2 Comments

  1. Maciej on 10 May 2018 at 3:59 pm

    Limiting the story just to cache might be overoptimistic. NUMA locality can matter a lot, memory or network banwidth consumption not mentioning oversubsciption of CPU cores and resources other than cache that are still shared between hyperthreads of single physical core.

  2. Russ on 13 May 2018 at 8:18 am

    Agreed — thanks for stopping by and commenting! 🙂

Leave a Comment





This site uses Akismet to reduce spam. Learn how your comment data is processed.