Taiwan Motherboards - NOT!
Paw Prints: Writings of the maddog
I was visiting a friend of mine, Dennis Jensen, while I was in Florianopolis, Brazil. During my visit I started telling Dennis about the period of 1975 to 1978 when some people were building computers from kits by soldering integrated circuits with hand soldering irons to the printed circuit boards. Many of my computer friends built these types of kits, including me.
I remember one kit in particular, purchased by my supervisor at Aetna Life and Casualty, where the directions on how to assemble the system often did not match the circuit diagram, and neither matched the holes in the printed circuit board. This led to some late-night head-banging, trying to figure out which of the three compoents was right. Typically “the holes” won out, but sometimes it was a close race.
I remember the night (after several weeks) that my boss finally got the entire board together, hooked up his television to the video out socket (yes, we used a real TV as the “monitor”), his cassette tape drive hooked up to be able to read the combination editor/assembler/debugger off the audio cassette tape and read it into the system, and finally type in his fifteen line assembly language program which “peeked” and “poked” data into registers and memory.
Unfortunately nothing at all worked, so he asked me to come over to his house to “have a beer”. That meant I was to help him debug the system.
Fortunately for both of us I had recommended that he use sockets for most of his integrated circuits, so he could easily remove the actual chips and test his solder joints to see if they were good. Our only test instrument was a “vacuum tube voltmeter” (itself a kit that I had built in high school) that also tested resistance, so I started using that instrument to see if his solder joints were good, or if there were any shorts.
As I kept testing, the closer I got to the bus pins at the end of the board, the lower the resistance became. This did not make any sense, since the pins were not connected, and as I got closer to the pins, and further away from the theoretical short circuits caused by his soldering, the resistance should have gotten higher, not lower.
Finally we had removed all the possible parts from their sockets that could have formed a short circuit, and the short circuit was still showing up. I started really staring at the board, and as I turned it in my hands a shiny glint of light shown across the bottom of the board...a very, very, VERY thin ribbon of gold connecting all of the pins. The gold was left over from the plating of the pins, and had not been removed. We cut away the gold strand, re-seated all the integrated circuits in their sockets, turned on the computer and everything worked fine, the TV monitor lighting up with instructions to load the editor/assembler/debugger.
Ecstatic, my boss typed in his first program....about 15 lines of assembly code. He then told the machine to assemble the code, and the screen flickered for about one-quarter of a second and stopped.
“Rats”, my boss said, “it still did not work.” “How do you know?”, I asked. “The screen did not do anything other than flicker,” he said.
I sighed. “Tom, you teach compiler theory...how many instructions do you think it took to assemble your program?”
“About 100,000 or so,” he guessed.
“How fast is the processor?”
“400,000 instructions a second,” he estimated.
“How long did the screen flicker?”
“One quarter of a second, or about....100,000 instructions....”
Tom started typing feverishly, and sure enough, the program binaries were all in the memory of the system, ready to be executed.
“Tom,” I said, “you are simply too used to mainframe computers, shared by dozens of other people, which have to pull data off very slow disks. You have no idea how fast these computers really are.”
Dennis listened to this story, then asked me “Why did he spend all that time building this thing? Why didn't he just buy a nice motherboard from Taiwan?”
“Dennis”, I said, drinking the last of my beer, “at that time there were no 'nice motherboards from Taiwan'”.comments powered by Disqus
A major setback for the Linux desktop.
Improved support for GPU in virtualization.
News site for the openSUSE community falls victim to a Wordpress exploit.
The source code is available online.
One out of three virtual machines on Microsoft Azure Cloud run Linux.
The form factor of the board makes it a drop-in replacement for Raspberry Pi.
Makes it easier for customers to move workloads into container-centric applications.
SUSE’s answer to container-centric operating systems.
Linux 4.9 is the biggest release in terms of number of commits.
The latest version of the official RHEL clone is here.