Advertisment

Scalability in Many-core Systems: Part 1 in a series

author-image
CIOL Bureau
Updated On
New Update

BANGALORE, INDIA: The traditional approach to application performance was to simply wait for the next generation of processor; most software developers did not need to invest in performance tuning, and enjoyed a “free lunch” from hardware improvements.

Advertisment

With the industry shift to multiple core systems, the situation has changed: performance has to be realized through concurrency, with applications designed to scale as the number of cores increases.

Click here to post your queries on Multicore

Blogger Joel Spolsky says: "As a programmer, thanks to plummeting memory prices, and CPU speeds doubling every year, you had a choice. You could spend six months rewriting your inner loops in Assembler, or take six months off to play drums in a rock and roll band, and in either case, your program would run faster. Assembler programmers don’t have groupies. So, we don’t care about performance or optimization much anymore. While anachronistically posted (in autumn of 2007, when the Serial Coding Era was well and truly over), the message accurately describes choices once available; performance improvements were free -- just wait a bit for new hardware."

Written by Michael Wrinn, senior course architect, Intel Software college, this white paper looks at aspect of predicting scalability and pose question: is parallel scalability the new free lunch?

tech-news