[Beowulf] SPEC CPU 2006 released

Robert G. Brown rgb at phy.duke.edu
Mon Sep 4 07:09:20 PDT 2006

On Sun, 3 Sep 2006, Ed Hill wrote:

> On Sun, 03 Sep 2006 16:08:05 +0200
> Toon Moene <toon at moene.indiv.nluug.nl> wrote:
>>> This raises a really "interesting" question about the depth at
>>> which open source GPL code is embedded in a tool when the viral
>>> clause kicks in.
>> You can't be serious here, Dr. Red-Green-Blue.
>> Gcc has been part of SPEC since, what, '95 ?
>> You can compile and run GPL'd code all you want - no restriction
>> there. It's only when you want to distribute changed code you have
>> have to supply source.
> Yes.  Its 2006.  I think we've all had ample time to read and
> (mostly?) digest the popular licenses.  :-)
> RGB's insinuations about viral nature are easily addressed.  If:
>  1) SPEC provides full source code for all the GPL-ed bits
>     to all the folks who purchase the SPEC suite, and
>  2) none of the non-GPL-ed SPEC stuff links against the
>     included GPL-ed bits
> then, in all likelihood, there are no violations.  So, no big deal or
> "interesting" questions here.  And if someone wanted to extract the
> GPL-ed bits from the SPEC suite and re-distribute them (per the GPL)
> then they'd have the ability and every right to do so.


Ooowww, taken behind the woodshed and schooled, that's what I am;-)

Sure, sure, as long as SPEC doesn't use the output of any of the
programs it uses but acts only as a timing shell, it is an edge case
PRECISELY on the boundary of the activity that triggers the viral
clause.  It has one foot on the slippery slope that leads to the ability
to embed GPL programs in non-GPL scripts to take advantage of the work
done by their authors (which is prohibited), but it is anchored well
enough by the other not QUITE to be at issue, unless the authors wrote
the program to have good timing characteristics so that the time it
takes to run IS considered to be a part of its productive use.

Just to give an interesting and provocative interpretation of a
program's "output", which needn't be just letters.  This is a nontrivial
example, btw, as there might well exist programs written for that
purpose, and whether any given program has that as ONE of its forms of
"significant output" is really pretty much up to the author.

However, I wasn't REALLY seriously suggesting this (certainly not in the
sense of trying to sow FUD:-).  Rather bringing it up was more an act of
mischief, tongue in cheek, to get people to think about the possibility
of cloning SPEC without the silly and expensive licensing so that
everybody could use it.  It's not exactly like writing a timing shell is
rocket science or something, especially not macro shells for entire
programs (microtiming shells are highly nontrivial, OTOH, as I
personally will attest), and no I don't believe that they can even
restrict the usage of the particular argument lists they feed into the
programs at issue or the compilers used.  The authors might have
something to say about somebody trying to "copyright" some particular
argument list to their program -- I know I would.

In fact, I TRUTHFULLY think that the "licensing" of the entire SPEC
suite is something more than a bit humorous in and of itself.  ROTFL
funny.  An "endorsement" company (non-profit or not) run by the very
companies whose products it tests.  Marketing in action, where the
"marketing" extends right dow to the "SPEC Associates" (the elite
universities and other supposedly publically spirited members that
"bless" the suite with objectivity) who gain exposure, maybe some grant
support, something for their CPS departments to do...:-)

This, in turn, was supposed to stimulate a discussion on whether or not
benchmarks of this sort "need" to be licensed and controlled (I would
argue a resounding "no") and if so controlled by whom, to prevent vendor
abuse or benchmark drift or the healthy discussion, criticism, and
modification that results in a true open source process.  This might
have lead to discussing e.g. the top500 rankings and linpack and other
(more open) macro benchmarks, whether the "member organizations" of SPEC
are above reproach and have no self-interest to promote (Oh wait!  The
member list, who doubtless pony up the actual cash that runs the
non-profit, is made up of all the top COMPANIES that sell HARDWARE?  No
hint of self-interest there...;-).

Oh, we could have had such fun!

But NOooo, instead we got FUD, and my bottom is all sore... and besides,

  I teach today
    and cannot play,
      even though it's "labor day".



Robert G. Brown	                       http://www.phy.duke.edu/~rgb/
Duke University Dept. of Physics, Box 90305
Durham, N.C. 27708-0305
Phone: 1-919-660-2567  Fax: 919-660-2525     email:rgb at phy.duke.edu

More information about the Beowulf mailing list