qemu-devel.nongnu.org archive mirror
 help / color / mirror / Atom feed
From: Richard Henderson <richard.henderson@linaro.org>
To: qemu-devel@nongnu.org
Subject: [PATCH 1/4] tcg: Don't free vector results
Date: Wed, 30 Aug 2023 19:57:26 -0700	[thread overview]
Message-ID: <20230831025729.1194388-2-richard.henderson@linaro.org> (raw)
In-Reply-To: <20230831025729.1194388-1-richard.henderson@linaro.org>

Avoid reusing vector temporaries so that we may re-use them
when propagating stores to loads.

Signed-off-by: Richard Henderson <richard.henderson@linaro.org>
---
 tcg/tcg-op-gvec.c | 39 ++++++---------------------------------
 1 file changed, 6 insertions(+), 33 deletions(-)

diff --git a/tcg/tcg-op-gvec.c b/tcg/tcg-op-gvec.c
index e260a07c61..f5cfd9bf99 100644
--- a/tcg/tcg-op-gvec.c
+++ b/tcg/tcg-op-gvec.c
@@ -561,7 +561,6 @@ static void do_dup(unsigned vece, uint32_t dofs, uint32_t oprsz,
             tcg_gen_dupi_vec(vece, t_vec, in_c);
         }
         do_dup_store(type, dofs, oprsz, maxsz, t_vec);
-        tcg_temp_free_vec(t_vec);
         return;
     }
 
@@ -1036,8 +1035,6 @@ static void expand_2_vec(unsigned vece, uint32_t dofs, uint32_t aofs,
         fni(vece, t1, t0);
         tcg_gen_st_vec(t1, cpu_env, dofs + i);
     }
-    tcg_temp_free_vec(t0);
-    tcg_temp_free_vec(t1);
 }
 
 /* Expand OPSZ bytes worth of two-vector operands and an immediate operand
@@ -1059,8 +1056,6 @@ static void expand_2i_vec(unsigned vece, uint32_t dofs, uint32_t aofs,
         fni(vece, t1, t0, c);
         tcg_gen_st_vec(t1, cpu_env, dofs + i);
     }
-    tcg_temp_free_vec(t0);
-    tcg_temp_free_vec(t1);
 }
 
 static void expand_2s_vec(unsigned vece, uint32_t dofs, uint32_t aofs,
@@ -1081,8 +1076,6 @@ static void expand_2s_vec(unsigned vece, uint32_t dofs, uint32_t aofs,
         }
         tcg_gen_st_vec(t1, cpu_env, dofs + i);
     }
-    tcg_temp_free_vec(t0);
-    tcg_temp_free_vec(t1);
 }
 
 /* Expand OPSZ bytes worth of three-operand operations using host vectors.  */
@@ -1105,9 +1098,6 @@ static void expand_3_vec(unsigned vece, uint32_t dofs, uint32_t aofs,
         fni(vece, t2, t0, t1);
         tcg_gen_st_vec(t2, cpu_env, dofs + i);
     }
-    tcg_temp_free_vec(t2);
-    tcg_temp_free_vec(t1);
-    tcg_temp_free_vec(t0);
 }
 
 /*
@@ -1134,9 +1124,6 @@ static void expand_3i_vec(unsigned vece, uint32_t dofs, uint32_t aofs,
         fni(vece, t2, t0, t1, c);
         tcg_gen_st_vec(t2, cpu_env, dofs + i);
     }
-    tcg_temp_free_vec(t0);
-    tcg_temp_free_vec(t1);
-    tcg_temp_free_vec(t2);
 }
 
 /* Expand OPSZ bytes worth of four-operand operations using host vectors.  */
@@ -1162,10 +1149,6 @@ static void expand_4_vec(unsigned vece, uint32_t dofs, uint32_t aofs,
             tcg_gen_st_vec(t1, cpu_env, aofs + i);
         }
     }
-    tcg_temp_free_vec(t3);
-    tcg_temp_free_vec(t2);
-    tcg_temp_free_vec(t1);
-    tcg_temp_free_vec(t0);
 }
 
