Missing computer company's

[WinterDeere]

We found modern CPUs with multi-threading slowed down older programs as the first thread would wait for the other threads to finish.
We turned off multi-threading for older programs and they really flew.

My primary simulation software still had instructions to manually disable hyper threading in the BIOS, as late as 2022. Just in the last few years, it has started automatically handling hyper threading, toggling it in/off automatically, depending on problem type and size.

I think hyper threading helps a lot when there are many small problems to solve in parallel, but the overhead of splitting the problem into parallel threads and then solving them on the same physical core will actually slow you down if solving single very large problems. Always better with more physical cores.

 

[MossRoad]

I was raised on UNIX on DEC PDPs and VMS on VAXes. I much preferred UNIX.

DOS came along and we ran that as terminal emulators for DEC VT100s and VT220s. We had very little use for windows. We started running SUN equipment and DEC Alpha's, multitasking way before windows, 64 bit way before windows as well.

Corporate tried forcing windows NT servers on us... so we made them use them (we were or owner's landlord and IT support). That didn't last long.

One good thing I did with windows workstations was make them into robot boxes. I could run a keystroke capture program, then have someone do their job all day. Then we'd take the captured keystrokes from the operator, tweak them through windows visual basic, and soon be able to automate that job. Then we'd just have to allow for anomalies or odd circumstances. That came it real handy for someone that would have to run a certain task multiple times a day. They no longer had to do it. The windows workstation would just sit there on a shelf and do it as required with no intervention. If it failed or found an outlier circumstance, it would throw an error, we'd figure the error out, and reprogram to allow for that error.

I had at least 50 of those robot boxes sitting there running 24/7/365 doing menial tasks, freeing up people to do other things, and filling jobs that were lost to attrition.

Single task. Single box. That's all windows was good for in the beginning.

 

[garry1p]

Sometime around 1966 RCA designed and sold the Spectra 70 computer with a unique 1st of it's kind virtual memory. IBM had 99% of the main frame business with the IBM 360 and convinced just about everyone virtual memory would not work and killed the Spectra sales forcing RCA to cancelled production.

IBM comes out a few years later selling the IBM 370 with virtual memory and sells it as the future of computers.

I'm not sure about the exact dates, but IIRC in 1971 I was offered a job at Disney World where the entire park was controlled by RCA Spectra computers. At the time I was unable to relocate and turned the offer down however, a few of my friends accepted and moved to FL, we crossed paths several years later working for Telex.

 

[BravoXray]

If you like videos about restoring old computer systems to operating order, give Usagi Electric's YT channel. He recntly got a Bendix G15 vacuum tube computer from the 1950s up and executing code read from paper tape.
He even built a 1 bit vacuum tube computer that is mounted on his wall, that loads 5 bit code off a paper tape reader built from scratch.

Usagi Electric

 

[kenmbz]

One thing I liked about the Vax was shared memory, so multiple nodes on a cluster could share program data securely.
I even had Unix running on the Vax machines to talk to the Sun boxes.
When PCs came out, had to start from scratch on communications.

Only issue was Vax to IBM mainframe transfers ASCII to EBCDIC.
The IBM box was about 150K plus SW maintenance costs.
So I wrote a translator and moved tape reels between the systems to move the financial data (IBM) to cover the inventory data (Ingress on Vax).

 

[JasperFrank]

Windows NT was the first true multitasking Microsoft OS, released July 27, 1993. Windows 95 was released July 14, 1995. OS7 was released May 13, 1991.

Not true. DR DOS had multi tasking.

 

[quicksandfarmer]

Not true. DR DOS had multi tasking.

DR DOS wasn't Microsoft, it was Digital Research.

 

[stuckmotor]

Gateway, Compaq, Xerox.

I loved my Gateway.

 

[ning]

Altos

 

[hayden]

Here... geek out....

Vintage Computing Dec Digital FOR SALE!

Vintage Computing Dec Digital FOR SALE!. Shop the Largest Selection, Click to See! Search eBay faster with PicClick. Money Back Guarantee ensures YOU receive the item you ordered or get your money back.

picclick.com

That's awesome. For a few years in college I bought, fixed, and resold a whole bunch of DEC PDP 11 equipment. Whole racks of stuff. It was a lot of fun, and I made good money off it.

 

[hayden]

Since we are going down memory lane, here is some obscure stuff I worked on over the years.

