The collected works of kramlq - Page 3

Sent you PM about an Indigo2 and O2.

ClassicHasClass wrote: Too bad you can't tell from the screen if it's just a prop or actually running. But we hope Tim asked for The Real Thing.

... from NeXT specialists http://www.blackholeinc.com/

Tim Berners Lee using a NeXT we integrated for him at the 2012 opening ceremonies of the Olympic Games!

hamei wrote:

NeXT specialists http://www.blackholeinc.com/ wrote: Tim Berners Lee using a NeXT we integrated for him ...

What, they bussed it to London ?

Its great attention to detail I think. As mentioned by others, most of the people watching probably don't even know the significance of the person who is at the computer, and I'd say 99% don't realise the significance of having him sitting at that particular type of computer. Yet they still seem to have gone to the effort of sourcing a NeXT to use.

They were fairly popular in academia. The platform hardware was relatively simple, and well documented (ftp links to the official docs are available here for anyone curious: http://www.linux-mips.org/wiki/DECstation ). And DEC didn't make it too difficult for Universities to get official Ultrix source code access. So it featured in a lot of OS research/development, networking (e.g. network stack research), OS benchmarking, and open source OS porting projects. The original MIT Exokernel was developed for the DECstation; it was also a major platform for CMU Mach, and even used for early Windows NT MIPS work.

I'm a little clueless when it comes to electronics - what exactly happens when they do get really bad... doesn't power on? fried motherboard and components? fire?

vishnu wrote: I've posted this before but from the capacitor plague page at wikipedia:

wikipedia wrote: As an electrolytic capacitor ages, its capacitance usually decreases and its equivalent series resistance (ESR) usually increases. The capacitance may abnormally degrade to as low as 4% of the original value, as opposed to an expected 50% capacity degradation over the normal life span of the component. When this happens, the capacitors no longer adequately serve their purpose of filtering the direct current voltages on the motherboard, and a result of this failure is an increase in the ripple voltage that the capacitor is supposed to filter out. This results in a system instability. Capacitors with high ESR and low capacitance can make power supplies malfunction, sometimes causing further circuit damage. In computers, CPU core voltage or other system voltages may fluctuate or go out of range, possibly with an increase in CPU temperature as the core voltage rises.

Thanks for the info guys... must check my SPARCstations for this capacitor problem.

ClassicHasClass wrote: My Solbourne S3000 needs Perl, but Perl 5 will not build with the built-in ancient K&R-only cc. Does someone have a binary around of this? I don't need the whole distribution, just the main perl binary for SunOS 4.1 (to be specific, OS/MP 4.1C, which is binary-compatible with SunOS 4.1.2 and 4.1.3).

Any version of Perl 5 accepted, 5.00504 or later preferred.

I do have gcc around here somewhere, but I don't want to install that whole pig just to build Perl if I can avoid it.

This site has a SunOS binary for Perl dated from 1996 (so 5.003 or earlier)
ftp://zealand.frontline-pcb.com/pub/dnl ... /perl_bin/

skywriter wrote: Fellow Emeritus for EMC Corp. Retired at 53 after 26 years. Work mostly on Symmetrix/DMX/VMX the last 5 years. HW design, System Architect, Chip/board designer, 48 patents. Even got my picture in the lobby, heh!

I liked the ASIC design stuff the best. I did the first ASIC for Symmetrix, a Dual port BUS and TAG controller. My favorite design was a 3 ASIC chipset that totally 256 chips on the largest sized system. It was a fault tolerant shared memory controller for 1st gen synch DRAM with chip kill. Although the basic control logic and datapath wasn't that complicated, trying to design chip level systems with a 'pair and spare' architecture with 'totally self checking' check logic, using logic synthesis and a layout tool with logic resynthesis that loved to remove all the redundancy was a lotta fun. I never worked on a chip 'as a team'. I did everything soup to nutz. As long as they threw everyone else on simulation I was happy.

I didn't want to retire late in life with my hearing and vision impaired. I planned on working on music and astronomy in my retirement. At 53, I'm in pretty good shape for that

Some heavy duty stuff there Did you worry you might eventually miss working on complex stuff like that... the challenges?

For the topics it set out to cover at the time, its a really excellent book. Obviously things have moved on, in terms of system architecture, processor architecture and kernel design. Many of the CPUs covered are gone, and some of the cache designs are no longer that relevant, but it still provides an excellent grounding in the type of issues that must be considered. I think you would need to have a fairly decent knowledge of kernel internals and system architecture to understand it all.

Some other interesting books that should be available very cheaply nowadays (again, the content gets dated, but you still get some interesting perspective on things):
- Unix Internals: The New Frontiers by Uresh Vahalia

- See MIPS Run - 1st and 2nd Editions. by Dominic Sweetman - Obviously highly MIPS specific, but discusses a lot of very low level kernel implementation stuff you don't normally see in other books. The second book is Linux orientated, but the first book is really good as well.

