Upcoming Events

Speaker:
Bob Colwell, Intel Fellow (retired)
The Validation Attitude - pdf

http://en.wikipedia.org/wiki/Bob_Colwell

Effective validation starts with the right attitude: ironclad skepticism bordering on cynicism.This talk will be a quasi-random walk through 30 years of design, validation, and management experience; ludicrous allusions will be proffered, wild allegations tossed about, and a good time had by all (except possibly upper management.)

Bio:
Bob Colwell was Intel's chief IA32 (Pentium) microprocessor architect from 1992-2000, and managed the IA32 Architecture group at Intel's Hillsboro, Oregon facility through the P6 and Pentium 4 projects. He was named the Eckert-Mauchly award winner for 2005, the highest honor in the field of computer architecture, for "outstanding achievements in the design and implementation of industry-changing microarchitectures, and for significant contributions to the RISC/CISC architecture debate." He was elected to the National Academy of Engineering in 2006 "for contributions to turning novel computer architecture concepts into viable, cutting-edge commercial processors". He was named an Intel Fellow in 1996, and an IEEE Fellow in 2006. Previously, Colwell was a CPU architect at VLIW minisupercomputer pioneer Multiflow Computer, a hardware design engineer at workstation vendor Perq Systems, and a member of technical staff at Bell Labs. He has published many technical papers and journal articles, is inventor or co-inventor on 40 patents, and has participated in numerous panel sessions and invited talks. He is the Perspectives editor for IEEE Computer Magazine, wrote the At Random column 2002-2005, and is author of The Pentium Chronicles, a behind-the-scenes look at modern microprocessor design. He is currently an independent consultant. Colwell holds the BSEE degree from the University of Pittsburgh, and the MSEE and PhD from Carnegie Mellon University.

Full Abstract:
It's a miracle that anything ever works. Although we are not orders of magnitude smarter than we were 30 years ago, we now design systems that are several orders of magnitude more complicated than what we could do then, with so many internal states that there's no hope of ever testing them all, let alone all the sequences that could occur or the interactions between them. Then we rush these designs into production, placing them in the hands of users who understand almost nothing about system capabilities, limitations, nor internal operations. Then to the surprise of nobody except upper management, operational bugs crop up, and who gets blamed? The designers who caused those bugs? No -- it was validation's job to prevent these escapes, so it's they get to wear the t-shirts with the targets painted on the back. Does that sound cynical? Hey I'm just getting started!

Presentation Notes:
PDF Slides