Thread: [GD-Linux] Intel's C++ compiler
Brought to you by:
vexxed72
Menu
▾
▴
gamedevlists-linux
|
From: Daniel V. <vo...@ep...> - 2001-11-06 23:07:14
|
Has anyone tried Intel's C++ compiler for Linux? I have some template code here that refuses to compile with any version of gcc >= 2.96 and am wondering what people's experience are with the Intel compiler with regard to C++ conformance (according to the German iX magazine it's supposed to be good)? - Daniel Vogel, Programmer, Epic Games Inc. |
|
From: Timothee B. <tt...@id...> - 2001-11-06 23:34:21
|
I think you have the first post here Daniel :-) I've been willing to, but haven't looked at Intel's C++ compiler for Linux yet .. but I hear they have a 30 days demo available. Anyone else? Is your template code related to STL, or fully custom? Do you know what exactly are the template features missing in gcc? I believe the official stable version of gcc is still a 2.95.X anyway? TTimo Not sure how many are subscribed to the list actually, this is a good time for a ping.. PS: SourceForge's new email policy of setting reply to to the poster blows. On Tue, 6 Nov 2001 18:05:47 -0500 "Daniel Vogel" <vo...@ep...> wrote: > Has anyone tried Intel's C++ compiler for Linux? I have some template code > here that refuses to compile with any version of gcc >= 2.96 and am > wondering what people's experience are with the Intel compiler with regard > to C++ conformance (according to the German iX magazine it's supposed to be > good)? > > - Daniel Vogel, Programmer, Epic Games Inc. > > > _______________________________________________ > Gamedevlists-linux mailing list > Gam...@li... > https://lists.sourceforge.net/lists/listinfo/gamedevlists-linux |
|
From: Daniel V. <vo...@ep...> - 2001-11-06 23:54:20
|
> I think you have the first post here Daniel :-) :) > I've been willing to, but haven't looked at Intel's C++ compiler for Linux > yet .. but I hear they have a 30 days demo available. Anyone else? I'm concerned about the time involved to get it up and running just to realize that there are either the same or different issues. > Is your template code related to STL, or fully custom? Do you know what Custom - nothing fancy just some very basic template usage. In one case I had to explicitely instantiate a template to make it compile and in another case no mojo worked to get it to compile with gcc >= 2.96. Also gcc >= 3.0 now warns about multi-line literals being deprecated and I couldn't find a way to turn off that warning - the code in question isn't even compiled (#ifdef'ed out for another platform) and apparently the warning in this case is by design... sigh. 2.95.3 (couldn't use 2.95.2 due to glibc incompatibilities) compiled the code fine except one occurance I solved by using explicit instantiation. > exactly are the template features missing in gcc? I believe the official > stable version of gcc is still a 2.95.X anyway? Yes but 2.95.x doesn't support 2 byte Unicode :) > PS: SourceForge's new email policy of setting reply to to the poster > blows. Yup. - Daniel Vogel, Programmer, Epic Games Inc. |
|
From: Bert F. <be...@is...> - 2001-11-07 08:06:11
|
On Wed, 7 Nov 2001, Timothee Besset wrote: > I've been willing to, but haven't looked at Intel's C++ compiler for Linux > yet .. but I hear they have a 30 days demo available. Anyone else? Yep, see http://developer.intel.com/software/products/compilers/c50/linux/ The German c't magazin had a review this week. They were quite pleased, although it doesn't compile the kernel, yet. > Not sure how many are subscribed to the list actually, this is a good time > for a ping.. Pong :) > PS: SourceForge's new email policy of setting reply to to the poster > blows. AWGTHTGTTA? http://www.unicom.com/pw/reply-to-harmful.html -- Bert |
|
From: Hal A. <ha...@cl...> - 2001-11-07 08:21:48
|
