[Cbe-oss-dev] SPU-GCC optimzation problem

[Ulrich Weigand]

Goffredo,

thanks for pointing this out, but this link discusses whether the 
instruction as specified is particularly useful or not -- not that it 
doesn't *work* as specified.   The bug reported by Machida-san is a real 
compiler bug where the code actually does *not* work as specified in the 
Language Extensions document.

Mit freundlichen Gruessen / Best Regards

Ulrich Weigand

-- 
  Dr. Ulrich Weigand | Phone: +49-7031/16-3727
  GNU compiler/toolchain for Linux on System z and Cell BE
  IBM Deutschland Entwicklung GmbH
  Vorsitzender des Aufsichtsrats: Martin Jetter | Geschäftsführung: 
Herbert Kircher
  Sitz der Gesellschaft: Böblingen | Registergericht: Amtsgericht 
Stuttgart, HRB 243294



"Goffredo Marocchi" <panajev at gmail.com> 
Sent by: cbe-oss-dev-bounces+ulrich.weigand=de.ibm.com at ozlabs.org
03/11/08 04:44 PM

To
feaf5ed90803110210t54f42bfcge8edfcc26705c258 at mail.gmail.com
cc
cbe-oss-dev at ozlabs.org
Subject
[Cbe-oss-dev]  SPU-GCC optimzation problem






I think this link (looking at the ASM instruction you highlight... not the 
same exact instruction as yours, but it should be the same problem, 
including the negative argument for the rotate instruction) might help 
you:

http://www.insomniacgames.com/tech/articles/0807/why_rotqmbybi_is_broken.php


Regards,

Panajev_______________________________________________
cbe-oss-dev mailing list
cbe-oss-dev at ozlabs.org
https://ozlabs.org/mailman/listinfo/cbe-oss-dev