[ExI] petaflop computers
stefano.vaj at gmail.com
Thu Nov 20 14:42:37 UTC 2008
On Thu, Nov 20, 2008 at 3:26 PM, Spargemeister <sparge at gmail.com> wrote:
> It's just a poorly-written article. Teraflops->petaflops isn't a
> paradigm shift. It may make some computations practical that weren't
> before, but, in general it just made existing codes run faster and
> made more cycles available to more researchers.
This is also my impression, and why it sounds silly that "petaflops
are not just faster computers, it is a qualitative barrier".
Of course, quantity becomes quality at specific thresholds - e.g., a
PC that becomes powerful enough for the first time to accept real-time
dictation may change to an extent the computing paradigm as far as
human interfaces are concerned. Only, "incrementally higher-resolution
climate models" or "snappier running of scramjet simulations" do not
seem to cut it.
More information about the extropy-chat