The History of DOE Nuclear Supercomputers: From MANIAC to AMD's EPYC Milan

About the author
Paul Alcorn

Paul Alcorn is a Contributing Editor for Tom's Hardware US. He writes news and reviews on CPUs, storage and enterprise hardware.

Read more
Create a new thread in the Photo reports comments forum about this subject
This thread is closed for comments
6 comments
Comment from the forums
    Your comment
  • alextheblue
    Quote:
    IBM itself fabbed the processor on its CMOS-6S2 process (roughly equivalent to 25nm).

    No. First off it was a hybrid process (0.25 µm feature sizes and 0.35 µm metal layers). Second, and more importantly, that's 250 nm. It's successor was built on a 0.22 µm process which is why they were able to boost the clocks so much.
  • PaulAlcorn
    5190 said:
    Quote:
    IBM itself fabbed the processor on its CMOS-6S2 process (roughly equivalent to 25nm).
    No. First off it was a hybrid process (0.25 µm feature sizes and 0.35 µm metal layers). Second, and more importantly, that's 250 nm. It's successor was built on a 0.22 µm process which is why they were able to boost the clocks so much.


    Thanks for catching the typo, fixed!
  • stdragon
    For those that are unaware, these computers are used primarily due to post Nuclear-Test-Ban Treaties. It's to ensure the reliability of existing stockpiles of warheads as they age. For that, you need lots of number crunching.

    Basically, making sure our spears are strong and the arrowheads remain sharp. All without actually throwing a single one.
  • matheo25
    Sooo... My 2012 i5-3570k pushes out about as much (109GFLOPS) as the world's best computer of 1994 (143.4GFLOPS). I remember my 386 back then and it was not impressive at all, I cannot even begin to comprehend the power of these new supercomputers. Thanks for the article!
  • BillV523
    it looks like an ordinary mobo for a pc
  • BillV523
    seems can't edit last post, I meant of the photo with the sign under it asking
    "what is it?"