This is the mail archive of the mailing list for the GCC project.

Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]
Other format: [Raw text]

Re: [PATCH] ARM half-precision floating point, 5/8 (detect constant overflow)

Andrew Pinski wrote:
On Sun, Apr 19, 2009 at 9:04 PM, Sandra Loosemore
<> wrote:
It looks like carrying out this level of testing would be a non-trivial
amount of effort.  I have discovered that, while we do in fact have some
Cell hardware at CodeSourcery, we've never tried to build or test an SPU
compiler before. And, as far as I can tell, the GNU SPU tools treat this as
a bare-metal/ELF/newlib target; can you even bootstrap GCC on this machine
at all?

You cannot bootstrap on the spu target but you can build a cross compiler and just run the testsuite like you would run a native compiler (as long as the spu-elf runner is turned on which normally is running). This is the way I test the spu-elf target and may other folks do (they also run the gdb testsuite this way).

I'm still really struggling with this. (I'm a guest on this Cell machine, which is used for testing on another project; it's not set up as a development machine and I keep finding that it's lacking bits of installed software required to build and test GCC, plus I'm not even sure I've got the right configure options for this target.) Is it possible that someone who already has a SPU build environment set up could test this patch for me? Otherwise it's likely to be blocked indefinitely while I thrash around.


Index Nav: [Date Index] [Subject Index] [Author Index] [Thread Index]
Message Nav: [Date Prev] [Date Next] [Thread Prev] [Thread Next]