New Intel Compiler Released

swsnyder writes "A new toy to play with: new Intel compiler (v7.0) available for download at Intel's site. The Windows version is a time limited eval version, but the Linux version is free for non-commercial use. This version is claimed to be even more GCC-compatible that the last version, though I haven't verified this yet."

Comments

Is there

[CableModemSniper]

an echo in here?

Redundant

[keesh]

"though I haven't verified this yet"

You're posting to slashdot. Of course you haven't verified it.

Kettle?

[Trusty Penfold]

How long does it take to make coffee in that neck of the woods?

someone still hasn't had his coffee...

[iosphere]

Twice in the same morning?

Re:someone still hasn't had his coffee...

[Anonymous Coward]

Thrice actually. Spam, Quarks and This.

Re:someone still hasn't had his coffee...

[sporty]

Emergency bugfix release? :)

If only companies worked this quick...

Re:someone still hasn't had his coffee...

[Anonymous Coward]

And this is at least the third time this morning that Hemos has sent a duplicate post. One had been posted three times!

compatible?

[DrSkwid]

even more GCC-compatible that the last version

shouldn't it be more standards compliant

?

Re:compatible?

[Anonymous Coward]

Well, as long as there already is lots of source out there taking advantage of gcc extensions, beeing gcc compliant is a good thing.
Also, its is very nice if it's "command line" compatible with gcc, that way you don't have to tveak the makefiles (or, horror , tha autoconf/make internals), just because icc has another parameter to do just the same.

Re:compatible?

[LizardKing]

shouldn't it be more standards compliant

Nope, being more compatible with gcc implies it supports more gcc extensions. That's one reason why the Linux kernel used to break so much between gcc versions, as gcc is chock full of neat and not so neat extensions to the C language.

Chris

Re:compatible?

[moodyj]

Wait, you mean to say that gcc EMBRACED and EXTENDED the C standard? God Forbid!

Re:compatible?

[PretzelAvenger]

Yes. they did. The nifty thing is that they actually list exactly what extensions, in their documentation, and if you don't mind YOUR compiler going under GPL, you can actually just use their code, all nicely pre-written for you. :-P

Repost

[IainHere]

I mean, it's still the third article on the developers page.

Thanks for Posting This Update!

[Koos Baster]

Wow. These Intel guys are fast: Three new C++ compilers within 48 hours!
...And the're all binary compatible - at least to each other.

--
The right to be heard does not include the right to be taken seriously

Repeat

[waffle zero]

Intel Releases Compiler Suite 7.0

kicks

[Anonymous Coward]

man, this will rock on my Alpha...
oh.. wait..

Dupe

[Xpilot]

Hasn't this already been posted?

http://developers.slashdot.org/article.pl?sid=02 /1 1/24/001248&mode=thread&tid=118

No way...

[PinkX]

TWO DUPES A DAY!

This reminds me of something...

[PD]

Are you starting an unrelated topic? [ynq]

Subject: q
Distribution:

This program posts news to thousands of machines throughout the entire
civilized world. Your message will cost the net hundreds if not thousands of
dollars to send everywhere. Please be sure you know what you are doing.

Are you absolutely sure that you want to do this? [ny]

Hemos, you should just pick the 'n' option I think.

Actually there have been three dupes today,

[amarodeeps]

...by my counting.

http://developers.slashdot.org/article.pl?sid=02/1 1/25/1559231&mode=thread&tid=118

http://science.slashdot.org/article.pl?sid=02/11/2 5/1226222&mode=thread&tid=134

http://slashdot.org/article.pl?sid=02/11/25/032422 1&mode=thread&tid=111

Don't worry Hemos, we all have days like that. ;) It is interesting stuff at least.

Hemos, you ignorant slut!

[Anonymous Coward]

Hemos, of all the slashdot "editors", you must be the most retarded. Now, who are the dumb fucks that *pay* for this shit?