- Operating Systems with Linux by John O'Gorman - OS theory illustrated by discussion of data structures/code excerpts from Linux kernel source. The Linux kernel version covered is maybe 15 years old, but the source code is probably still a useful way to illustrate the fundamental ideas in practice (just not a guide to how the modern Linux kernel works). Disclaimer: This book is basically a packaged version of the lecture notes I used as a student (though the version we used was about Digital UNIX).

PS: You should be able to pick up a copy of Schimmel's book for way less than what you mentioned. Mine was very heavily used, so I bought a replacement copy in 2009 or so for €7 including postage I think.

Thanks!

krfkeith wrote: Since you're clearly teetering on the edge of functional illiteracy, I will try and put this simply: no, you didn't.

Relax. He posed some "devils advocate" questions and talking points, which in some circles might have led to some intelligent or interesting debate. Not many of us have Prisms, so at worst any properly thought out answer by TeamBlackFox or any Prism owners was likely to have informed us all a little more about these systems. That is the reason I opened and read this thread - like most other posters, I don't have any Prism to sell.

Since you are the expert on literacy, the question mark at the end of this text should have been your first clue he posed a question.

ivelegacy wrote: but it's Itanium inside, what about Applications ?

Not everyone here is a native English speaker you know.

Hi,
I am interested in upgrading to a faster Compaq/HP UNIX workstation, so if anybody has one of the following machines they no longer use or want, let me know the exact specs and how much:
- Compaq AlphaStation XP1000 with 667MHz CPU
- HP C8000 with Dual Core 1.1GHz CPU(s)
- HP zx2000 Itanium

I'm located in Ireland, but I'm used to buying machines from overseas... EU countries with reasonable shipping costs preferred obviously.

gocram wrote: The zx2000 is a fairly rare and wanted system, in its niche of course. It's definitely a very nice system, as a rather quiet and energy-efficient one (compared to other systems in its class), which would accommodate a large number of platforms and fairly recent and up-to-date ones, too, I might add. VMS guru Stephen Hoffman also proclaimed that he considers the zx2000 one of the nicest VMS (I64, needless to say) workstations you could lay your hands onto.

Even up to two or three years ago, the few people who owned zx2000s were only interested in offering them for rather large sums and, specification-wise, uneven trades... So, it will be tricky to find one, I think. But, maybe you'll get lucky.

I would keep an eye on auction sites like eBay, which is what you're probably already doing. I assume when VMS V9 comes along, the zx2000s will begin to appear and for more sensible sums...

Hi, thanks for the heads up. I actually haven't really been actively looking too much in the past. I currently have a PWS 433 workstation for Tru64 and a C3750 for HP-UX, and I just got the idea recently it would be nice to upgrade. I didn't realise the later/final generation workstations are that hard to find for the hobbyist.

I saw the odd system offered for crazy money as Buy-It-Now on eBay, but I thought that was just the usual eBay thing - resellers/recyclers trying to hit the jackpot. I've also seen Sun Ultra 1s offered on there for four figure sums, but whether anyone of sound mind would actually buy one for that is another matter Perhaps I'm not in tune with the true commercial value of these things though.

@ivelegacy You were probably asking gocram, but nevertheless... my interest was due to the fact that it is pretty much the end-of-the line for HP-UX on a true workstation form-factor system. The zx6000 seems to be essentially a rackmount system that's not in a rack.

DECwest had some interesting technology in development (PRISM, Mica etc.) that sadly seemed to get canned for business/political rather than technical reasons. In one sense, part of the designs later made it into Alpha and Windows NT at least. Did you ever get to see any of that? Or did it even progress much beyond design documents (which someone has kindly scanned and uploaded to bitsavers a few years ago)?

cesss wrote: Are there any known details on programming the ASICs responsible for managing the CPUs in multiprocessor SGIs? I mean, details from an OS-development point of view (ie: ASIC commands needed to create a new process on a certain CPU, or for scheduling each CPU, etc...).

I don't mind if the details are for Challenge, for the Octane, or for NUMA or later MP boxes.

The purpose is that I feel curious about the protocol of commands used by SGI for controlling MP systems. I'm beginning a pet project which needs (simple) MP and maybe I could get some inspiration from older SGI designs...

If you are interested in experimenting a bit with multiprocessors, inter-processor interrupts etc. then Stanford's SimOS/MIPS may also be of interest. Its related to the FLASH architecture ( http://mprc.pku.edu.cn/mentors/training ... kuskin.pdf ) which many Stanford DASH people (and future VMware people) were involved in. Its a partial implementation though and probably about 50% of its commands and registers are no-ops. Documentation isn't great, but you do however have the source code to see how the 'hardware' works, making it possible to understand enough to do CPU startup/shutdown, interrupt setup, interprocessor interrupts, timers etc, and it doesn't take a lot of code - maybe 10% of what an x86/Intel MPS multiprocessor from that same era would have taken. If learning basic SMP operations is your goal, its probably a good starting point. For example, it would be a good exercise to try to port an existing OS like Linux to it.

