This is the mail archive of the
mailing list for the GCC project.
Re: optab handler for floating point to interger "fix" is not generated
- From: Segher Boessenkool <segher at kernel dot crashing dot org>
- To: David Kang <dkang at isi dot edu>
- Cc: GCC <gcc at gcc dot gnu dot org>
- Date: Tue, 25 Nov 2014 14:25:47 -0600
- Subject: Re: optab handler for floating point to interger "fix" is not generated
- Authentication-results: sourceware.org; auth=none
- References: <20141125182138 dot GA29715 at gate dot crashing dot org> <1929286565 dot 1075249 dot 1416941106326 dot JavaMail dot root at zm dot isi dot edu>
On Tue, Nov 25, 2014 at 10:45:06AM -0800, David Kang wrote:
> Is the logical operators described similarly?
> I tried "seqf2", "cmpseqf2", and "one_cmplseqf2" for "eq" of two floating point numbers as it is shown below.
> But none of them work.
> (define_expand "seqf2"
> [(set (match_operand:SI 0 "register_operand" "")
> (eq:SI (match_operand:SF 1 "register_operand" "")
> (match_operand:SF 2 "register_operand" "")))]
This is done via cstoresf4. Please see the `gccint' manual, available
online or via e.g. `make pdf'; or read gcc/doc/md.texi directly.