Either that, or intel paid extra for the "article" to get posted twice.

From the too-busy-pulling-my-pud-to-check-for-duplicates dept.
Score: -1 redundant

GCC Compatable?

[TerryAtWork]

Why not just run GCC?

Re:GCC Compatable?

[pythorlh]

Having some mod points today, I almost modded this as +1 Funny. But then I realised that I don't get the joke. I assume that there's some good reason(s) to run the intel compiler vs gcc. I would guess that it involves specific optimizations for Intel chip(set)s. Am I wrong? Is there more(or less) to it than that? Not being a proffessional programmer, I don't really know.

Re:GCC Compatable?

[Reality Master 101]

GCC's big advantage is portability, not quality of optimization. I wouldn't go so far as to say that GCC sucks, but it's mediocre at best.

Re:GCC Compatable?

[RupW]

I assume that there's some good reason(s) to run the intel compiler vs gcc. I would guess that it involves specific optimizations for Intel chip(set)s. Am I wrong? Is there more(or less) to it than that?

Yeah, that's about it. GCC continues to produce mediocre code for the P4 and Itanium. (The latter's very significant - some reports of 50% slower.) There are efforts to fix this but Intel has a significant lead.

The other advantage is vectorization - Intel's compiler makes much better use of MMX/SSE/SSE2 than GCC does. I don't know how ICC compares to VectorC but I'd guess they'd have the edge back by now.

Re:GCC Compatable?

[WasterDave]

The Intel compiler is a fscking great piece of code. My video compression stuff drops out 20% faster than it does under GCC 3.1. On an AMD chip, and it has specific flags to compile for P4 so god alone knows how fast that would be. All this is before we start to take advantage of the built in vectorisation (that actually sucks a fair bit), or change my source in any way.

Something that's fun, for the geeks out there (what am I doing, this is slashdot) is to turn the optimisation up to full whack but leave the debug symbols in .... Jeez, it does some clever shit.

Dave

The Award for Redundancy Award

[Anonymous Coward]

The day will come when stories and trolls go head-to-head in a deathmatch redundancy competition

Is Intel ICC 7.0 ABI compatible with GCC 3.2 ABI?

[truth_revealed]

The GCC 3.2 C++ compiler is known to have some ABI conformance issues that prevent it from claiming 100% IA32 ABI compability. So the question is: does ICC 7.0 match GCC 3.2's C++ ABI exactly (bug for bug) or does it match the offical IA32 C++ ABI standard (thus rendering it incompatible with GCC 3.2)?

Simple answer

[Anonymous Coward]

Unlike GCC, the Intel compiler doesn't suck.

Re:Is Intel ICC 7.0 ABI compatible with GCC 3.2 AB

[Anonymous Coward]

3.2.1 was released last week, which isn't even 100% compatible with 3.2.0

Open Source: Eventually we'll get it right!

Compile Kernel

[Screaming Lunatic]

Has anyone attempted to compile the kernel with the Intel Compiler?

Re:Compile Kernel

[WasterDave]

It's apparently been done within Intel, but they had to do some wacky patches to make it a goer.

Dave

Re:Compile Kernel

[Anonymous Coward]

I'd be interested if the Intel-compiled kernel was significantly faster than the gcc-compiled one. Anybody know of any benchmarks for this type of thing?

Re:Compile Kernel

[IrvineHosting]

Yeah, it appears ICC generated code is 30-50% faster than code generated by GCC.

check out this benchmark

http://www.coyotegulch.com/reviews/intel_comp/in tel_gcc_bench2.html

Re:Is Intel ICC 7.0 ABI compatible with GCC 3.2 AB

[rhost89]

I would think that it would match GCC bug for bug, if it didnt match those "Features" one for one, then it wouldnt be 100% compatible. Besides, if they wanted to make it 100% IA32 ABI compatible, they would have just written the compiler that way and say its more standards complient then GCC.