Linux SMP isn't that hard to understand (though admittedly I haven't kept up with it for many years now, and it tends to grow in complexity each year). Even though they were obsolete, I found it easier to look at simpler early SMP architectures first (e.g. SPARC32, Alpha) - PROM and hardware handled many of the really low level details and so you can concentrate on understanding how stuff like how tracking and managing CPUs, IPIs, cache operations work. I certainly think diving straight into something like a high end NUMA architecture is going to be a more difficult approach.

Also, I'd recommend you get a really good grounding in cache/TLB coherency, atomic locking, memory ordering, and interrupts on multiprocessor systems first. Otherwise you will probably have to redesign all your code when it doesn't work as you expect on real hardware

I shouldn't be posting this in the SGI section I guess, but as we're on the topic of cringeworthy, here HP's equivalent:

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

Did you retry any downloads again just to be sure? I downloaded about 60 scanned PDFs of old DEC manuals from archive.org recently, but I'd say about 15 were truncated. On a second attempt at the bad files, some were ok, but some were truncated again. So it could be due to bad downloads at that site.

Seeing as the thread is already resurrected.... if any of the original posters are reading, there is a short paper on the Indigo2 design, but mostly from a thermal/airflow perspective:
http://s3.mentor.com/public_documents/s ... tation.pdf

surrealdeal wrote:

Shiunbird wrote: I'm itching. =)))))))))))

just keep chanting "it's just another x86 box" and apply calamine lotion

I have one, and as x86 goes its a really unique system, so I'll probably never sell it. But it really is love-hate. Its the most frustrating X86 and possibly even the most frustrating system you will ever own. Putting together a maxed out configuration (2 x 1GHz, 1Gb, FPA, DVD, Zip100, 2 x SCSI HDD), I went through 4 motherboards, about 7 ram kits, 3 1GHz CPUs, 2 VRMs, several PSUs etc. trying to find a combination of components that would play nicely together.

And then when you have it working, you still have several strange hardware bugs - plug in an exotic USB or firewire device and it can reset the system. the graphics driver or hardware has occasional consistency issues, the mouse sometimes stutters while certain USB devices are plugged in, it doesn't like some USB-PS2 adapters, network throughput is not what it should be. Its also strange to have an X86 box but only have about three OS choices - really old windows (NT4, Win2k) or really old Linux. The various SGI patches/updates are essential, and trying to keep a newly installed system stable long enough to install the patches via a network share was an interesting challenge. Having said all that, if you work within its limitations - Windows 2000, fully patched, USB keyboard and mouse, no fancy USB or firewire peripherals, it does the job fine.

If you want to own it because its a very unique X86 design (in terms of platform architecture and outer case), and also an SGI, by all means get it. But if you just want an everyday x86 machine, get a Dell or whatever.

smj wrote: Nuts. Oh well - thanks for the feedback.

Just be thankful the vintage hardware gods have blessed you with a mk1 Omron. Some of us have none

spiroyster wrote:

uunix wrote: What is this? http://www.ebay.co.uk/itm/SGI-550-Workstation-Arrows-F1-livery-fully-working-/122181014777?hash=item1c728e54f9:g:4h8AAOSwZJBX~-t1
Just a bit of cosmetic dust?

I wonder if that paint job was done by a fan, or this hardware was actually used by the Arrows F1 team?

Incidentally didn't Arrows crash out of F1 (no pun intended) about 2002

Its almost certainly a real ex-Arrows machine and not a fan's artistic tribute. There was an ex-Arrows SGI VW320 available on sgidepot many years ago, still with all the cad designs, data, simulators etc. on the HDD, so they were definitely users of SGI X86 machines for car development work.

smj wrote:

kramlq wrote: Just be thankful the vintage hardware gods have blessed you with a mk1 Omron. Some of us have none

I humbly acknowledge the truth of what you say. I was indeed very fortunate back in 2002...

But this seems like the best opportunity I've had to date to find other Japanese workstations, so I've got to make at least a small effort...

Definitely. Please post some info/pics you do find any - it would be interesting to see CMU Mach 2.5 in action, or UniOS.

jan-jaap wrote: When I moved everything to/from a storage unit when I constructed my computer annex, I rented a truck with a tail gate lift.

