SPEC CPU2006 announced

J

John.L.Henning

The new CPU benchmark from the Standard Performance Evaluation
Corporation is announced

http://www.spec.org/cpu2006/

Readers of comp.arch and comp.benchmarks may
recall that I posted a call for benchmark candidates
a few years ago. Thank you to those who responded.
There were many contributions of real applications,
as well as a variety of freely-available programs,
that were included (see "credits.html" at the URL
above).

The new benchmark is a challenge to compilers: mostly
new code, derived from real applications, and over
3 million lines of code. Nevertheless, a wide variety of
compilers have been used to produce initial results.

- John Henning
Performance Engineer, Sun Microsystems
and
Vice-Chair/Secretary, SPEC CPU Subcommittee
 
F

Flash Gordon

The new CPU benchmark from the Standard Performance Evaluation
Corporation is announced

<snip spam>

and your C question is? Of for those on comp.lang.c++ your C++ question?
Or your Fortran question? The main group where it looks like, from the
name, it *might* be topical is comp.benchmarks.
- John Henning
Performance Engineer, Sun Microsystems
and
Vice-Chair/Secretary, SPEC CPU Subcommittee

If you really are in those positions you should know better than to
bring your companies in to disrepute by spamming.
 
C

Chris F Clark

Flash Gordon said:
<snip spam>

and your C question is? Of for those on comp.lang.c++ your C++
question? Or your Fortran question? The main group where it looks
like, from the name, it *might* be topical is comp.benchmarks.


If you really are in those positions you should know better than to
bring your companies in to disrepute by spamming.

I don't think it is as off-topic for the C/C++/Fortran groups as you
seem to suggest. As I recall, the spec suite of tests includes at
least C (and perhaps some FORTRAN) benchmarks that are typically used
when evaluating compiler (and architecture) performance. Thus, when I
was working on C and C++ compilers, I cared about those tests. I'm
more surprised, that comp.compilers wasn't in the list of cross-posted
groups.

Just my opinion,
-Chris

*****************************************************************************
Chris Clark Internet : (e-mail address removed)
Compiler Resources, Inc. Web Site : http://world.std.com/~compres
23 Bailey Rd voice : (508) 435-5016
Berlin, MA 01503 USA fax : (978) 838-0263 (24 hours)
------------------------------------------------------------------------------
 
E

Eugene Miya

No it's not off topic.

I agree with below. On the other hand some people don't hide behind aliases.
But that's your choice.
I don't think it is as off-topic for the C/C++/Fortran groups as you
seem to suggest. As I recall, the spec suite of tests includes at
least C (and perhaps some FORTRAN) benchmarks that are typically used
when evaluating compiler (and architecture) performance. Thus, when I
was working on C and C++ compilers, I cared about those tests. I'm
more surprised, that comp.compilers wasn't in the list of cross-posted
groups.

Levine has to approve it as moderator. Some people think moderated
are slow. We shield you guys from a lot of spam.

--
 
J

jacob navia

Flash said:
<snip spam>

and your C question is? Of for those on comp.lang.c++ your C++ question?
Or your Fortran question? The main group where it looks like, from the
name, it *might* be topical is comp.benchmarks.



If you really are in those positions you should know better than to
bring your companies in to disrepute by spamming.

I do not agree.

This is an interesting post. This benchmark is widely known and
it is used to evaluate compilers.
 
F

Flash Gordon

Eugene said:
No it's not off topic.

It is not discussing C it is advertising a commercial product. Therefore
it is off topic in comp.lang.c and as it is commercial it comes under
what at least some people currently consider to be spam.
I agree with below. On the other hand some people don't hide behind aliases.
But that's your choice.

Anyone can easily find my real name with either a whois query of
checking back through my posting history. Anyone who knows the perpose
of .me.uk will know this. I think it would even be possible to track
down my current and previous employer for someone skilled in the use of
Google.
Levine has to approve it as moderator. Some people think moderated
are slow. We shield you guys from a lot of spam.

