat wrote: Given that 2 of the VBOBs appear to be reporting ok, it might be either my DM2 board or cables that is causing the problem.

Are you using two of the 'black' LVDS cables? As I'm sure you're aware, the DMediaPro manual mentions using a black and a white cable <with different part numbers>, so I wonder if there's a difference that extends beyond the color. Perhaps we could get someone with one of each to check to see if the pin out differs between the white and black cables.


EDIT #1: This post seems to confirm that the only difference in the cables is color: viewtopic.php?f=16&t=16720867& - I've PM'd the OP in the linked thread and asked him to join the discussion.


EDIT #2: As a follow up to one of the jwhat's questions in his original post......

jwhat wrote: Are people aware of compatibility issue with different versions of VBOB and DM2/DM3 boards?

......while doing a search on 'vbob_flash' I ran across this mention that there are four versions of the DMediaPro VBOB that don't appear to be compatible <or at least as far as firmware goes>. It might also suggest caution should be used before choosing <or running> a flash update on a VBOB:

Added a Graphics Compositor to the system.

TechPubs wrote: Product Description -The compositor is a hardware graphics compositor that can receive two, three, or four digital video inputs, and then combine them into a single video output to increase graphics performance. Each input is from one pipe residing in an SGI graphics device such as a Silicon Graphics Prism XG2N module, Silicon Graphics Prism Deskside system (with Image Sync installed), Silicon Graphics UltimateVision Onyx4 G2/G2N, SGI Onyx 350 compute/graphics module, or an SGI 3000 series V-brick. The video output can display on digital and analog monitors at the same time.

General Capabilities - The compositor can perform spatial compositions, pixel averaging, and stereo rendering. The compositor receives two to four input signals and outputs a single signal either in analog or digital format. The following items are noteworthy regarding the compositor's capabilities:

A selectable pixel-averaging feature (antialiasing) combines the inputs from two or four graphics pipes into the compositor and averages these inputs to smooth jagged edges on displayed images.

For every output pixel, the compositor averages all values from all the pipes. Among other things, this provides applications with the means to do full-scene antialiasing (FSAA) in hardware.

A spatial composition/tiling feature enables each pipe connected to the compositor to render a predefined portion of the display.

DB-9 stereo sync port connector connects to an emitter to provide stereo effect for LCD shutter glasses. This stereo function has one pipe providing images for the left and right eyes in alternating framebuffers. You can also select a Stereo mode through the Initial Tiling Mode field so that one graphics pipe provides 24-bit images for the left eye and another pipe provides images for the right eye in alternating frames to enhance viewing performance.

One advanced csync connector and a GENLOCK 1 connector connect with the genlock connectors on the graphics pipes to synchronize the video refresh of the compositor output with the graphics pipes. These connections enable the compositor to receive an external sync signal, which locks the timing of the output video picture.

One swap ready port connector synchronizes the compositor buffer swaps to the graphics pipe buffer swaps.

With InfinitePerformance graphics, there are four swap ready port connectors labeled IP SWAP READY 0, 1, 2, and 3 connect to the corresponding swap ready connector on the InfinitePerformance pipes 0, 1, 2, and 3. These connections synchronize the compositor buffer swaps to the InfinitePerformance graphics pipe buffer swaps.

Stereo is supported only for analog output.
When Xinerama is used to overlap screen regions on an edge-blended display or compositor-based system, the cursor will seem to disappear when it enters the overlapped or uncomposited regions of the display.

By upgrading to IRIX 6.5.20 or later, you can use an X server feature that prevents the cursor from disappearing in these cases. It causes additional cursor images (not real cursors) to appear on all pipes contributing to the overlapped regions. To enable this feature, add the –phantomcursors flag to the X server command line in the /var/X11/xdm/Xservers file. For more information about the –phantomcursors option, see the Xsgi(1) man page.

Now the hunt for another Fuel V12 begins.....

nekonoko wrote: Any pics of the SGC? I've never run across one of those.

