Home > Archive > Fortran > June 2004 > Tales for Intel
You are viewing an archived Text-only version of the thread.
To view this thread in it's original format and/or if you want to reply to
this thread please [click here]
|
|
| Gerry Thomas 2004-06-03, 7:24 pm |
| You don't have to port from CVF to IVF: CVF is way and by far a more
robust, complete, and stable product than IVF.
If you go for IMSL 5, forget MKL, most of it is accessible via IMSL and
what's not is already part of IMSL.
--
Ciao,
Gerry T.
| |
|
| Gerry Thomas wrote:
>
> You don't have to port from CVF to IVF: CVF is way and by far a more
> robust, complete, and stable product than IVF.
Indeed, and for the next few years, at least, you'll be better off to
leave that Intel *OUTSIDE*!
| |
| Gerry Thomas 2004-06-29, 8:57 am |
|
"kia" <kia@sparrow.com> wrote in message
news:40DE5E9D.75F4C90C@sparrow.com...
> Gerry Thomas wrote:
>
> Indeed, and for the next few years, at least, you'll be better off to
> leave that Intel *OUTSIDE*!
>
Like Intel .NOT. Inside, sounds good to me, but those (academic scroungers
and scavengers) who get Intel and its software (in addition to free 3rd
party software like IMSL) for free are as happy as hogs in mud as long as
Windows' users are underwriting this charity.
--
You're Welcome,
Gerry T.
______
"A cynic is one who knows the price of everything and the value of
nothing." -- Oscar Wilde.
| |
| Jason Nielsen 2004-06-29, 8:57 am |
| > "kia" <kia@sparrow.com> wrote in message
> news:40DE5E9D.75F4C90C@sparrow.com...
>
> Like Intel .NOT. Inside, sounds good to me, but those (academic scroungers
> and scavengers) who get Intel and its software (in addition to free 3rd
> party software like IMSL) for free are as happy as hogs in mud as long as
> Windows' users are underwriting this charity.
>
IMSL being free software is news to me! I find it strange that you think
that Windows' users are underwriting any charity. Windows is a private
company devoted to its own market share and making sure it keeps it at any
cost. If they are underwriting anything it is part of their own long term
interests and strategy. If this behavior disturbs you as you seem to
indicate perhaps you should stop giving your money to a corporation that
perpetuates these 'charitable' donations....
Cheers,
Jason
| |
| Gerry Thomas 2004-06-29, 8:57 am |
|
"Jason Nielsen" <jdn@cs.sfu.ca> wrote in message
news:Pine.LNX.4.58.0406290221300.13341@janus...
more[color=darkred]
scroungers[color=darkred]
as[color=darkred]
>
> IMSL being free software is news to me! I find it strange that you think
> that Windows' users are underwriting any charity. Windows is a private
> company devoted to its own market share and making sure it keeps it at
any
> cost. If they are underwriting anything it is part of their own long
term
> interests and strategy. If this behavior disturbs you as you seem to
> indicate perhaps you should stop giving your money to a corporation that
> perpetuates these 'charitable' donations....
>
these 'charitable' donations...' are courtesy of Microsoft and while
they're freebees to freeloading recipients, someone is footing the bill and
it's evidently not such cheapskates. How much has the Linux guy put in the
pot for Africa this year or any. Zip, so zip up.
--
You're Welcome,
Gerry T.
______
"A cynic is one who knows the price of everything and the value of
nothing." -- Oscar Wilde.
| |
| Paul Van Delst 2004-06-29, 4:08 pm |
| Jason Nielsen wrote:[color=darkred]
Hello,
I stuck GT in my killfile a couple of w s ago (sorry, Gerry, I couldn't take it anymore
:o) but having used the noncommercial version of the Intel linux compiler for a couple of
ws now, I have to comment on GT's statement above (received via Jason Nielsen's post).
I think it's the linux users who are underwriting the "free" Intel linux compiler. They
may not be paying cash, but they're paying with their time. Some are quite willing to pay
that price... people like myself would rather pay $$ for a product that consistently
works. My experience with the Intel compiler is that when it works it works *really,
Really* well. Compilation and run times on some of my code were blindingly fast, and the
compiler diagnostics pointed out a lot more issues in my code than every other compiler
I've used (PGI, SGI, IBM, Sun). But when it doesn't work, it really doesn't work - failing
on relatively simple code (which all six f90/95 books on my shelf tell me is legal code).
I was about 3 mouse clicks from purchasing a license until someone popped into my office
asking for something in a hurry. No worries, I said, I just need to recompile the code
since I replaced all my pgf90-compiled netCDF libraries with Intel f85 compiled ones. The
code, however, would not compile. Ever other system I've used over the last two years has
had no issues. I grabbed the Intel noncom compiler so I could test it properly over a
couple of months (as opposed to a one-month trial) not so I could revel in the joy of
having screwed Intel over coz I got meself a free compiler. I submitted a question (here
and in the user forum) and got a quick response, but I just don't have the time to debug
that compiler on code that I've been using for 2-3+ years.
