The dissipation bottleneck, which slowed the progress of clock frequency and shifted computing systems towards multi-cores, was a reminder that the smooth evolution of technology we have enjoyed for decades may not last forever; the recent focus on reliability issues suggests that "technology pressure" will only increase over time. While architecture, compiler and programming language researchers try to find ways to circumvent these technology constraints one by one, it also feels like architecture efficiency decreases with each new constraint to be factored in, while the design task becomes each time more complex. Therefore, investigating radically new architecture paradigms intrinsically compatible with these CMOS technology constraints, and/or future and alternative technologies should not be labeled as exotic (if not useless) research: it comes out of a very practical, if not industrial, concern to anticipate drastic changes soon enough to be ready when needs be.
Meetings:
- A first informal and short introductory meeting took place at the May clusters meeting in Munich.
- The next meeting will take place at the October clusters meeting in Poland.