Talk:CSC 456 Spring 2012/ch4b

"He initially expected the first winner to have a speedup close to ten times, and that it would be difficult to advance beyond that." I recall that when we wrote the Cm* book (Parallel Processing: the Cm* Experience), we routinely saw speedups in the 10-20 range. Though the book was c. 1987, most of it was written in 1983 and 1984. I will have to check once I get back to my office.

Oh, please display the equation in math mode. I'm not quite sure how to do this in Mediawiki, but I see enough math that I'm sure there's a way.

I suggest the following references for scaled speedup (I got them from scholar.google.com, searching for "Gustafson scaled speedup"):

• Reevaluating Amdahl's law

[PDF] from sjtu.edu.cn JL Gustafson - Communications of the ACM, 1988 - dl.acm.org ... GM Validity of the single-processor approach to achieving large scale computing capabilities. ... RE. Gustafson, JL, and Montry ... parallel processors General Terms: Theory Additional Key Words and Phrases: Amdahl's law, massively paral- lel processing, speedup Author's Present ... Cited by 843 - Related articles - All 64 versions

• Development of parallel methods for a 1024-processor hypercube. [This is their original paper]

[PDF] from ameslab.gov JL Gustafson, GR Montry… - SIAM J. Sci. Stat. Comput., 1988 - scl.ameslab.gov ... HYPERCUBE* JOHN L. GUSTAFSON, GARY R. MONTRY, AND ROBERT E. BENNER Abstract. ... fixed. The latter measure, denoted scaled speedup, is developed and contrasted with the traditional measure of parallel speedup. ... Cited by 435 - Related articles - View as HTML - All 18 versions

• For superlinear speedup* the first several hits for "superlinear speedup" in Google Scholar seem good. I would suggest reading the introduction & conclusions of the papers before delving into the details (which may not be necessary).