|
|
Subscribe / Log in / New account

Stallman on GCC, LLVM, and copyleft

Stallman on GCC, LLVM, and copyleft

Posted Jan 25, 2014 14:23 UTC (Sat) by Cyberax (✭ supporter ✭, #52523)
In reply to: Stallman on GCC, LLVM, and copyleft by robert_s
Parent article: Stallman on GCC, LLVM, and copyleft

Sure, and it's open. You can browse it right here: http://llvm.org/svn/llvm-project/llvm/trunk/lib/Target/NV...

So I stand by my assertion that closed-source compilers are simply a non-issue. They are either pointless or of little interest. Usually both.


to post comments

Stallman on GCC, LLVM, and copyleft

Posted Jan 25, 2014 15:49 UTC (Sat) by deepfire (guest, #26138) [Link] (8 responses)

Is that the whole thing, though, or it's a nvidia.ko-like shim to some secret sauce?

Stallman on GCC, LLVM, and copyleft

Posted Jan 25, 2014 17:22 UTC (Sat) by Cyberax (✭ supporter ✭, #52523) [Link] (7 responses)

It's the whole thing. Though PTX is an intermediary format, not the actual machine code.

Stallman on GCC, LLVM, and copyleft

Posted Jan 25, 2014 23:52 UTC (Sat) by deepfire (guest, #26138) [Link] (3 responses)

Isn't this how you would shim GCC? In fact, the best way to shim it, since this totally insulates the IR-receiver from GCC API volatility.

Stallman on GCC, LLVM, and copyleft

Posted Jan 26, 2014 12:59 UTC (Sun) by deepfire (guest, #26138) [Link] (2 responses)

I obviously messed up GCC with LLVM above, but the point, obviously, still stands.

Which makes me think -- wouldn't _exactly_ same be possible to do with GCC - today?

Can't think of anything to stop it.

Stallman on GCC, LLVM, and copyleft

Posted Jan 26, 2014 20:49 UTC (Sun) by stevenb (guest, #11536) [Link] (1 responses)

It's been discussed at length on GCC mailing lists. A port for PTX exists and probably will be included for the next development cycle (after GCC 4.9).

More information:
http://gcc.gnu.org/ml/gcc/2013-11/msg00410.html
http://gcc.gnu.org/ml/gcc-patches/2013-12/msg00842.html

Stallman on GCC, LLVM, and copyleft

Posted Jan 26, 2014 20:59 UTC (Sun) by deepfire (guest, #26138) [Link]

I'm stumped -- doesn't Stallman seek to prevent exactly this shimming of GCC?

What's the remaining practical difference between restrictions that GCC and LLVM put on extensions then?

Heck, what's the fuss all about?

Stallman on GCC, LLVM, and copyleft

Posted Jan 26, 2014 11:39 UTC (Sun) by robert_s (subscriber, #42402) [Link] (2 responses)

> Though PTX is an intermediary format, not the actual machine code.

Exactly - not usable without nonfree software.

Stallman on GCC, LLVM, and copyleft

Posted Jan 26, 2014 22:26 UTC (Sun) by renox (guest, #23785) [Link] (1 responses)

> > Though PTX is an intermediary format, not the actual machine code.
> Exactly - not usable without nonfree software.

Given that some CPUs are also not usable without non free software (microcode)
I'm not sure that there is really a difference..

s/some/nearly all/

Posted Jan 27, 2014 18:39 UTC (Mon) by Wol (subscriber, #4433) [Link]

n/t

Cheers,
Wol


Copyright © 2024, Eklektix, Inc.
Comments and public postings are copyrighted by their creators.
Linux is a registered trademark of Linus Torvalds