Worked at Bell Labs designing digital PBXes which are the phone switching systems for businesses as opposed to central office switches. Just smaller versions of central office switches. I did the development work for their first ISDN phone that allowed you to plug a computer directly into the phone and send digital data straight through the phone system without any modem. Sent the first data over it.

Also did a bunch of work on making the system work even when it wasn't working. It involved continuous testing of hardware, and auditing of internal data to detect problems, then removing the hardware from service if a problem was detected, and correcting the data if there was any corruption. All the hardware (including big circuit boards) was hot-pluggable and the system would adapt on the fly and of course keep running without interruption. When a new board was installed it would be put into service seamlessly. All this is something that didn't appear in PCs until 10-20 years later, and never really worked.

Did the technical training on all the internal design and implementation of the system for other groups that were adapting it to specialized applications. The most common was Hotel/Motel to provide the phone service in those businesses. The other was this cool new thing call cellular telephones. They were using our switches at each cell tower, and handing off calls from tower to tower without dropping them. Look where that went...

Needless to say, everything was done usign UNIX and C, both invented down the hall by Kernigan and Ritchie. And it all ran on a row of VAX 11/780 computers from DEC.

Moved to DEC to build the first fiber optic data network, FDDI. Anyone remember that? It never really took off. Anyway, my team wrote all the software to control various FDDI network devices. And when we got our first chips back from the foundry, I cobbled together the first system which was an ethernet to FDDI bridge. Then I sent the first data message over FDDI which was a "we did it" email to Ken Olsen. A bunch of other companies were also building FDDI, but we were way ahead of them all because we had our own silicon where everyone else was waiting for Motorola. I also found the first (and only) bug in one of our chips when I built that prototype. My name can be found in the ANSI and IETF (internet) standards for FDDI. I also got my first patents in this area.

Switched from networking to PCs and operating systems at DEC. Was responsible for the first "multimedia" products in DEC's PC line, which included a CD player (it seemed like a big deal then), a motion JPEG encoder/decode option board, and a set of software MPEG encoder/decoders. All this let you create and play live video on a PC. I remember on exec asking me what anyone would use this for. Now he knows. I also devised a way to use ethernet (or really any network) to broadcast multiple video channels and let arbitrary viewers tune in, kind of like cable TV. At the time, running video, let alone multiple simultaneous video streams across a data network was the proverbial 10 lbs of sh$t in a 5 lb bag. I came up with a way to dynamically allocate multicast addresses for "channels" that were being viewed, and viewing stations could then tune into a channel by listening on the appropriate multicast address. Totally unnecessary now, but got me a handful more patents.

Then I got involved with porting Windows NT to DEC's Alpha systems. The NT guys were all ex-DEC people, so there was a tight relationship among developers. It was a valiant effort to make Alpha relevant in the market, but few of us thought it would save DEC. What it really was for most of use was a good exit ticket out of DEC.

My big contribution was coming up with a way to use off-the-shelf PC option cards in Alpha system. Many option cards had "BIOS Extensions" which were Intel code on the card that ran when you booted the system up. They were essentially drivers to operate the card until the OS was running and full drivers were installed. The problem is that the extensions were intel code and we had an Alpha process that couldn't run it. So I wrote an Intel CPU + PC hardware emulator that was able to run the bios extension code. Suddenly, you could use any video card, or any disk controller card in an Alpha system, vs one card sold by DEC and understood by the Alpha boot code. It opened up a big door for Alpha and is in every system they shipped. And I got a number of other patents and a big promotion off that project.

Then I moved to a startup that got bought by Adaptec. Anyone remember Adaptec? They were hot at the time. My team built the first 64bit PCI raid controller, including custom silicon. I brought both Dell and HP on as customers, and it became a big business for Adaptec. We had a significant cost and performance advantage because of our custom silicon, but Adaptec managed to fumble and fail, and thankfully I got out at the top rather than on the down slide.

Then I went and started EqualLogic building networked RAID storage boxes. They were ethernet connected, self-grouping, and self-managing, and became a huge hit for small to medium business who needed big storage but could never afford to do it like the Fortune 500 companies. We filed and were approved to go public in 2007, and the day before doing so we inked a deal with Dell to buy us. On closing day, I said good by to everyone and walked out the door to start my first day of retirement.

Since then I dabbled with board positions on a couple of startups, but decided that wasn't for me. In fact, my one recurring nightmare is that I find myself with a job, realize I don't need to be there, and start trying to plan an extraction. Retirement suits me much better.

