The Changelog avatar

Dexa/The Changelog

Learn more
Published Feb 16, 2018

Moore's Law and High Performance Computing

Dive into the evolution of high-performance computing with Todd Gamblin, as he examines Moore's Law's lasting impact, the critical applications in scientific advancements, and the transformative role of open source software in supercomputing.
Episode Highlights
The Changelog logo

Topics covered

Popular Clips

Episode Highlights

  • Historical Context

    Moore's Law, initially an observation by Gordon Moore in 1965, noted the doubling of transistors on a chip approximately every 18 to 24 months. explains that this process, which involves etching transistors onto silicon, has significantly influenced computing power over the decades 1. clarifies that while Moore's Law is often misunderstood, it remains relevant, albeit with some slowing in recent years 2.

    Moore's Law is not dead. Although it's fair to be confused because there have been a lot of articles written about this.

    ---

    This historical perspective sets the stage for understanding its impact on technology development.

       

    Impact on Computing

    The influence of Moore's Law on computing power is profound, as it has driven the industry to continually enhance chip performance. notes that while single-thread performance has plateaued, parallelism allows for greater utilization of chip capabilities 1. The breakdown of Dennard scaling around 2006 marked a shift towards multicore processors, which describes as "massively parallel machines" 3.

    You don't get as much single thread performance as you used to. That's kind of capped out.

    ---

    This shift underscores the ongoing evolution in computing architecture.

       

    Potential Limitations

    As Moore's Law approaches its physical limitations, the future of chip technology faces significant challenges. predicts a slowdown in the rate of transistor doubling, leading to a need for more specialized hardware solutions 4. highlights that while the number of transistors may become constant, the focus will shift to maximizing their efficiency 1.

    If the number of transistors that you can fit on a chip becomes constant, then the only way that you can get more speed is to make more effective use of them.

    ---

    This transition marks a pivotal moment in the evolution of high-performance computing.

Related Episodes