I have agreed to the 30-day demo. I've even downloaded it....I just haven't actually compiled anything with it yet. Sorry, just adding to the noise to signal ratio :) Hal. -----Original Message----- From: gam...@li... [mailto:gam...@li...]On Behalf Of Bert Freudenberg Sent: 07 November 2001 08:02 To: gam...@li... Subject: Re: [GD-Linux] Intel's C++ compiler On Wed, 7 Nov 2001, Timothee Besset wrote: > I've been willing to, but haven't looked at Intel's C++ compiler for Linux > yet .. but I hear they have a 30 days demo available. Anyone else? Yep, see http://developer.intel.com/software/products/compilers/c50/linux/ The German c't magazin had a review this week. They were quite pleased, although it doesn't compile the kernel, yet. > Not sure how many are subscribed to the list actually, this is a good time > for a ping.. Pong :) > PS: SourceForge's new email policy of setting reply to to the poster > blows. AWGTHTGTTA? http://www.unicom.com/pw/reply-to-harmful.html -- Bert _______________________________________________ Gamedevlists-linux mailing list Gam...@li... https://lists.sourceforge.net/lists/listinfo/gamedevlists-linux |
|
From: Ryan C. G. <ic...@cl...> - 2001-11-07 08:24:36
|
> AWGTHTGTTA? http://www.unicom.com/pw/reply-to-harmful.html I am dying to know what "AWGTHTGTTA" means. --ryan. |
|
From: Daniel V. <vo...@ep...> - 2001-11-12 07:34:45
|
I installed the evaluation today and I'm really impressed. The compile speed is amazing and I haven't even played with precompiled headers yet! It only took me about half an hour to convert from gcc to icc but we have a nice abstraction layer and don't use non ANSI C++ code or weird template stuff so YMMV. - Daniel Vogel, Programmer, Epic Games Inc. > -----Original Message----- > From: gam...@li... > [mailto:gam...@li...]On Behalf Of > Daniel Vogel > Sent: Tuesday, November 06, 2001 6:06 PM > To: gam...@li... > Subject: [GD-Linux] Intel's C++ compiler > > > Has anyone tried Intel's C++ compiler for Linux? I have some template code > here that refuses to compile with any version of gcc >= 2.96 and am > wondering what people's experience are with the Intel compiler with regard > to C++ conformance (according to the German iX magazine it's > supposed to be > good)? > > - Daniel Vogel, Programmer, Epic Games Inc. > > > _______________________________________________ > Gamedevlists-linux mailing list > Gam...@li... > https://lists.sourceforge.net/lists/listinfo/gamedevlists-linux > |
|
From: Joe V. <jva...@li...> - 2001-11-12 14:26:39
|
On Mon, Nov 12, 2001 at 02:32:15AM -0500, Daniel Vogel wrote: > I installed the evaluation today and I'm really impressed. The compile speed > is amazing and I haven't even played with precompiled headers yet! It only > took me about half an hour to convert from gcc to icc but we have a nice > abstraction layer and don't use non ANSI C++ code or weird template stuff so > YMMV. Does it have its own linker? If not what are the link times like? -- J. Valenzuela -- jva...@li... Only those who leisurely approach that which the masses are busy about can be busy about that which the masses take leisurely. -- Lao Tsu |
|
From: Daniel V. <vo...@ep...> - 2001-11-13 04:05:57
|
> Does it have its own linker? If not what are the link times like? Yes (or I at least think it does have it's own linker). It feels a bit faster than with gcc but linking time never was an issue with the Unreal engine as opposed to *cough*SC3K*cough* ;-) - Daniel Vogel, Programmer, Epic Games Inc. |
|
From: Daniel V. <vo...@ep...> - 2001-11-13 07:35:24
|
Nope, I'm wrong. It seems to use ld for linking. - Daniel Vogel, Programmer, Epic Games Inc. > -----Original Message----- > From: gam...@li... > [mailto:gam...@li...]On Behalf Of > Daniel Vogel > Sent: Monday, November 12, 2001 11:04 PM > To: gam...@li... > Subject: RE: [GD-Linux] Intel's C++ compiler > > > > Does it have its own linker? If not what are the link times like? > > Yes (or I at least think it does have it's own linker). It feels a bit > faster than with gcc but linking time never was an issue with the Unreal > engine as opposed to *cough*SC3K*cough* ;-) > > - Daniel Vogel, Programmer, Epic Games Inc. > > > _______________________________________________ > Gamedevlists-linux mailing list > Gam...@li... > https://lists.sourceforge.net/lists/listinfo/gamedevlists-linux > |
|
From: Daniel V. <vo...@ep...> - 2001-11-13 08:57:00
|