 

[kenmbz]

I still have my Kernigan and Ritchie C book somewhere.
The PBX I worked on was the last of the 8" floppy ones, was already quite obsolete by the time I worked on it.
Thanks for the history, mimics mine a bit, though I spent more time originally on Apple/Atari/Commodore doing game software before moving into hardware and real time control software.
And I have no patents, my disk editor and resonance calculator were in use all over, but I never had my name attached. Oh well.

 

[riptides]

Then I got involved with porting Windows NT to DEC's Alpha systems. The NT guys were all ex-DEC people, so there was a tight relationship among developers. It was a valiant effort to make Alpha relevant in the market, but few of us thought it would save DEC. What it really was for most of use was a good exit ticket out of DEC.

Sad.

I was in software at that time. We forged deeper relationships with MS due to the Windows NT work.
It kept a lot of us employed for a very long time.

Recall some of those Alphas could run DEC's UNIX variant, DEC's OpenVMS and NT. And the clustering tech was way beyond anything else on the market.

 

[WinterDeere]

I still have my Kernigan and Ritchie C book somewhere.

In my first house, which I bought when I was still single, my office was built into the biggest bedroom on the second floor. Wall-to-wall industrial shelving, filled with old manuals, IEEE Proceedings, Transactions, and Letters, IC and processor data sheet books (remember those?), catalogs... you name it. It seemed every engineer's "power" could be measured by the size of their literature collection.

I was working my way through that house, renovating each room, which because the joists had sagged since being built in the 1870's, meant tearing out carpet, tearing up old damaged flooring, re-leveling the joists, and then laying down new flooring I'd mill from purchased caches of old yellow pine. Some of the flitches I bought were sawn at the same time my house was built!

Anyway, the office was next in the renovation schedule, and this happened to be about the time I started dating my wife, so she was there helping me to dismantle all of the computers and empty the room out to start that process. She's looking at this huge wall of shelving full of books, makes a big sweeping motion with her arms as if painting the sky, and says "See this? There's a thing called 'The Internet', and all of this is on it."

I filled a dumpster with books and catalogs and journals, that weekend. The shelving got repurposed for basement storage of her crap, when she eventually moved out of her apartment and into my house, some of which we're still carting around with us today.

 

[hayden]

Sad.

I was in software at that time. We forged deeper relationships with MS due to the Windows NT work.
It kept a lot of us employed for a very long time.

Recall some of those Alphas could run DEC's UNIX variant, DEC's OpenVMS and NT. And the clustering tech was way beyond anything else on the market.

Yes, the Alpha systems were generally available with your choice of OS.

Oh, and I forgot about clusters. That was huge for DEC in the mini computers where they invented it. While I was doing the emulator, another team in the same organization was creating Clusters for NT. After I finished the emulator I picked up responsibility for that project which was horribly late, and finally got it out the door. Then for some reason that evades me, DEC wanted to further suck up to Microsoft and decided to give MS the Clusters technology, and that became Microsoft Clusters for NT.

Another cool project was a PDP-11 emulator that could run PDP-11 programs on an Alpha NT system. The goal was to capture all those legacy systems and sell them Alpha NT machines. We contracted a couple of Russian PhDs who had built a PDP-11 clone during the cold war and knew the systems inside and out. For something like 50 cents they wrote the whole thing, then retired to the life of kings.

 

[hayden]

In my first house, which I bought when I was still single, my office was built into the biggest bedroom on the second floor. Wall-to-wall industrial shelving, filled with old manuals, IEEE Proceedings, Transactions, and Letters, IC and processor data sheet books (remember those?), catalogs... you name it. It seemed every engineer's "power" could be measured by the size of their literature collection.

I was working my way through that house, renovating each room, which because the joists had sagged since being built in the 1870's, meant tearing out carpet, tearing up old damaged flooring, re-leveling the joists, and then laying down new flooring I'd mill from purchased caches of old yellow pine. Some of the flitches I bought were sawn at the same time my house was built!

Anyway, the office was next in the renovation schedule, and this happened to be about the time I started dating my wife, so she was there helping me to dismantle all of the computers and empty the room out to start that process. She's looking at this huge wall of shelving full of books, makes a big sweeping motion with her arms as if painting the sky, and says "See this? There's a thing called 'The Internet', and all of this is on it."

I filled a dumpster with books and catalogs and journals, that weekend. The shelving got repurposed for basement storage of her crap, when she eventually moved out of her apartment and into my house, some of which we're still carting around with us today.

