Bekker's Blog

Blog archive

Microsoft Research Sees a Brownout for Moore's Law

Ever wonder where Microsoft's much-touted billions in R&D spending goes?

At least part of the money funded what's emerging to be an influential research paper that projects a radical slowdown on progress in Moore's Law, the idea that the number of transistors that can be inexpensively fit on an integrated circuit doubles roughly every two years.

Doug Burger, a computer scientist at Microsoft Research, co-wrote the paper called, "Dark Silicon and the End of Multicore Scaling" (PDF). Burger and the other authors postulate that chipmakers are reaching limits on getting electrical power to all of the ever-smaller transistors that are getting jammed onto chips.

According to an account in The New York Times today:

"In their paper, Dr. Burger and fellow researchers simulated the electricity used by more than 150 popular microprocessors and estimated that by 2024 computing speed would increase only 7.9 times, on average. By contrast, if there were no limits on the capabilities of the transistors, the maximum potential speedup would be nearly 47 times, the researchers said."

Not all experts are as pessimistic that the technical challenges won't be overcome, but the research paper seems to have struck a nerve.

Posted by Scott Bekker on August 01, 2011


Featured

  • Report: Cost, Sustainability Drive DaaS Adoption Beyond Remote Work

    Gartner's 2025 Magic Quadrant for Desktop as a Service reveals that while secure remote access remains a key driver of DaaS adoption, a growing number of deployments now focus on broader efficiency goals.

  • Windows 365 Reserve, Microsoft's Cloud PC Rental Service, Hits Preview

    Microsoft has launched a limited public preview of its new "Windows 365 Reserve" service, which lets organizations rent cloud PC instances in the event their Windows devices are stolen, lost or damaged.

  • Hands-On AI Skills Now Outshine Certs in Salary Stakes

    For AI-related roles, employers are prioritizing verifiable, hands-on abilities over framed certificates -- and they're paying a premium for it.

  • Roadblocks in Enterprise AI: Data and Skills Shortfalls Could Cost Millions

    Businesses risk losing up to $87 million a year if they fail to catch up with AI innovation, according to the Couchbase FY 2026 CIO AI Survey released this month.