Irene Zhang@ashley @dan Wow, I’ve never heard anything more true. Also reminds me of when women actually had to weave computers, like this: https://www.sciencenews.org/article/core-memory-weavers-navajo-apollo-raytheon-computer-nasa
Dan Ports@dan @ashley Yes, here is the Cray article that I was looking for! Little old ladies that wired the backplane because the wires were so dense that no one else could do it! https://digibarn.com/collections/systems/crays/cray1/portrait/
Peter Desnoyers@irene @dan @ashley on a tangent, not related to gender roles in early CS (how did the Cray 1 get to be ”early”? Damn, I’m getting old)
In Tom Knight’s high performance computer architecture class in the 80s he claimed that one of the few patents on the machine had to do with those Freon cooling pipes. They tried boring holes through the aluminum racks, but enough Freon leaked to carry some oil, and the circuit board speed relied on having air dielectric on one side of the traces; oil had a much higher dielectric coefficient and slowed the signals too much. (A related anecdote of his described how Cray or someone else using the same trick got a PCB board vendor who “helpfully” added a layer of solder mask, destroying a whole run of boards)
Fun class from a long-gone era.
@irene @dan @ashley there was one useful rule of thumb I learned in that class.
Air cooled heat sinks need turbulent flow for efficiency, because laminar flow basically provides an insulating blanket of air. How do you know it’s fast enough? There are no simple equations to figure it out, but all you need to do is to crank up the fans until it makes noise, preferably a lot of noise.
Unless of course you’re reading LFM numbers from an IPMI console on a remote machine, or worse yet getting them over Slack from a student I’m working with. Then you just curse, surf the internet trying to figure out how many LFM your stupid SmartSSD needs, and discover this gem: https://forum.digikey.com/t/a-thermal-management-example-part-2-when-datasheets-attack/13252, where someone works out the cooling needed for the “absolute maximum” current rating in a MOSFET data sheet. It involves quite a bit of liquid nitrogen.
A Thermal Management Example (Part 2: When Datasheets Attack)
In part 1, the process of making basic thermal management calculations was described. This page builds on that discussion, and demonstrates the importance of such analyses as an inflation gauge for certain component specifications. The previous example assumed use of an IRL3713PBF transistor (TO-220 package), HS278-ND heatsink, and BER183-ND thermal pad, and showed the calculation of device temperatures using (for sake of simplicity) an assumed 4 watts of electrical power dissipation in the d...