That does not apply to comp.lang.c or at least one of the other groups
since they are not moderated. If it is acceptable on whichever group you
are reading, fine, but it is not appropriate to all groups and thus
gives a bad impression of the company on at least some of the groups.

I will not be posting further in this thread.
 
G

glen herrmannsfeldt

(someone wrote)
That does not apply to comp.lang.c or at least one of the other groups
since they are not moderated. If it is acceptable on whichever group you
are reading, fine, but it is not appropriate to all groups and thus
gives a bad impression of the company on at least some of the groups.

I have replied to articles crossposted between comp.compilers and
other groups. It seems that they don't come through any group until
they are approved by the moderator. I don't understand the process
any more than that, but that is the way I see it.

-- glen
 
E

Eugene Miya

Levine has to approve it as moderator. Some people think moderated
I have replied to articles crossposted between comp.compilers and
other groups. It seems that they don't come through any group until
they are approved by the moderator. I don't understand the process
any more than that, but that is the way I see it.

This is true.

A single moderated group holds propagation of a cross post.
Levine and I have an agreement (as we each over with other groups)
that we will approve posts to each others groups (multiple moderated groups).

--comp.parallel moderator

--
 
E

Eugene Miya

The new CPU benchmark from the Standard Performance Evaluation
It is not discussing C it is advertising a commercial product. Therefore
it is off topic in comp.lang.c and as it is commercial it comes under
what at least some people currently consider to be spam.

First off, there is no prohibition of Usenet for commercial use.
There is no central authority to prohibit such. The only authority to
prohibit any use are given ISPs. If you have a problem with prohibitions,
change ISPs.

It was a small post and they have been around some time.

That does not apply to comp.lang.c or at least one of the other groups
since they are not moderated. If it is acceptable on whichever group you
are reading, fine, but it is not appropriate to all groups and thus
gives a bad impression of the company on at least some of the groups.

SPEC, well intended, has a pretty good, really amazing reputation
among computer firms. I have attend an early cook off and I also
contributed to the usually mis-run NASA7 benchmark.

That's the problem with unmoderated groups.
I will not be posting further in this thread.

Your choice.

--
 
?

=?ISO-8859-15?Q?Juli=E1n?= Albo

Eugene said:
A single moderated group holds propagation of a cross post.
Levine and I have an agreement (as we each over with other groups)
that we will approve posts to each others groups (multiple moderated
groups).

Then approving this type of messages you are encouraging threads like this
one, with very little value to any of the groups, moderated or not.
 
C

CBFalconer

Chris said:
I don't think it is as off-topic for the C/C++/Fortran groups as
you seem to suggest. As I recall, the spec suite of tests
includes at least C (and perhaps some FORTRAN) benchmarks that
are typically used when evaluating compiler (and architecture)
performance. Thus, when I was working on C and C++ compilers, I
cared about those tests. I'm more surprised, that comp.compilers
wasn't in the list of cross-posted groups.

I consider it much more questionable than I did originally. The
benchmarks are being sold; they are not freely available. This
makes the post a commercial advertisement. I would have no problem
if the benchmarks were downloadable.
 
J

John Mashey

CBFalconer said:
I consider it much more questionable than I did originally. The
benchmarks are being sold; they are not freely available. This
makes the post a commercial advertisement. I would have no problem
if the benchmarks were downloadable.

If anyone thinks that SPEC makes a lot of money *selling* this stuff,
or thinks that SPEC expects that, they are living in fantasy-land.

Please read http://www.spec.org/spec/faq/, for example; SPEC is a
non-profit organization, and the sales cover some of the administrative
costs, although of course the largest real expenses are the amounts of
staff time that various organizations devote to doing all of the work.

