This is the mail archive of the
gcc-patches@gcc.gnu.org
mailing list for the GCC project.
Re: fsel patch for rs6000
- To: dje at watson dot ibm dot com
- Subject: Re: fsel patch for rs6000
- From: Geoff Keating <geoffk at geoffk dot org>
- Date: Fri, 25 May 2001 16:26:20 -0700
- CC: gcc-patches at gcc dot gnu dot org
- References: <200105252302.TAA23876@makai.watson.ibm.com>
- Reply-to: Geoff Keating <geoffk at redhat dot com>
> cc: gcc-patches@gcc.gnu.org
> Date: Fri, 25 May 2001 19:02:22 -0400
> From: David Edelsohn <dje@watson.ibm.com>
>
> Why is the fseldfsf4 pattern now the only one that compares a
> DFmode operand with an SFmode operand?
>
>
> (define_insn "*fseldfsf4"
> [(set (match_operand:SF 0 "gpc_reg_operand" "=f")
> (if_then_else:SF (ge (match_operand:DF 1 "gpc_reg_operand" "f")
> (match_operand:SF 4 "zero_fp_constant" "F"))
> (match_operand:SF 2 "gpc_reg_operand" "f")
> (match_operand:SF 3 "gpc_reg_operand" "f")))]
> "TARGET_PPC_GFXOPT && TARGET_HARD_FLOAT"
> "fsel %0,%1,%2,%3"
> [(set_attr "type" "fp")])
>
>
> The other three variants perform the comparison between two arguments of
> the same mode (although the selected value may have a different mode than
> the comparison. I think this is a cut-and-paste a typo.
Yup, it is. I'll commit this when the testrun on powerpc-eabisim
completes:
2001-05-25 Geoffrey Keating <geoffk@redhat.com>
* config/rs6000/rs6000.md (fseldfsf4): Correct mode of operand 4.
Index: rs6000.md
===================================================================
RCS file: /cvs/gcc/gcc/gcc/config/rs6000/rs6000.md,v
retrieving revision 1.118
diff -p -c -r1.118 rs6000.md
*** rs6000.md 2001/05/21 18:38:25 1.118
--- rs6000.md 2001/05/25 23:24:20
***************
*** 4879,4885 ****
(define_insn "*fseldfsf4"
[(set (match_operand:SF 0 "gpc_reg_operand" "=f")
(if_then_else:SF (ge (match_operand:DF 1 "gpc_reg_operand" "f")
! (match_operand:SF 4 "zero_fp_constant" "F"))
(match_operand:SF 2 "gpc_reg_operand" "f")
(match_operand:SF 3 "gpc_reg_operand" "f")))]
"TARGET_PPC_GFXOPT && TARGET_HARD_FLOAT"
--- 4879,4885 ----
(define_insn "*fseldfsf4"
[(set (match_operand:SF 0 "gpc_reg_operand" "=f")
(if_then_else:SF (ge (match_operand:DF 1 "gpc_reg_operand" "f")
! (match_operand:DF 4 "zero_fp_constant" "F"))
(match_operand:SF 2 "gpc_reg_operand" "f")
(match_operand:SF 3 "gpc_reg_operand" "f")))]
"TARGET_PPC_GFXOPT && TARGET_HARD_FLOAT"
--
- Geoffrey Keating <geoffk@geoffk.org>