Re:Ways to Fool the Masses Giving Performance Results on Vec...
by andrewl on Saturday July 13, 2019 @09:13AM
(#20427885)
I'd never heard of him. He's done a lot of interesting stuff, mostly in high performance computing. From his personal page:
"Gustafson has recently finished writing a book, The End of Error: Unum Computing, that presents a new approach to computer arithmetic: the unum. The universal number, or unum format, encompasses all IEEE floating-point formats as well as fixed-point and exact integer arithmetic. This approach obtains more accurate answers than floating-point arithmetic yet uses fewer bits in many cases, saving memory, bandwidth, energy, and power."
Re:Ways to Fool the Masses Giving Performance Results on Vec...
by mark-r on Saturday July 13, 2019 @06:34PM
(#20430558)
Too bad Seymour Cray is no longer around to pick on, or to inspire.
I was lucky enough to tour the Chippewa Falls facility early on - serial number 5 of the Cray-1 was undergoing final tests before delivery.
We were lucky enough to see Seymour in his office. He pointed to a pumpkin on his desk, and declared that it was the Cray-3. His daughter had grown it, and since she knew he was already working on a Cray-2, she decided this would be the Cray-3.
Re:Ways to Fool the Masses Giving Performance Results on Vec...
by tangus on Saturday July 13, 2019 @10:02AM
(#20428093)
Burning it, actually. It's a legend about how Seymour Cray would design and build a sailboat every year only to burn it afterward, so he could design a better one, or something like that.
"Gustafson has recently finished writing a book, The End of Error: Unum Computing, that presents a new approach to computer arithmetic: the unum. The universal number, or unum format, encompasses all IEEE floating-point formats as well as fixed-point and exact integer arithmetic. This approach obtains more accurate answers than floating-point arithmetic yet uses fewer bits in many cases, saving memory, bandwidth, energy, and power."
Has anybody read it?
https://www.amazon.com/dp/1482239868
http://www.johngustafson.net/pubs/pub13/amdahl.htm
“Speedup should be measured by scaling the problem to the number of processors, not fixing problem size”
http://www.johngustafson.net/presentations/Unums2.0.pdf
I was lucky enough to tour the Chippewa Falls facility early on - serial number 5 of the Cray-1 was undergoing final tests before delivery.
We were lucky enough to see Seymour in his office. He pointed to a pumpkin on his desk, and declared that it was the Cray-3. His daughter had grown it, and since she knew he was already working on a Cray-2, she decided this would be the Cray-3.
If you read any biographies of US physicists of the time, it seems they either all sailed, or had a close scientist friend that owned a boat.