For most readers of these newsgroups, the relevance of the announcement
is so that people can go read the website and see the results *for
free*, i.e., it is a public service announcement to a relevant group of
people, not a money-making SPAM thing.

Disclosure: I haven't had any formal association with SPEC for a years,
but I was one of the original founders, and I still visit them every
once in a while.
 
C

Christopher Benson-Manica

In comp.lang.c Eugene Miya said:
First off, there is no prohibition of Usenet for commercial use.
There is no central authority to prohibit such.

Granted, but presumably you would agree that in the absence of such
central authority that prevailing norms on individual newsgroups
should be regarded as, at minimum, indicative of "best practice"
regarding posting. comp.lang.c in particular views commercial
postings of any kind whatsoever as inappropriate, a fact prospective
posters would do well to remember.
 
G

glen herrmannsfeldt

Granted, but presumably you would agree that in the absence of such
central authority that prevailing norms on individual newsgroups
should be regarded as, at minimum, indicative of "best practice"
regarding posting. comp.lang.c in particular views commercial
postings of any kind whatsoever as inappropriate, a fact prospective
posters would do well to remember.

First of all, the SPEC post made no mention, or even hint of a
mentioning sales of the SPEC suite.

If you really want to avoid any suggestion of commercial interest
I think we have to ban all posts from .com addresses, as we never
know if they might try to sell something, not to mention that some
people with .edu or .org addresses might want to sell something.

Personally, I consider the 'cost' of reading a post compared to
the possible advantage to readers for anything that might possibly
be considered commercial. For some rare items some newsgroups
might be the most appropriate place to find people, not that others
don't abuse the privilege.

The $200 educational/non-profit price seems very reasonable to me,
though I don't expect to actually buy it. Even for the $800 commercial
license, I don't expect them to get rich. Note that it is spec.ORG.

-- glen
 
?

=?ISO-8859-1?Q?Jan_Vorbr=FCggen?=

The new CPU benchmark from the Standard Performance Evaluation
<snip spam>
and your C question is? Of for those on comp.lang.c++ your C++ question?
Or your Fortran question? The main group where it looks like, from the
name, it *might* be topical is comp.benchmarks.

If you really are in those positions you should know better than to
bring your companies in to disrepute by spamming.

You really shouldn't be caught pontificating on a subject you apparently know
nothing about.

Every commercial and many non-commerical developers of compilers for the C,
C++ and Fortran languages will be studying the details of these benchmarks
closely, if they haven't been working with them in the past three or four
years anyway while the suite was being developed. They will change their
compilers substnatially based on what they find. Language features that are
found in these benchmarks will be better supported than they were before as
a result. Thus, the announcement is quite on topic for all of the groups.

Oh, but I forget - the most vocal people from c.l.c I have heard via cross-
posts seem to think that a programming language should only be discussd in
the abstract, with no reference to actual implementations and programs.
Sheesh.

The "commercial" point has been asnwered by others. You should perhaps
say "TANSTAAFL" to yourself more often. I don't think anybody who really
wanted access to the suite in the past had a problem in getting it. As SPEC
needs to enter into a contract with a user anyway for legal reasons (copy-
right etc.), I find it reasonable that they recover at least the adminis-
trative costs.

Jan
 
G

glen herrmannsfeldt

Jan Vorbrüggen wrote:

(snip)
Every commercial and many non-commerical developers of compilers for the C,
C++ and Fortran languages will be studying the details of these benchmarks
closely, if they haven't been working with them in the past three or four
years anyway while the suite was being developed. They will change their
compilers substnatially based on what they find. Language features that are
found in these benchmarks will be better supported than they were before as
a result. Thus, the announcement is quite on topic for all of the groups.

The point was that people in these groups had submitted programs
from a previous request, and to let them know the new suite was out.
A service to the readers of these groups. I agree that it is on topic.
Oh, but I forget - the most vocal people from c.l.c I have heard via cross-
posts seem to think that a programming language should only be discussd in
the abstract, with no reference to actual implementations and programs.
Sheesh.

