Become a MacRumors Supporter for $50/year with no ads, ability to filter front page stories, and private forums.
Compiler Cost

I can't imagine Apple would supply the compiler for free, but it sure would be nice to get a big discount on it. I have been using CodeWarrior since it first came out, and though it is still better overall than XCode, since Motorola bought CodeWarrior it seems like they are losing their focus on Macs and going the way of Symantec's compilers for Mac.
 
get excited

Originally posted by macshark
I think some of us are getting a bit too excited about this. I have no doubt that IBM's new compilers will increase the SPEC results, etc. for the G5, however, I wouldn't expect huge performance increases for most real world applications. Despite being free, gcc is still a very good compiler and many people have been working on G4/G5 optimizations for gcc for several years.

Everyone should get REALLY DAMN EXCITED over this. :)

people I work with have seen ~35% speed increases on their code. IBM has a PDF up of XLC/XLF v. GCC/F77 compiled SPECfp running on a Power4 (the daddy of the G5). The results were, on average, about twice as fast with XL.

If 35% was a typical speed gain on the G5, It'd be the difference between a dual 2GHz and a dual 2.7GHz. That's something to get excited about IMHO.

As for gcc, most people don't think that much about its ability to generate fast code. On x86, Intel's compiler generates much faster code. In fact, gcc still isn't significantly optimised for the P4.
The group in charge of gcc development, as a rule, steers the project with the ultimate goal of producing a well rounded compiler. They don't, by design, provide extensive optimization for any one architecture. Right now, GCC 3.3 is aware of the 970 but it hasn't really been optimized for it yet. gcc 3.3 actually treats the 970 as a different Power processor because it tends to product better code than if it treated it as a G4.
 
Small Comment on IBM Compilers and GCC

:rolleyes: i think that it has two sides

ibm's compiler might not be VERY expensive but for medium and small size developers it is just too expensive :mad:

apple for sure can pay for the liscense and the feeling of using mac os x will get even faster (or stay the same fast but will get more enchanted) with this new compiler.

even if there is only a 10% speed increase in real world performance.


though its kind of bad for gcc because apple and ibm for sure put some work in the gcc compiler to make it fit and work for the ppc 970

this work wont be continued like before ,i guess.

and that is bad. :(


someone knows if the compiler is 970 specific and runs under linux as well?

if that is the case we MIGHT see the ibm 970 compiler build into xcode, because then its a general development for ibms low cost linux servers.

if not - well then we can hope for xcode implementation for edu and private use...

better than nothing.
 
Re: Small Comment on IBM Compilers and GCC

Originally posted by ionas
ibm's compiler might not be VERY expensive but for medium and small size developers it is just too expensive :mad:
I'd have to differ here. And yes, I help run a software shop. $500/seat may be too expensive for hobbyists - who generally don't need anything over gcc. For any commercial shop, pretty much of any size, getting a 35% speed boost in exchange for $500 and 3-4 hours of time to integrate it is a sensational deal.

-Richard
 
hmmm

does the compiler support server/client options (can you compile via tcpip)?

if you work in a 6 ppl software company you ll only have to buy one liscense then?
 
Re: Re: Rebuild of Panther?

Originally posted by macrumors12345
Unless the XL compilers can be integrated into the XCode IDE, I seriously doubt that you will see Apple using them anytime soon. Let's be realistic here.
Oh, but they can...:)
 
Originally posted by ITR 81
I wonder why anyone would vote negative on this??

It happened to me once. I didn't read it all and thought that I was going to see negative comments about the subject. I didn't actually know that I was voting.
 
Re: Objective-C Compatibility

Originally posted by MattMass
But, does anyone know if/when these compilers will support Obj-C? Without this functionality I won't be able to use them myself, and I suspect that many developers will be in the same boat. [/B]

From the website linked in the article:

Technical preview — Selected functions

The following selected functions are offered in this initial release of XL C/C++ Advanced Edition V6.0 for Mac OS X, as a "technical preview." Technical Previews provide insight into IBM plans and directions and information and functions are provided "as-is" and without warranty or condition of any kind. Availability of the selected functions will be announced at a later date.
OpenMP
.
.
.