icc is well documented but for those who don't want to browse the 361 page pdf here is a small summary of differences gcc: 1. -export-dynamic 2. -fPIC -shared -Wl,-soname,$SONAME 3. #if __GNUG__ icc: 1. -Qoption,link,--export-dynamic 2. -KPIC -shared -Qoption,link,-soname,$SONAME 3. #if __ICC BTW, icc supports #pragma pack and -Zp[n] which should make porting from Windows easier :) - Daniel Vogel, Programmer, Epic Games Inc. > -----Original Message----- > From: gam...@li... > [mailto:gam...@li...]On Behalf Of > Daniel Vogel > Sent: Tuesday, November 13, 2001 2:34 AM > To: gam...@li... > Subject: RE: [GD-Linux] Intel's C++ compiler > > > Nope, I'm wrong. It seems to use ld for linking. > > - Daniel Vogel, Programmer, Epic Games Inc. > > > -----Original Message----- > > From: gam...@li... > > [mailto:gam...@li...]On Behalf Of > > Daniel Vogel > > Sent: Monday, November 12, 2001 11:04 PM > > To: gam...@li... > > Subject: RE: [GD-Linux] Intel's C++ compiler > > > > > > > Does it have its own linker? If not what are the link times like? > > > > Yes (or I at least think it does have it's own linker). It feels a bit > > faster than with gcc but linking time never was an issue with the Unreal > > engine as opposed to *cough*SC3K*cough* ;-) > > > > - Daniel Vogel, Programmer, Epic Games Inc. > > > > > > _______________________________________________ > > Gamedevlists-linux mailing list > > Gam...@li... > > https://lists.sourceforge.net/lists/listinfo/gamedevlists-linux > > > > _______________________________________________ > Gamedevlists-linux mailing list > Gam...@li... > https://lists.sourceforge.net/lists/listinfo/gamedevlists-linux > |
|
From: Daniel V. <vo...@ep...> - 2001-11-14 09:08:22
|
Does anyone know what I have to do get the printf in the below code example
to be executed on Linux? Our current approach to achieve that uses signal
handlers, macros and setjmp which is ugly, slow and an overall PITA.
#include <stdio.h>
int main()
{
try
{
int* pointer = NULL;
int crash = *pointer;
}
catch( ... )
{
printf("I want to be executed\n");
}
return 0;
}
- Daniel Vogel, Programmer, Epic Games Inc.
|
|
From: Ryan C. G. <ic...@cl...> - 2002-01-10 01:59:17
|
...sucks. I've been plagued for a long time with three debugging problems: 1) When threads terminate, for example, when calling SDL_CloseAudio() (which is also called by SDL_Quit()), GDB reports that the "program terminated normally," and stops debugging, when in fact the remaining thread should have continued. Not sure what happened to that remaining thread, except that perhaps gdb killed it itself in a fit of confusion. This is mildly annoying during shutdown on Serious Sam, since there's more to clean up after the audio subsystem goes down, but it's downright infuriating when changing a cvar in-game that wants to restart the audio subsystem. 2) GDB lets you use TAB to auto-complete symbol names. Many people complained that the code for walking the symbol list in gdb 4.18 is too slow, and it is, if you work on something that has as many symbols as, say, Mozilla. The GDB maintainers rewrote it to be lightning fast in version 5.0, but be careful; some autocompletion efforts make gdb go into an infinite loop. It locks hard, and you have to hit CTRL-Z to suspend the process and then kill it with a -9. That's nasty. It happens to me way more than it should, and usually at a really critical point where you can't afford to kill the process and start over, too. 3) awatch doesn't work. I think I mentioned this before. This is a very helpful function if you are working in a big project and find yourself going "when does THAT variable get used". It's also very helpful if you know that a variable is getting hit by memory corruption, and you want to know exactly when it happens so you can fix it. Doesn't work at all. I have fixed all these problems today. On the advice of one of the XMMS developers, I downgraded. Take the stock gdb-4.18.tar.gz from GNU's website, and apply this patch (so that it supports Linux thread signals, etc): http://sources.redhat.com/ml/gdb-patches/1999-q3/msg00177.html And, poof! All three problems go away. Actually, the infinite loop during some symbol auto-completions is still there, but in 4.18 you can hit CTRL-C and it will give up, putting you back at the gdb command prompt where you can continue debugging or try the symbol lookup again. Best decision I've made all year (er...all nine days of it), and I would advise anyone working on Linux to do the same. I was using gdb 5.0, and was told that 5.1 has the same problems. I can't think of a single benefit in 5.0 that I've lost by downgrading. I'll have to check the changelog. --ryan. |
×
Want the latest updates on software, tech news, and AI?
Get latest updates about software, tech news, and AI from SourceForge directly in your inbox once a month.
Thanks for helping keep SourceForge clean.
X