From d135c23203827e4f9d7a05e6f3b268d2e724440b Mon Sep 17 00:00:00 2001
From: kyukhin <kyukhin@138bc75d-0d04-0410-961f-82ee72b054a4>
Date: Wed, 24 Sep 2014 08:06:30 +0000
Subject: [PATCH] AVX-512. Add cvtps2 insn patterns.

gcc/
	* config/i386/sse.md
	(define_insn
	"<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode><mask_name>"):
	Add masking.
	(define_insn "fix_truncv8sfv8si2<mask_name>"): Ditto.
	(define_insn "fix_truncv4sfv4si2<mask_name>"): Ditto.



git-svn-id: svn+ssh://gcc.gnu.org/svn/gcc/trunk@215545 138bc75d-0d04-0410-961f-82ee72b054a4
---
 gcc/ChangeLog          | 16 ++++++++++++++++
 gcc/config/i386/sse.md | 30 +++++++++++++++---------------
 2 files changed, 31 insertions(+), 15 deletions(-)

diff --git a/gcc/ChangeLog b/gcc/ChangeLog
index a9707a3..c227481 100644
--- a/gcc/ChangeLog
+++ b/gcc/ChangeLog
@@ -8,6 +8,22 @@
 	    Michael Zolotukhin  <michael.v.zolotukhin@intel.com>
 
 	* config/i386/sse.md
+	(define_insn
+	"<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode><mask_name>"):
+	Add masking.
+	(define_insn "fix_truncv8sfv8si2<mask_name>"): Ditto.
+	(define_insn "fix_truncv4sfv4si2<mask_name>"): Ditto.
+
+2014-09-24  Alexander Ivchenko  <alexander.ivchenko@intel.com>
+	    Maxim Kuznetsov  <maxim.kuznetsov@intel.com>
+	    Anna Tikhonova  <anna.tikhonova@intel.com>
+	    Ilya Tocar  <ilya.tocar@intel.com>
+	    Andrey Turetskiy  <andrey.turetskiy@intel.com>
+	    Ilya Verbin  <ilya.verbin@intel.com>
+	    Kirill Yukhin  <kirill.yukhin@intel.com>
+	    Michael Zolotukhin  <michael.v.zolotukhin@intel.com>
+
+	* config/i386/sse.md
 	(define_c_enum "unspec"): Add UNSPEC_PSHUFHW, UNSPEC_PSHUFLW.
 	(define_insn "<mask_codefor>avx512bw_pshuflwv32hi<mask_name>"): New.
 	(define_expand "avx512vl_pshuflwv3_mask"): Ditto.
diff --git a/gcc/config/i386/sse.md b/gcc/config/i386/sse.md
index d1c399c..287fd11 100644
--- a/gcc/config/i386/sse.md
+++ b/gcc/config/i386/sse.md
@@ -3937,13 +3937,13 @@
 (define_mode_attr sf2simodelower
   [(V16SI "v16sf") (V8SI "v8sf") (V4SI "v4sf")])
 
-(define_insn "<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode>"
+(define_insn "<sse2_avx_avx512f>_fix_notrunc<sf2simodelower><mode><mask_name>"
   [(set (match_operand:VI4_AVX 0 "register_operand" "=v")
 	(unspec:VI4_AVX
 	  [(match_operand:<ssePSmode> 1 "nonimmediate_operand" "vm")]
 	  UNSPEC_FIX_NOTRUNC))]
-  "TARGET_SSE2"
-  "%vcvtps2dq\t{%1, %0|%0, %1}"
+  "TARGET_SSE2 && <mask_mode512bit_condition>"
+  "%vcvtps2dq\t{%1, %0<mask_operand2>|%0<mask_operand2>, %1}"
   [(set_attr "type" "ssecvt")
    (set (attr "prefix_data16")
      (if_then_else
@@ -4031,20 +4031,20 @@
    (set_attr "prefix" "evex")
    (set_attr "mode" "XI")])
 
-(define_insn "fix_truncv8sfv8si2"
-  [(set (match_operand:V8SI 0 "register_operand" "=x")
-	(fix:V8SI (match_operand:V8SF 1 "nonimmediate_operand" "xm")))]
-  "TARGET_AVX"
-  "vcvttps2dq\t{%1, %0|%0, %1}"
+(define_insn "fix_truncv8sfv8si2<mask_name>"
+  [(set (match_operand:V8SI 0 "register_operand" "=v")
+	(fix:V8SI (match_operand:V8SF 1 "nonimmediate_operand" "vm")))]
+  "TARGET_AVX && <mask_avx512vl_condition>"
+  "vcvttps2dq\t{%1, %0<mask_operand2>|%0<mask_operand2>, %1}"
   [(set_attr "type" "ssecvt")
-   (set_attr "prefix" "vex")
+   (set_attr "prefix" "<mask_prefix>")
    (set_attr "mode" "OI")])
 
-(define_insn "fix_truncv4sfv4si2"
-  [(set (match_operand:V4SI 0 "register_operand" "=x")
-	(fix:V4SI (match_operand:V4SF 1 "nonimmediate_operand" "xm")))]
-  "TARGET_SSE2"
-  "%vcvttps2dq\t{%1, %0|%0, %1}"
+(define_insn "fix_truncv4sfv4si2<mask_name>"
+  [(set (match_operand:V4SI 0 "register_operand" "=v")
+	(fix:V4SI (match_operand:V4SF 1 "nonimmediate_operand" "vm")))]
+  "TARGET_SSE2 && <mask_avx512vl_condition>"
+  "%vcvttps2dq\t{%1, %0<mask_operand2>|%0<mask_operand2>, %1}"
   [(set_attr "type" "ssecvt")
    (set (attr "prefix_rep")
      (if_then_else
@@ -4057,7 +4057,7 @@
      (const_string "*")
      (const_string "0")))
    (set_attr "prefix_data16" "0")
-   (set_attr "prefix" "maybe_vex")
+   (set_attr "prefix" "<mask_prefix2>")
    (set_attr "mode" "TI")])
 
 (define_expand "fixuns_trunc<mode><sseintvecmodelower>2"
-- 
2.7.4