I just added a few - one to the original hinv post, and one to the newer post. The Graphics Compositor looks kind of nice when powered up. The SGI logo is illuminated by blue LEDs.

The do appear very similar. I noticed that it identifies itself as a "cbob" <compositor break out box?> in gfxinfo. Here's a photo of the rear panel and the interior:

ramq wrote: You have inconcistent amounts of cache on your CPU-boards. I demand that you immediately send the 8MB boards to me for exchange to standard 4MB variety.

I'll let you know when I get around to packing those up.

ramq wrote: Seriously, that's a very cool piece of hardware.
The possibilites... another two O350 bricks (and NUMAlink brick) with V12 boards and you have 4-to-1 pipe config and a buttload of graphics performance. Yum.

Thanks! Just like most of us here I suffer from the need for speed, so that thought has crossed my mind more than once. Limiting the set up to two modules does offer one <less powerful> fringe benefit. I can run the two-module set up 24 hours a day year round without having to make any special provision for cooling. That size makes for a everyday IRIX desktop with a more than reasonable amount of power. A single two or four processor O350 with a V10 or V12 would probably be an even better balance of those same factors - until I ran across the compositor I was even considering the idea of moving one O350 to run the Onyx 300 G-Brick, and splitting off <or selling> the 16 processor O300 compute set up.

jan-jaap wrote: So hamei's fantasy of using this to connect to some 3840x2400 display isn't going to happen.

That would actually be the reverse of how the Compositor works. The compositor takes the output of <up to four> displays and squeezes them into a single display no larger then any one of the four. So instead of creating more resolution, you assign each V12 to smaller section of a standard display and combine their power. BTW, hamei is definitely a fan of the IBM T221, but he was well aware that the Compositor wouldn't be the device to drive his dream display: viewtopic.php?f=3&t=16723528&

jan-jaap wrote: It will be interesting to seel how well this device works, and how transparent it is from a software perspective. SGI has a long history of graphics pipe multiplexers (all the way back to the Skywriter cyclops mode), but it always required explicit support from the applications.

It will be interesting, and that's a big part of why I collect and use SGI hardware - I like to figure out what makes things tick. There's a mention in the programmer's guide that the Compositor:

transparently distributes rendering across multiple hardware pipes and combines their outputs by either feeding them to a hardware compositor device or through software composition

Whether than means complete transparency for any application or only through custom programming I don't know <and since this a hobby rather than a livelihood, am not really bothered one way or the other>. But I intend to have fun finding out.

smj wrote: I'm new to linked nodes so I'm wondering about hostid changes and how one deals with licenses that no longer match... RTFM PMs and redirects to other threads welcome.

If they have previously been linked they'll already have the same system serial number. If they haven't previously been linked, the pair will assume the serial number of which ever becomes the master module <though you might have to use the "serial clear" command in the L1 of the *unlinked* slave so when it is linked it's ready to receive the serial number from the master>. So examining what licenses <if any> came with each module before selecting the master would be a good idea.

If you have an L2 controller <or emulator>, then the attached modules will assume the serial number assigned to the L2. A very nice fringe benefit of having an L2 controller is that can easily change the L2 serial number at will using the L2 command "serial set <desired_serial_number>". The 'serial set' command doesn't work at the compute module level <with the L1 controller>.

The following excerpts from an L2 "serial all" command example shows the serial numbers assigned to 01c001 <the master compute module>, 001c002 <a linked slave module>. Both also list the <same> L2 "reference system serial number". These particular compute modules didn't come from the same source and originally had different 'local system serial numbers'. One had a full set of MIPSPro licenses, the other didn't. The serial number from the licensed system was assigned to the L2, which in turn wrote the licensed serial number it to the <previously> unlicensed slave.
An L2 controller can be used to set the system serial number of any of the O3x0 modules *except* the Router Module, which has an enforced serial number security routine coded into NVRAM <that prevents end-user system serial number changes> - though there are now a couple of known workarounds for the NUMALink <Router> Module <and the R-Brick>. This would also be a good time to mention that you can't link more than two O3x0 compute modules without using a router - a NUMAlink Module <or a hacked R-Brick>.