Objective-C
XL C/C++ Advanced Edition compiler supports Objective-C.

Objective-C is an object-oriented programming language based on standard C and is popular in the Mac OS X programming environment. The Objective-C programming is a basis for writing to Cocoa frameworks.
 
Considering that Apple has already been philanthropic with their developer tools compared to the $3000+ they cost from NeXT, I doubt they're going to pay IBM and give us the XL compilers for free.

A discount would be appreciated and we could say goodbye to MetroWerks (and their arrogance) forever.
 
Originally posted by bousozoku
A discount would be appreciated and we could say goodbye to MetroWerks (and their arrogance) forever.

I don't remember Metrowerks being arrogant until after Motorola bought them. I used to get phone calls from Metrowerks people regularly asking about how things were going, suggestions, etc. That all stopped when they were bought out. I made some money, not a lot, on their stock that I owned, but I would gladly trade that back to prevent their deterioration.
 
Re: Small Comment on IBM Compilers and GCC

Originally posted by ionas
:rolleyes: i think that it has two sides
ibm's compiler might not be VERY expensive but for medium and small size developers it is just too expensive :mad:
I think that anyone who can't afford $500 bucks to develop software probably doesn't need anything better than gcc and xCode (which is a very nice overall package).
though its kind of bad for gcc because apple and ibm for sure put some work in the gcc compiler to make it fit and work for the ppc 970.
this work wont be continued like before ,i guess.
Apple and IBM both have reason to continue adding to GCC. They both support open source projects (darwin and linux). Furthermore, Apple will continue to ship GCC with xCode.. you just can't beat the price and it's in Apple's interest to ship free development tools to as many people as possible.

someone knows if the compiler is 970 specific and runs under linux as well?
xlc has support flags for the PPC970 and the G5 specifically. At this point, I think they do pretty similar things but that may change as IBM starts shipping PPC 970 based machines.
There is a linux version of xl which I'm sure will support 970 optimizations since IBM is releasing 970 based Linux blades soon.

if that is the case we MIGHT see the ibm 970 compiler build into xcode, because then its a general development for ibms low cost linux servers.

if not - well then we can hope for xcode implementation for edu and private use...
Built in? Maybe bundled... but xCode won't be released for Linux if that's what your getting at. xCode is a cocoa app. I'm not exactly sure what you're getting at.
There is an edu discount for xl and xl does integrate with xCode. It is even compatible with gcc headers.

:)
 
Re: Re: Small Comment on IBM Compilers and GCC

Originally posted by ffakr
I think that anyone who can't afford $500 bucks to develop software...

The amount is trivial to pay for any decent commercial projects, as they will generate enough revenue to subsidize the cost, but this hurts freeware developers. Sure you might say if I make freeware/opensource software I might not need a different compiler from gcc, but that is not always true. There a few larger freeware projects that would benefit from the speed the XL compilers would provide, and this in the end benefits Mac OS X users. I think Apple should seriously consider bundling XL with Xcode, as this will make the platform much better when all apps run a good deal faster, and therefore generate sales and revenue.
 
Re: Re: Re: Small Comment on IBM Compilers and GCC

Originally posted by kherdin
I think Apple should seriously consider bundling XL with Xcode, as this will make the platform much better when all apps run a good deal faster, and therefore generate sales and revenue. [/B]
Do you realize that this would cost Apple millions? The $500 is just for a single license. Having Apple give it away means that IBM will want compensation for all the licenses they won't sell anymore. Do you really really think they will make all that money back?
 
Originally posted by macshark
I think some of us are getting a bit too excited about this. I have no doubt that IBM's new compilers will increase the SPEC results, etc. for the G5, however, I wouldn't expect huge performance increases for most real world applications. Despite being free, gcc is still a very good compiler and many people have been working on G4/G5 optimizations for gcc for several years.
The other thing is - I bet Apple knows where MacOSX can do with better optimisation. There are certain sections of the OS that are continually used, and areas that are bottlenecks... Apple doesn't just compile code for these sections, they manually optimise the machine code itself.

