This is the mail archive of the
mailing list for the GCC project.
Re: PATCH (SPU): Add enhanced Double Precision (eDP) support on SPU
- From: trevor_smigiel at playstation dot sony dot com
- To: Sa Liu <SALIU at de dot ibm dot com>
- Cc: gcc-patches at gcc dot gnu dot org, Steven Bosscher <stevenb dot gcc at gmail dot com>, Ulrich Weigand <Ulrich dot Weigand at de dot ibm dot com>, Ira Rosen <IRAR at il dot ibm dot com>
- Date: Tue, 10 Jul 2007 14:52:43 -0700
- Subject: Re: PATCH (SPU): Add enhanced Double Precision (eDP) support on SPU
- References: <20070419030901.GH30047@playstation.sony.com> <OFBA12729E.B3DD14C0-ONC12572F3.0060806C-C12572F3.0062D0D7@de.ibm.com>
Sorry for the delay in replying.
Is there a reason you didn'y make this change?
> Can you rewrite the celledp handling in spu_emit_branch_or_set. Instead
> of special casing in there, move all special casing into a
> (define_expand "ceq_df") and (define_expand "cgt_df").
* Sa Liu <SALIU@de.ibm.com> [2007-06-07 11:00]:
> > Can you rewrite the celledp handling in spu_emit_branch_or_set. Instead
> > of special casing in there, move all special casing into a
> > (define_expand "ceq_df") and (define_expand "cgt_df"). This patch
> > includes an implementation of ceq_v2df and cgt_v2df, can you make scalar
> > versions of those to use in those expands? Preferably using mode macros
> > to avoid duplication, but that's not a big deal. Test cases would be
> > nice.
> > Except for the last suggestion and the test cases, I've done most of
> > these changes in the IBM CELL SDK sources. You might want to grab that
> > to compare with.
> > Trevor
> Sorry for the delay. Attached is a new version of this patch.
> In this patch the scalar and vector version of ceq_v2df_celledp and
> cgt_v2df_celled are combined using mode macros.
> A test case intrinsics-2.c is added to test spu intrinsic functions on
> Two test cases dfcmeq.c and dfcmgt.c are used to test generation of dfcmeq
> and dfcmgt instructions combined from double precision compare and
> absolute operations.
> In addition, vcond patterns are added to support vector conditional
> expression. Test case fast-math-vect-reduc-7.c is switched on to test
> vcond pattern.