Re NSA says China's supercomputing advances put US at risk

["Dave Farber" <dave () farber net>]

---------- Forwarded message ---------
From: Jerry Leichter <leichter () lrw com>
Date: Fri, Mar 17, 2017 at 8:54 AM
Subject: Re: [Cryptography] NSA says China's supercomputing advances put US
at risk
To: Christian Huitema <huitema () huitema net>
Cc: <cryptography () metzdowd com>



> > > NSA, DOE say China's supercomputing advances put U.S. at risk
> > The tl;dr version:
> >
> > Fnord fnord China fnord supercomputers fnord more money fnord.  Did we
mention
> > China fnord?  And MORE MONEY?
>
> A few years ago, s/China/Soviet Union/. Same pitch.
OK, we're all properly cynical.  Just because someone will potentially make
a great deal of money out of solving a "problem", doesn't mean there isn't
an actual problem hiding in there somewhere.  And some of the people who
would get the money actually *are* experts in the appropriate field.  (Of
course, I guess you can subscribe to our President's attitude that
"so-called experts" are all liars anyway.)

> Somehow, the massive data centers built by Google, Amazon, Microsoft or
> IBM do not seem to count...
Actually ... they don't.  I can only speak to what Google builds.  It's
indeed absolutely massive at an unprecedented scale that few on the outside
can even begin to wrap their heads around ... but it's optimized for a
large but certainly not universal set of problems.  Want to run a highly
parallel, loosely-coupled, data-intensive, mainly integer and string
manipulation computation?  We've got something perfect for you.  Want to
run a massive, tightly-coupled, data-light (relative to computation)
floating point computation?  There are better designs out there.

We can debate whether problems of this sort are important.  We can debate
whether Google-like approaches, even if slower, are "good enough" by now to
deal with them.  We can certainly debate, even if it's important to develop
machines to solve such problems, whether government money should be
invested in them - one can certainly argue that if there's actual demand,
private money will be available.  (Note that if you look at the history of
super-computing, even when commercial enterprises like CDC and IBM where
involved, many, perhaps most, of their customers were relying on government
contracts one way or another, so (a) the line is rather fuzzy; (b) it's not
a trivial question.)

But just outright dismissing the whole issue as "more pandering for a
bigger place at the government trough" just doesn't cut it.

BTW, I'm sure Baidu has Google-like data centers in China, too.

                                                        -- Jerry

_______________________________________________
The cryptography mailing list
cryptography () metzdowd com
http://www.metzdowd.com/mailman/listinfo/cryptography



-------------------------------------------
Archives: https://www.listbox.com/member/archive/247/=now
RSS Feed: https://www.listbox.com/member/archive/rss/247/18849915-ae8fa580
Modify Your Subscription: https://www.listbox.com/member/?member_id=18849915&id_secret=18849915-aa268125
Unsubscribe Now: 
https://www.listbox.com/unsubscribe/?member_id=18849915&id_secret=18849915-32545cb4&post_id=20170317092028:7B5A4960-0B14-11E7-AF5B-9C8F47BB072B
Powered by Listbox: http://www.listbox.com