"Got a Tezro today" is something I've long wanted to say. Well I can because I did, but it took a little doing to be able to say I had a Tezro that worked.

This one appeared in a Craigslist ad a couple of hundred miles away. I answered immediately but didn't get a reply even after a second post. The next week a link to the ad was posted here, and I mentally wrote the Tezro off. Later the same day I got a response from the seller with his phone number. When I called the seller seemed to be a straight up guy, who told me he didn't know a lot about SGI hardware - he bought the Tezro at an auction. I asked if it ran - his answer was the little display panel says L-something is running. I asked if any fans ran when the power button was pushed the answer was no. Didn't sound too reassuring <there have been a few posts on nekochan about Tezros with dead motherboards>, but the price left a little room for repairs, so I agreed to buy it.

It arrived this afternoon, nicely packed. The case is a little scuffed up, but intact with no broken parts. The insides were filthy. Every fan and ventilation perforation was full of clumps of black dust - looked like it had run non-stop for several years at floor level in a place that under-paid the janitorial staff.

Cleaned out the worst of it <outdoors>, then carried it to the work bench and inspected the interior to see that everything appeared to be seated. Connected a serial terminal and applied power. The fans didn't start, and
appeared in the serial terminal console. That was it, couldn't get a response from the L1 or the PROM. A few thoughts came to mind, all of them four letters long.

I went back over the system, and this time noticed that there was a slight bit of deformation to the VHDCI SCSI port on the back of the IO9, and a the bulkhead next to the SCI port was was slightly bent. I pulled the IO9 and with the help of a magnifying glass noticed that the impact that had cause the deformation to the SCSI port compressed the array of solder pins on the back of the port so that some were broken and others were in contact. I had a spare IO9 on hand, so I stuck in the replacement and reconnected power. This time I got an L1 prompt. an L1 'env' and 'serial all' looked ok, so I powered the system up and switched the L1 to console mode.

Got 'widget 15 present but not alive during' the power on diagnostics, no display on a connected LCD monitor, and a 'gfx config error' on the L1 display - even though the PROM was set to console=g and a keyboard and mouse were connected. Stopped in the PROM and ran an hinv, no graphics board appeared.

The graphics board was fully seated, so I pulled it and inserted a spare Fuel V10 I had around. This time it went through power on diagnostics without and error, and the Tezro splash screen appeared on the LCD monitor and in a PROM hinv.

The Tezro came without any drives or sleds, so I used a clone of the root drive from my O350/V12 and was able to boot IRIX. Here's an hinv <with the V10 test board still installed>:


Update: I replaced the V12 and posted an hinv and some photos the hinv forum.

A few photos - some have said they don't care for the tower Tezro's case design, and I wonder if that partially due to the fact most photos are taken head-on. Now that I've seen one in person, I don't think most photos accurately depict the visual strength of the design. I found the Tezro's most striking feature to be it's proportions - particularly the less often seen profile:

BigD wrote: #@!$U*$#@%&!!!!

I hear ya, but you're still over there right in the middle of the land of SGI milk and honey. From the stand point of bang for the buck/amount of high-end gear included, that Onyx 350 IR4 over your way is probably the best SGI hardware deal I've seen in a long time.

PymbleSoftware wrote: Congratulations, I too, am envious.

Thanks - what I left out of my post was right after I found out it was dead enough not to bring up an L1 prompt, I had to go and sit through several hours of a small-talk social event. Things were slow enough that I had time to mentally run through every possible gloom-n-doom scenario at least four or five times - regular life of the party I was.

smj wrote: I would've had to go hunting the elusive IO9...

Yeah, that could have been real ugly - check out the $500 price quotes in this IO9 wanted thread . I got lucky and grabbed one last year when one of the eBay sellers was running an SGI firesale. Didn't really need it at the time, but it was going down the second time without a bid.

