This is the mail archive of the crossgcc@sources.redhat.com mailing list for the crossgcc project.
See the CrossGCC FAQ for lots more information.
Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
---|---|---|
Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |
Other format: | [Raw text] |
Hi, I've made a patch for GCC using soft-float primitives for PPC 4xx taken from the IBM PowerPC Performance Libraries Project (http://www-124.ibm.com/developerworks/projects/ppcperflib/).
The project is licensed with BSD License terms, so I don't know the chances of this patch being merged into the GCC project.
It replaces the soft float emulation of GCC used for PPC targets. I did not run the GCC regression tests and I'm still working on this. But I did ran some of the IEEE 754 floating-point tests (i.e. paranoia) on a GCC 3.3.2 crosstool-0.25 generated toolchain for 405 and everything seemed fine. I do not have access to a 440 right now, but I've modified the config.gcc file so it applies to both 405 and 440.
You can expect an average 300% performance gain for float ops
Sounds great! In order to contribute this to the main gcc tree, both you and IBM would need to fill out a copyright assignment form for this work. IBM has done that in the past for GCC contributions, so perhaps you could contact the IBM'ers that have submitted patches to GCC and see what they say.
Once you have the gcc regression tests passing, you should definitely put together a little web page about this patch and the test results, and post a note with the URL both here and to gcc@gcc.gnu.org. - Dan
------ Want more information? See the CrossGCC FAQ, http://www.objsw.com/CrossGCC/ Want to unsubscribe? Send a note to crossgcc-unsubscribe@sources.redhat.com
Index Nav: | [Date Index] [Subject Index] [Author Index] [Thread Index] | |
---|---|---|
Message Nav: | [Date Prev] [Date Next] | [Thread Prev] [Thread Next] |