The 35% boosts are on GCC compiled code. So those hand-optimised bottlenecks etc won't be seeing any improvements. Still, there's got to be huge amounts of the MacOS and applications that will benefit from the overall speed boost.

[edit] err... just to be clear.... no source for my assertions... might not be right.
 
Great News (for me at least)

I'm a developer, so I'll definitely have a use for XLC. I don't have TONS of money (and I don't have a Mac right now), so I won't get it right away. I'm sure, though, that I'll appreciate having it later. Kudos goes out to IBM for developing XLC and reporting the impressive performance gains it produced over GCC.
 
Re: Re: Small Comment on IBM Compilers and GCC

Originally posted by rjstanford

Originally posted by ionas
ibm's compiler might not be VERY expensive but for medium and small size developers it is just too expensive

I'd have to differ here. And yes, I help run a software shop. $500/seat may be too expensive for hobbyists - who generally don't need anything over gcc. For any commercial shop, pretty much of any size, getting a 35% speed boost in exchange for $500 and 3-4 hours of time to integrate it is a sensational deal.

-Richard

I have to agree with ionas here. I am a hobbyist developer and i make shareware games. High framerates are essential and ever since OS9 is gone, it was pretty hard to make this happen. So I want to get my hands on this compiler, but since i don't expect to make more than $500 with my game, how can i afford it?

I think it would be great if there was a $100 to $200 solution, wich i would really consider.

If i am correct the compiler will be out 16 january 2004 (tomorrow). Anyone with a G5 here who wants to run the SPEC benchmark. I am really interested in the new scores.
 
what i like to see recompiled

I am thinking that IBM will probably self come with new SPEC scores for the dual G5. What i like to see is comparison between PowerMac and XServe on the GCC3.3 compiler and the IBM compiler.

Now just for fun a list of things that i want to see recompiled

  • OpenGL
  • All the other frameworks
  • photoshop
  • virtualPC
  • UT2003 and other First Person Shooters
 
Re: Rebuild of Panther?

Originally posted by MikeAtari
Will we have to wait till next year's 10.4 or could Apple release a new build of Panther?

why not 10.3.3?
 
Re: what i like to see recompiled

Originally posted by isgoed
I am thinking that IBM will probably self come with new SPEC scores for the dual G5. What i like to see is comparison between PowerMac and XServe on the GCC3.3 compiler and the IBM compiler.

Now just for fun a list of things that i want to see recompiled

  • OpenGL
  • All the other frameworks
  • photoshop
  • virtualPC
  • UT2003 and other First Person Shooters
Why not everything? ;)
Originally posted by wfzelle
Do you realize that this would cost Apple millions? The $500 is just for a single license. Having Apple give it away means that IBM will want compensation for all the licenses they won't sell anymore. Do you really really think they will make all that money back?
Xcode Pro with a cost. The only diffrence would be the compiler. I think it would be a smart move from Apple...
Originally posted by isgoed
I'd have to differ here. And yes, I help run a software shop. $500/seat may be too expensive for hobbyists - who generally don't need anything over gcc. For any commercial shop, pretty much of any size, getting a 35% speed boost in exchange for $500 and 3-4 hours of time to integrate it is a sensational deal.

-Richard


I have to agree with ionas here. I am a hobbyist developer and i make shareware games. High framerates are essential and ever since OS9 is gone, it was pretty hard to make this happen. So I want to get my hands on this compiler, but since i don't expect to make more than $500 with my game, how can i afford it?

I think it would be great if there was a $100 to $200 solution, wich i would really consider.

If i am correct the compiler will be out 16 january 2004 (tomorrow). Anyone with a G5 here who wants to run the SPEC benchmark. I am really interested in the new scores.
Well... It would be an investment for the future.. So why not go for it? All your future games would benefit, and your sales would probably be up anyway. You could also recompile your old games, do some re-advertisement and see what happens.
 