What was worrying me before I figured out it was the IO9 was the initial VRM error messages. I mailed my last spare node board <with VRMs attached> to Scotland last week.

nekonoko wrote: Wow, nice machine! Hopefully you can get the graphics sorted

Thanks! I stuck a Fuel V12 in it for now. The original Tezro v12 wasn't stone cold dead, it would appear in an L1 serial all. I'll go over it with a magnifying light to see if maybe I can find a loose solder joint.

nekonoko wrote: Yeah, my Tezro was missing its IO9 completely.

I suspect this one originally had a DM3 in it - the cover plate was missing from the XIO2 slot for the DM3.

jan-jaap wrote: Since you have both, maybe you can tell me: how loud is a Tezro compared to an Origin/Onyx 350 module?

Been a while since I ran a single O350, and the O350s are now in an SGI short rack which really muffles the noise <the short rack is built out of double-wall 3/16" steel - even the rear door is 3/16" steel>. But going solely from memory I'd say the sound level of the Tezro is similar to one O350, but maybe slightly lower in pitch.

I do have the Tezro and a Fuel running side by side. If you stand to the rear of the two the sound levels are pretty close. From the front or sides the Tezro is noticeably quieter. While the Tezro has many more fans, they seem to run at lower speed <the heavier construction of the Tezro probably plays into that too>.

Looks like the 800MHZ nodes generate a little more heat. After a few hours of run time I noticed the temps reported by an 'L1 env' were several degrees C warmer than these nekonoko posted for his Tezro.

To make the comparison a little more balanced I ran the 800MHz Tezro overnight with the same amount of RAM and number of PCI boards as nekonoko's Tezro <though I'm clearly short of whatever heat nekonoko's DM3 generates>. Here's the fan and temperature portion of nekonoko's Tezro:

and the same section of the 800MHz Tezro <with approximately 8-hours of run time>:

If the temperature increases were anywhere near proportional to the speed increases of the R16 processor, it's easy to see why SGI more than doubled the size of the heatsinks on the 1GHz IP53 node boards <though I'm sure adding another 12MB of processor scache played into that too>.

kshuff wrote:

#@!$U*$#@%&!!!!

My thoughts exactly

I'm not a fan of posting third-party auctions or sales of SGI gear on nekochan - especially when that item has the potential to be a very good deal. Unless you're the seller of the item, I can't see the logic in posting what could or might have been a good deal on the one place that will almost certainly ensure *no one* gets a good deal.

Last time I tried to buy a Tezro, pretty much the same thing happened. Got the first bid in, and the all was quiet for several days until some kind soul decided to post a link <see "Saw a tezro on ebay, ready to be snatched away" for the gory details>. Might that earlier Tezro have still ended up beyond my budget? Sure it *might* have, but posting it here removed any doubt.

The Tezro that's the subject of this thread was listed on Craigslist on November 10th. I routinely search *all* of Craigslist, and found the listing the same day, and emailed the seller asking for details. Didn't get an immediate reply, so I emailed him again on the 12th, and again didn't get an immediate reply <I don't think he gets in a hurry, it was six days before that Tezro shipped>.

On November 16th, a link to the Craigslist ad was posted on nekochan. I figured that post pretty much let the air out of the idea that *anybody* would get a Tezro for a hobbyist friendly price. The same day as the nekochan post, the Craigslist seller sent me a reply with a phone number, and I called him. During our conversation he mentioned he was rather surprised that the listing had gone from 'almost dead' to receiving 12 or 15 email inquiries that day.

nekochan's membership is probably the single largest block of SGI hardware purchasers in the world, and at the same time is a community who's members have formed a social bond. Mixing those seeming opposing elements with everyone's desire to get that once-in-a-lifetime deal doesn't always produce the best results. Keith's obvious disappointment is just another reason why I'll renew my suggestion that members refrain from posting what could be a 'good deal' for SGI hardware on this list.