I've successfully moved tall racks over long distances using a horse trailer. They have a ramp to load the equipment, and it's possible to secure a rack inside the trailer with straps. You can stand inside and unlike most other trailers they usually have a solid roof. Just get rid of the straw and shit before you hit the road

Moving lots of supercomputer equipment into an innocuous looking suburban house from a horsebox... You must have had the neighbours talking about what you really must be doing for a living. Just erect about 5 prominent large satellite dishes pointing in various directions and start calling it your lair or the base to complete the illusion

Y888099 wrote:

smj wrote: NCD-19b, -16c, etc - 68k-based models, before any MIPS or 88k models)

Weren't 68Ks (020? 030? 040?) a bit slow ? Never seen any 88K model, what does it run? VxWorks?


HP Envizex, here it is a video, it shows a unit booting up (ENWARE X-station software. Release 7.0, maybe)
What about HP-Agilent Entria? Any feedback? I know they exist, never seen/tried one.

p.s. the brochure says HP ENVIZEX "p" series and ENTRIA"Plus" users who purchase an optional MPEG accelerator card can now view full-speed MPEG videos with synchronized CD-quality audio.

I used an Envizex with a 19" monitor for software development/admin in my first real job. Great system, and cost a lot of money at the time - as much as the standalone HP 712 workstations the group next to ours had I was told. I remember one of my co-workers even mentioned it had a SCSI bus and could attach a floppy locally for example, but I never had any external SCSI peripherals to test and see if this was true or not.

We also had loads of NCD 14" and I think a few Tektronik as well - Probably 100 in total. They were used for running interfaces to applications on servers that controlled production line machinery (laser etches, zebra/IPL printers, RF programming stations, ASRS, box packing machines etc) - mostly running quite simple Motif or menu based apps, so graphics performance was never a huge requirement, and they all did the job quite nicely. Memory was sometimes an issue with the odd program (e.g. in TCL/TK). But most importantly for us, bringing a new XTerm online was very simple - edit a few config files on the server, and set a few entries in the XTerminal's menu system, compared with the hours needed to install and configure Windows on a PC.

I think the M88k was a very common CPU used in X terminals from NCD/Tektronik in the 90s. The Envizex was i960 based. The later Envizex II was MIPS R4300 based - almost like a MIPS based PC without a disk, and a price to match.

The Sun XTerminal 1 should also get an honourable mention in any thread about Xterms

kjaer wrote: I'm not arguing that the xp217c isn't a better choice today (not least being almost five years newer tech during a time of very busy progress), I'm arguing that a 68k is perfectly capable of running a totally serviceable standalone X terminal. Lots of people were happily using scavenged Sun 3/50s as X terminals as late as 1996, and the primary limitation on those was never the 16 MHz 68020 (it was the 4 MB RAM limit). As you noted, the advent of the WWW changed what people expected from an X terminal (for about three instants before it made them completely obsolete). The old stuff is still just as capable of doing the old jobs as it ever was.

We'd have loved XTerms to stay using slower M68k/M88k rather than the newer 133MHz MIPS R4300i etc., but then again, the company I was working for was Motorola

Raion-Fox wrote: It didn't kill off MIPS, just HPC MIPS, since MIPS is still used in network and embedded applications.

I don't mourn MIPS, it's a dirty architecture with nasty register spill tendency, similar to SuperH.

Itanium has its own problems - strictly in-order, difficult to optimize without the right compilers, code density issues.

I wholeheartedly disagree. MIPS was a great CPU to write kernel level code for - you could be executing C code from a 1:1 mapped virtual window within a few hundred lines of boot assembly code. Virtual memory and the TLB were quite straightforward, and the CP0 model of kernel support features was quite clean. The only thing I didn't like was that if you needed to write code across many vendors MIPS processors, there was little on chip support for explicit feature detection. The MIPS32/MIPS64 series addressed this main shortcoming.

I never worked on compilers, but did plenty of disassembly and debugging, and never thought code generated was too bad either. I studied many processor architectures over the years, and always though the "features" like visible register windows on SPARC (and Itanium to an extent) were overkill and had the potential to add inefficiency to critical paths.

johnnym wrote: A HP 9000 Model 712/80 (w/maxed out RAM and additional VRAM module installed) thanks to FlasBurn .

[/code]

Nice. I've got a 712/100 with 192Mb which is quite slow, but probably one of my favourite machines. They were never speed demons - even back in the days these were current, they were mainly used as glorified xterminals where I worked (i.e. an xterminal but with a floppy drive, which is why they were chosen over proper xterminals by the guys in testing).

In terms of maxing it out, beyond RAM, its quite difficult out as the parts are rare. I've got probably the most useful GIO option: the 2nd VGA and Ethernet GIO Card with breakout cable, but you don't see them for sale much. The second (teleshare) slot is also impossible to find anything for, but pretty useless these days anyway. Even the floppy + matching bezel is very difficult to find outside of buying a complete system with one installed. There is also a nice matching external CDROM and HDD if you want to go for the full setup