Man, this is all bringing back great memories. I had a "Dexter" room when I met my wife, and she still married me. At the time we had a barn with a big loft space - probably 40 x 60. There was a propane wall heater at one end that could heat the whole space, mostly, sort of. I wanted a "lab", but didn't want to heat the whole space, so I built a make-shift room around the heater. It was widely spaced 2x4 studs, a framed door, and plastic skins on both sides of the studs creating an air gap in between. It was super cheap to build, and stayed nice and warm for relatively little propane cost.

As projects, I hacked a number of different hard disk interfaces to various external disk drives. There was one for 8" floppies, then I got a surplus hard disk and built an interface for that. Then shortly after the IBM PC came out, ATT started selling a clone and offered a super discount to employees (I was at Bell Labs at the time). I bought one and the first program I wrote was a bypass to crack the protection lock on Lotus 123 so I didn't have to buy it.

Also in my first week or so at Bell Labs, I bought an HP16C programmers calculator which was awesome for doing binary, octal, and hex math and manipulation. I still have it, it works great, and I have only changed the batteries once is 42 years.

 

[MossRoad]

I loved my Gateway.

At least the Gateway Cows transitioned to new jobs...

 

[kenmbz]

My senior project was an A/D - D/A for an IBM PC.
Made an ISA card and got it working using basic.
I still have the card, all hand wired on a blank Card with added pins and sockets.
Back then bus speed was pretty slow and most of the chips were gates, registers and seperate A/D and D/A.
To make it simple I only had it measure 0-4 volts as the supply to the bus was 5 volts.
Though I could have gotten 12 v to the board, it would have meant a lot more chips to do the comparison.
Now I use an Arduino and that whole board is about the size of just the sampling chips.

When I worked big cameras for the military, the sensor was about the size of my current phone and so was the the A/D chip.

Most of the design around the chips was to lower noise in a mil grade environment.
Did a lot of tempest testing back then, had to do it at 2am to lower the outside noise from the base and town.

Our 2k x 2k monitors took at least 2 people to lift.
And out frame buffer was about 12 RAM boards with a high speed tape drive for storage.
That drive cost 250K even back then and the printer was a high res laser to film.

I did the tape driver SW and the output stream to the printer.

Needed a small stand on forklift to move the printer around.

 

[MossRoad]

I loved my Gateway.

At my old job I had a web server running on a Gateway PC using Windows NT with an auto reboot every evening at midnight. I kept an internal intranet on it with all of my notes about everything in the company. How systems worked. What was their purpose. All their specs. Where every PC in the corp was located. Each one's specs, programs, etc...

User instructions for the programs they used. Tutorials. You name it, I had it on that box.

I accessed it remotely from any PC inside the company firewall and kept it updated religiously. I set it up on a high shelf in a corner of the IT Department, removed the keyboard, mouse and monitor and pretty much never physically touched it for well over 10 years. Anytime I was anywhere in the building I could go to any web browser and type MIS for the URL and it would pop up. Users could do the same thing. They loved it because they didn't have to call for support. Of course, everything that was on it is now at least a decade or more out of date.

Every few year they'd hire network security companies to come in and survey the network. It would show up as something like R2D2. They'd flag it as a security risk. But no one would take the time to trace down its ethernet port and then follow a 250' wire under an IT room floor.

For all I know, that thing is still running to this day.

 

[MossRoad]

I had several DEC Alpha clusters and RAID arrays as well as Dell clusters and RAID arrays. All of them were great. High availability. We had tape backups at the time. You had to buy a license for each node you wanted to backup to tape. So instead, I'd have each node backup to a file, then transfer the file to a node that had a tape backup license and that was that.

We had similar issues with phone and data access at our remote offices. They were long distance calls. Anything more than two towns over, or across state line was long distance. So we set up a multi line multiplexer. Multiple voice and data lines. We'd rent space in a gas station restroom ceiling (yes, you read that right) that had local calling to us and our remote office. Then we'd put call forwarding on those lines. 2 in and 1 out. No more long distance charges. We had that set up for three different offices. Saved a ridiculous amount of phone charges.

The other thing we did was run a land line to a tower behind a fire station near the state line. We ran a wireless lan at the tower with an antenna about 100' up the tower. At our warehouse across the state line, we did the same thing. Now we could cross state lines with no land-line tariffs. Ran a T1 from the warehouse to the news office about 10 miles up the road. That worked for years until the interned did away with the need.