Re: Re: Re: Small Comment on IBM Compilers and GCC

Originally posted by isgoed
I have to agree with ionas here. I am a hobbyist developer and i make shareware games. High framerates are essential and ever since OS9 is gone, it was pretty hard to make this happen.
Well, that has a lot to do with graphics performance as well, but anyway, its still not that much money for good software.

Besides, aren't you (and your customers) already, by definition, people who are willing to pay hundreds of dollars more for higher quality products (ie: Macs in general)? When the point that Macs cost more than Windows machines is brought up, everyone is so quick to defend that pricing - you get what you pay for and all that. Why should this be any different than the Mac concept itself?

And, if you look at it purely finanancially, either the faster graphics, etc, are worth $500 collectively to your customers (in which case, buy it), or they're not that big a deal (in which case, don't). Either way, nobody's hurting.

-Richard, sticking by his claim
 
Think about $500

$500 is $1 of overhead for a product that only ships 500 copies. $0.10 for a product that ships only 5000 copies. If your shareware is compelling and it ships just 500-1000 [licensed] copies, could you justify a $0.50 - $1.00 increase per copy?
edit: arg, I can't multiply.. fixed.

xlc and xlf for AIX run over $2000 per seat per compiler. I've heard recently that xlf is $2,9xx per seat on AIX! Mac users are already getting a huge discount. This is likely due to the fact that there is a larger market, but it also may have something to do with Apple's involvement. There is also an Edu discount though I've not gotten pricing yet (got the number, didn't call yet).

xCode with gcc is a very nice package. One argument was that shareware game developers need the fastest compilers but don't have the $$. This isn't any different from other platforms. Intel's compiler isn't free either Intel's C++ compiler for Windows is $399 and they have a much larger market.

Other development environments are expensive, but people don't complain about the option of using them if they choose. What does Metrowerks go for these days? Around $500.

We'd all love to have the best possible compiler (or car, or TV, or Computer) but we don't have a perfect redistribution of wealth so some people have to use GCC and some have the ferarri in the garage. :-(
I totally understand that some shareware developers want this but are short on $$. I just want to point out that we shouldn't poo-poo the release of xlc just because we can't all afford it right now.

on a slightly different note, I love the idea of Apple bundleing them and selling an xCode eXtreme or some such thing. :)
 
Recompiled Software

Don't expect big performance improvements with performance-critical commercial software by recompiling it with these new compilers. Most of this stuff (mp3-/divx-codecs, Photoshop, 3D-software, Quartz-drawing and other critical parts of Mac OS X) is optimally hand-tuned already.

In addition, Apple has been working with these compilers for months. Even though they didn't release any XLC-compiled software yet -- at least not officially --, I'm confident they have been using it to learn from the generated binary code.

But this release is still very, very good news for software developers. With good optimizing compilers, you can in many cases achieve almost optimal results without getting your hands dirty (assembler coding, that is). Many projects just don't have the resources for optimization efforts like Apple or Adobe do.

BTW, in my own projects (numerics), XL C++ beta gave me a minus 5% to plus 30% performance increase over gcc.
 
Re: Recompiled Software

Originally posted by leo
Don't expect big performance improvements with performance-critical commercial software by recompiling it with these new compilers. Most of this stuff (mp3-/divx-codecs, Photoshop, 3D-software, Quartz-drawing and other critical parts of Mac OS X) is optimally hand-tuned already.

In addition, Apple has been working with these compilers for months. Even though they didn't release any XLC-compiled software yet -- at least not officially --, I'm confident they have been using it to learn from the generated binary code.

But this release is still very, very good news for software developers. With good optimizing compilers, you can in many cases achieve almost optimal results without getting your hands dirty (assembler coding, that is). Many projects just don't have the resources for optimization efforts like Apple or Adobe do.

BTW, in my own projects (numerics), XL C++ beta gave me a minus 5% to plus 30% performance increase over gcc.

I'm one of those small developers who can't be bothered to hand-optimize code. I'd much rather buy XLC than write PowerPC assembly. (I can do nothing until I get a Mac, though!)
 
Register on MacRumors! This sidebar will go away, and you'll see fewer ads.