And, GT, lest you still think I'm a scungy freeloader, yesterday I bought a Lahey f95
license despite that compiler lacking in some features that the Intel compiler has (e.g.
the "-convert big_endian" is a compiler switch to die for and AFAICT, the Lahey f95 linux
compiler doesn't have an equivalent).
And, to hopefully assuage SteveL or any other Intel folks reading this, let me say I lack
the eloquence to convey the disappointment I felt when the I realised I couldn't use the
latest Intel compiler. Once the Intel compiler has matured a bit more I fully intend to
revisit it.
Apologies for the monologue.
cheers,
paulv
| |
| Jason Nielsen 2004-06-29, 4:09 pm |
| On Tue, 29 Jun 2004, Gerry Thomas wrote:
>
> "Jason Nielsen" <jdn@cs.sfu.ca> wrote in message
> news:Pine.LNX.4.58.0406290221300.13341@janus...
> more
> scroungers
> as
> any
> term
>
> these 'charitable' donations...' are courtesy of Microsoft and while
> they're freebees to freeloading recipients, someone is footing the bill and
> it's evidently not such cheapskates. How much has the Linux guy put in the
> pot for Africa this year or any. Zip, so zip up.
>
These 'charitable donations' as I previously said are not charitable. I
doubt Windows gives a rats ass about helping out anyone one way or
another... well other than how it may at a future date make them more
money. I think their corporate behavior is a testament to this...
How you equate someone who uses Linux and someone who doesn't contribute
to charitable organizations is ridiculous at best ... like much of the
tripe that you spew in your relatively constant diatribes on this mailing
list! Linux is not a product like Windows, it is a collection of programs
supported in no small part by the individuals who donate their time to
create much of the great software under the GNU and OSS banner and those
users that submit bug reports to make it better... this is why Linux fairs
so well and why in my opinion so much excellent freeware exists. The
Intel compiler does not fall under that banner and their free
non-commercial license even if supported by Windows funding... which as
far as I can tell hasn't been established.... is following a much
different agenda than would any open source project. So how much have you
donated to Africa this year.... and making a donation as a tax write off
doesn't really count..... if you care about Africa and its people you
would need to do much more than send a little cash!!
Cheers,
Jason
| |
| Dr Chaos 2004-06-29, 4:09 pm |
| On Tue, 29 Jun 2004 10:55:24 -0400, Steve Lionel <Steve.Lionel@intel.com> wrote:
> On Tue, 29 Jun 2004 09:17:04 -0400, Paul Van Delst <paul.vandelst@noaa.gov>
> wrote:
>
>
> Thanks for the comments. We're aware that we didn't make the best first
> impression, but things are a lot better now. Please do try us again when you
> get a chance. If you have specific issues that you want to check on, feel
> free to send me mail.
I used the non-commercial versions of Intel compilers for a while now.
At the beginning, Intel's compiler was much buggier than others that I
tried (now only moderately buggier). I had the feeling that it was a
SPECmark compiler, designed to correctly compile and quickly execute
up to 10 known programs. Notably the run-time checking facilities
were unusably buggy.
I ended up submitting many confirmed bugs to their support system, including
extracting small test cases from larger codes. The sum value of this
time is larger than any reasonable licensing fee.
It's a fair trade, I think. If Intel wants to donate to me a very fast
pentium motherboard upgrade, I wouldn't say no either.
| |
| Gerry Thomas 2004-06-30, 4:01 am |
|
"Jason Nielsen" <jdn@cs.sfu.ca> wrote in message
news:Pine.LNX.4.58.0406291050030.21773@janus...
> On Tue, 29 Jun 2004, Gerry Thomas wrote:
>
off to[color=darkred]
3rd[color=darkred]
long[color=darkred]
think[color=darkred]
private[color=darkred]
at[color=darkred]
that[color=darkred]
and[color=darkred]
the[color=darkred]
>
> These 'charitable donations' as I previously said are not charitable. I
> doubt Windows gives a rats ass about helping out anyone one way or
> another... well other than how it may at a future date make them more
> money. I think their corporate behavior is a testament to this...
[puelile rant elided]
Neielsen brings to mind Milton's
"HENCE loathed Melancholy
Of Cerberus and blackest midnight born,
In Stygian Cave forlorn
'Mongst horrid shapes, and shreiks, and sights unholy.
Find out som uncouth cell,
Where brooding darknes spreads his jealous wings,
And the night-Raven sings;
There, under Ebon shades, and low-brow'd Rocks,
As ragged as thy Locks,
In dark Cimmerian desert ever dwell."
From just what boulder did you make your verminious emergence? Never mind.
Here's a mental lightweight studying in the actuarial 'science' department
of a 3rd rate institution on a par with Toronto's Rochdale College,
proselytizing at taxpayer's expense on behalf of a marginalized OS whose
primary appeal is to tinkerers, hobbyists, rank amateurs, and other of such
ilk. And he's complaining about Microsoft making money. This dumb dolt
doesn't realize that Microsoft probably employs more actuaries than he's
ever met. Besides, for all his posturing, Neielsen isn't an actuary, and is
not likely to be, J.
