From: Andrew Stubbs <ams@codesourcery.com>
Date: Thu, 22 Sep 2022 11:48:30 +0000 (+0100)
Subject: amdgcn: Add vector integer negate insn
X-Git-Tag: basepoints/gcc-14~3983
X-Git-Url: http://git.ipfire.org/cgi-bin/gitweb.cgi?a=commitdiff_plain;h=bf6b5c74a6f1927174091c73aa51401895ef92f0;p=thirdparty%2Fgcc.git

amdgcn: Add vector integer negate insn

Another example of the vectorizer needing explicit insns where the scalar
expander just works.

gcc/ChangeLog:

	* config/gcn/gcn-valu.md (neg<mode>2): New define_expand.
---

diff --git a/gcc/config/gcn/gcn-valu.md b/gcc/config/gcn/gcn-valu.md
index f708e587f387..00c0e3be1ea4 100644
--- a/gcc/config/gcn/gcn-valu.md
+++ b/gcc/config/gcn/gcn-valu.md
@@ -2390,6 +2390,19 @@
   [(set_attr "type" "vop2,ds")
    (set_attr "length" "8,8")])
 
+;; }}}
+;; {{{ Int unops
+
+(define_expand "neg<mode>2"
+  [(match_operand:V_INT 0 "register_operand")
+   (match_operand:V_INT 1 "register_operand")]
+  ""
+  {
+    emit_insn (gen_sub<mode>3 (operands[0], gcn_vec_constant (<MODE>mode, 0),
+			       operands[1]));
+    DONE;
+  })
+
 ;; }}}
 ;; {{{ FP binops - special cases