TE22: Talking FPGAs with Dr. Ray Hoare

Here are some of the topics we touched on while talking FPGAs with Dr. Ray Hoare President, and CEO of Concurrent EDA:

  • How are FPGAs differ from normal chips?
  • It’s all about bringing new applications to FPGA.
  • The advantages of FPGAs over generic CPUs.
  • FPGAs have more repeatable performance because they are dedicated, and don’t get interrupted with system tasks.
  • Aligning processing performance between a 22-core Xeon and a current FPGA using a factory analogy.
  • How do we see FPGAs attacking everyday problems like encryption?
  • What’s up with this system-on-a-chip (SoC) FPGA approach?
  • Huge amounts of bandwidth coming into the chip, demand more compute to offload the host CPU.
  • How is programming an FPGA different from a generic CPU?
  • Today programming FPGAs is still more art than science, while compilers are fantastic, the tools for FPGAs are not yet at that level.
  • When you have a gigabit or more of raw data per second coming in you’re better off pushing that through an FPGA.
  • To move your application or algorithm into an FPGA it needs to be mature, and well understood, ex. electronic trading, encryption, or data deduplication.
  • Why are cloud environments so excited about moving to FPGAs?
  • Are FPGAs going to be how we jump forward into artificial intelligence?

Leave a Reply