Canada has just gone through an acrimonious national election in which
highly profound issues were on the table while Neielsen in his supreme
overt narcissism can only use taxpayer's monies to champion a cause about
which the mass of Canadians couldn't give a muse's turd. Sadly, Neielsen
has neither the intellectual or chronological maturity to comprehend the
irrelevancy of his misdirected obsession. It is here that a thesis advisor
is supposed to intervene but apparently not at sfu where graduate students
like a Neielsen are as loose as a muse, and with even a lesser cerebral
contribution.
We're going to get back to you and sfu on this one: "So how much have you
donated to Africa this year.... and making a donation as a tax write off
doesn't really count..... if you care about Africa and its people you would
need to do much more than send a little cash!"
Happy Canada Day,
Gerry T.
| |
| Jason Nielsen 2004-06-30, 4:01 am |
| > [puelile rant elided]
>
I think you meant puerile 'good' sir! Boy, I really must of hit a cord
with you... I'm very impressed with your ability to track me down.. very
skillful... LOL! I find you very amusing, I can't wait for you to respond
with some more of your witty remarks. I know you won't be able to stop
yourself either which is a trifle  however amusing I find you.
Cheers,
Jason
| |
| Gerry Thomas 2004-06-30, 4:01 am |
|
"Paul Van Delst" <paul.vandelst@noaa.gov> wrote in message
news:cbrra9$k0i$1@news.nems.noaa.gov...
> I stuck GT in my killfile a couple of ws ago (sorry, Gerry, I couldn't
take it anymore
> :o)
That's OK Paul, I don't take it personally, :-))))), but keep in touch, I'm
neither the dullest nor the dumbest on the block, :-)))))).
> but having used the noncommercial version of the Intel linux compiler for
a couple of
> ws now, I have to comment on GT's statement above (received via Jason
Nielsen's post).
Com'on Paul, who are you kidding?
>
> I think it's the linux users who are underwriting the "free" Intel linux
compiler. They
> may not be paying cash, but they're paying with their time. Some are
quite willing to pay
> that price... people like myself would rather pay $$ for a product that
consistently
> works. My experience with the Intel compiler is that when it works it
works *really,
> Really* well. Compilation and run times on some of my code were
blindingly fast, and the
> compiler diagnostics pointed out a lot more issues in my code than every
other compiler
> I've used (PGI, SGI, IBM, Sun). But when it doesn't work, it really
doesn't work - failing
> on relatively simple code (which all six f90/95 books on my shelf tell me
is legal code).
> I was about 3 mouse clicks from purchasing a license until someone popped
into my office
> asking for something in a hurry. No worries, I said, I just need to
recompile the code
> since I replaced all my pgf90-compiled netCDF libraries with Intel f85
compiled ones. The
> code, however, would not compile. Ever other system I've used over the
last two years has
> had no issues. I grabbed the Intel noncom compiler so I could test it
properly over a
> couple of months (as opposed to a one-month trial) not so I could revel
in the joy of
> having screwed Intel over coz I got meself a free compiler. I submitted a
question (here
> and in the user forum) and got a quick response, but I just don't have
the time to debug
> that compiler on code that I've been using for 2-3+ years.
>
There's some merit in your discourse.
> And, GT, lest you still think I'm a scungy freeloader, yesterday I bought
a Lahey f95
> license despite that compiler lacking in some features that the Intel
compiler has (e.g.
> the "-convert big_endian" is a compiler switch to die for and AFAICT, the
Lahey f95 linux
> compiler doesn't have an equivalent).
>
Paul (scungy, new to me, have updated vast repotoir of nasties, thanks) and
this is the rub, you're acting in the best interests of NOAA (more useful
than NASA IMO, but I'm a US outsider, nevertheless, in matters Fortran
we're all legally family, kind-of-sort-of, depends if your on the payroll
of a US domiciled enterprise, all others need not apply even in French can
you believe, which the IBM designate for F2003 found amusing, although only
he knows why) and a mouse click here and there and it's on your lap; in
reality space, that's not how it works, it's not my money so I first get my
cat to go eat my mouse, and if that doesn't work, the deal's done. Do I
hate cats, but there on my tab!
> And, to hopefully assuage SteveL or any other Intel folks reading this,
let me say I lack
> the eloquence to convey the disappointment I felt when the I realised I
couldn't use the
> latest Intel compiler. Once the Intel compiler has matured a bit more I
fully intend to
> revisit it.
>
> Apologies for the monologue.
>
> cheers,
>
> paulv
>
Intel is improving, maybe that's progress. For the long haul, sink or swim,
I'm IN. (Doesn't mean I'll take any shit that comes my way, no surrybob).
--
Ciao,
Gerry T.
|
|
|
|
|