I would agree, except that c.l.c really does get an excessive number
of posts that really don't belong there, making it too difficult to
read in a reasonable amount of time. I don't read it much because
it takes too long. Still, I think this is reasonable for c.l.c.
The "commercial" point has been asnwered by others. You should perhaps
say "TANSTAAFL" to yourself more often. I don't think anybody who really
wanted access to the suite in the past had a problem in getting it. As SPEC
needs to enter into a contract with a user anyway for legal reasons (copy-
right etc.), I find it reasonable that they recover at least the adminis-
trative costs.

I don't know that they always had the educational/non-profit rate, at
one quarter the commercial rate. I probably won't buy it, but I
can't really complain, either.

-- glen
 
?

=?ISO-8859-15?Q?Juli=E1n?= Albo

Jan said:
Oh, but I forget - the most vocal people from c.l.c I have heard via
cross- posts seem to think that a programming language should only be
discussd in the abstract, with no reference to actual implementations and
programs. Sheesh.

You seem to not think. People can discuss anything they need or want, they
just must avoid to do it in, or cross-post to, inadequate newsgroups.
 
C

Christopher Benson-Manica

In comp.lang.c CBFalconer said:
I consider it much more questionable than I did originally. The
benchmarks are being sold; they are not freely available. This
makes the post a commercial advertisement. I would have no problem
if the benchmarks were downloadable.

It's worth noting that official copies of the C standard are also not
freely available, presumably for similar reasons.
 
D

Default User

glen said:
First of all, the SPEC post made no mention, or even hint of a
mentioning sales of the SPEC suite.

Yet they are selling, right? In some ways that's worse.
If you really want to avoid any suggestion of commercial interest
I think we have to ban all posts from .com addresses, as we never
know if they might try to sell something, not to mention that some
people with .edu or .org addresses might want to sell something.

This in nonsensical. If the posts contain no commercial content, then
where they are posted from makes no difference.





Brian (thinks the whole thing is pretty dumb)
 
E

Eugene Miya

First of all, the SPEC post made no mention, or even hint of a
mentioning sales of the SPEC suite.

Correct me if things have changed but SPEC is a non-profit consortium.

If you really want to avoid any suggestion of commercial interest
I think we have to ban all posts from .com addresses, as we never
know if they might try to sell something, not to mention that some
people with .edu or .org addresses might want to sell something.

I think Usenet works and remains, because it didn't have the restrictive
Acceptable Use Policies from the Defense Communications Agency which
took over the ARPAnet, the academic IBM mainframe mentality of BITNET,
the staid academic mentality of CSnet, throw JANET and Alvey in there,
etc. First unix-wizards as a mailing list and the recognition that
govt. and academics use commercial products made the net.* and comp.*
groups successful as a model to others (Gore info-superhighway, world
experts comments aside). Usenet is an optional protocol. An admin
choses to take it even if it's a protocol recognized by practically all
web browsers. Similarly Gilmore created the alt.* hierarchy for even
more complaints (you can't really complain with effect there).

And you have to use all tools. Or complain locally for them.
Personally, I consider the 'cost' of reading a post compared to
the possible advantage to readers for anything that might possibly
be considered commercial. For some rare items some newsgroups
might be the most appropriate place to find people, not that others
don't abuse the privilege.
;^)

The $200 educational/non-profit price seems very reasonable to me,
though I don't expect to actually buy it. Even for the $800 commercial
license, I don't expect them to get rich. Note that it is spec.ORG.

8^)
--
 

Ask a Question

Want to reply to this thread or ask your own question?

You'll need to choose a username for the site, which only take a couple of moments. After that, you can post your question and our members will help you out.

Ask a Question

Members online

Forum statistics

Threads
473,744
Messages
2,569,484
Members
44,903
Latest member
orderPeak8CBDGummies

Latest Threads

Top