On Mon, Mar 02, 2026 at 01:07:36PM +0100, Jakub Jelinek wrote:
> On Mon, Mar 02, 2026 at 08:04:53PM +0800, Hongtao Liu wrote:
> > You are correct. There is no place that calls
> > gen_avx512fp16_mov{v8hf,v8bf,v8hi}. The original pattern‘s name is
> > avx512fp16_vmovsh which is added in r12-3407-g9e2a82e1f9d2c4, there's
> > also another pattern named *avx512fp16_movsh . At that time, the * was
> > added to distinguish between these two patterns.
> > And yes, we can add* to the pattern name.
>
> Will test that tonight.
Here it is.
Bootstrapped/regtested on x86_64-linux and i686-linux, ok for trunk?
2026-03-03 Jakub Jelinek <[email protected]>
* config/i386/sse.md (avx512fp16_mov<mode>): Rename pattern to...
(*avx512fp16_mov<mode>): ... this.
--- gcc/config/i386/sse.md.jj 2026-03-02 10:45:52.035766839 +0100
+++ gcc/config/i386/sse.md 2026-03-02 14:48:33.787524473 +0100
@@ -12435,7 +12435,7 @@ (define_insn_and_split "*vec_setv2di_0_z
DONE;
})
-(define_insn "avx512fp16_mov<mode>"
+(define_insn "*avx512fp16_mov<mode>"
[(set (match_operand:V8_128 0 "register_operand" "=v,v")
(vec_merge:V8_128
(match_operand:V8_128 2 "register_operand" "v,v")
Jakub