 /*
@@ -1191,10 +1174,6 @@ static void expand_4i_vec(unsigned vece, uint32_t dofs, uint32_t aofs,
         fni(vece, t0, t1, t2, t3, c);
         tcg_gen_st_vec(t0, cpu_env, dofs + i);
     }
-    tcg_temp_free_vec(t3);
-    tcg_temp_free_vec(t2);
-    tcg_temp_free_vec(t1);
-    tcg_temp_free_vec(t0);
 }
 
 /* Expand a vector two-operand operation.  */
@@ -1732,7 +1711,6 @@ void tcg_gen_gvec_dup_mem(unsigned vece, uint32_t dofs, uint32_t aofs,
             TCGv_vec t_vec = tcg_temp_new_vec(type);
             tcg_gen_dup_mem_vec(vece, t_vec, cpu_env, aofs);
             do_dup_store(type, dofs, oprsz, maxsz, t_vec);
-            tcg_temp_free_vec(t_vec);
         } else if (vece <= MO_32) {
             TCGv_i32 in = tcg_temp_ebb_new_i32();
             switch (vece) {
@@ -1766,7 +1744,6 @@ void tcg_gen_gvec_dup_mem(unsigned vece, uint32_t dofs, uint32_t aofs,
             for (i = (aofs == dofs) * 16; i < oprsz; i += 16) {
                 tcg_gen_st_vec(in, cpu_env, dofs + i);
             }
-            tcg_temp_free_vec(in);
         } else {
             TCGv_i64 in0 = tcg_temp_ebb_new_i64();
             TCGv_i64 in1 = tcg_temp_ebb_new_i64();
@@ -1796,7 +1773,6 @@ void tcg_gen_gvec_dup_mem(unsigned vece, uint32_t dofs, uint32_t aofs,
             for (i = (aofs == dofs) * 32; i < oprsz; i += 32) {
                 tcg_gen_st_vec(in, cpu_env, dofs + i);
             }
-            tcg_temp_free_vec(in);
         } else if (TCG_TARGET_HAS_v128) {
             TCGv_vec in0 = tcg_temp_new_vec(TCG_TYPE_V128);
             TCGv_vec in1 = tcg_temp_new_vec(TCG_TYPE_V128);
@@ -1807,8 +1783,6 @@ void tcg_gen_gvec_dup_mem(unsigned vece, uint32_t dofs, uint32_t aofs,
                 tcg_gen_st_vec(in0, cpu_env, dofs + i);
                 tcg_gen_st_vec(in1, cpu_env, dofs + i + 16);
             }
-            tcg_temp_free_vec(in0);
-            tcg_temp_free_vec(in1);
         } else {
             TCGv_i64 in[4];
             int j;
@@ -3137,14 +3111,14 @@ static void expand_2sh_vec(unsigned vece, uint32_t dofs, uint32_t aofs,
                            void (*fni)(unsigned, TCGv_vec, TCGv_vec, TCGv_i32))
 {
     TCGv_vec t0 = tcg_temp_new_vec(type);
+    TCGv_vec t1 = tcg_temp_new_vec(type);
     uint32_t i;
 
     for (i = 0; i < oprsz; i += tysz) {
         tcg_gen_ld_vec(t0, cpu_env, aofs + i);
-        fni(vece, t0, t0, shift);
-        tcg_gen_st_vec(t0, cpu_env, dofs + i);
+        fni(vece, t1, t0, shift);
+        tcg_gen_st_vec(t1, cpu_env, dofs + i);
     }
-    tcg_temp_free_vec(t0);
 }
 
 static void
@@ -3722,16 +3696,15 @@ static void expand_cmp_vec(unsigned vece, uint32_t dofs, uint32_t aofs,
 {
     TCGv_vec t0 = tcg_temp_new_vec(type);
     TCGv_vec t1 = tcg_temp_new_vec(type);
+    TCGv_vec t2 = tcg_temp_new_vec(type);
     uint32_t i;
 
     for (i = 0; i < oprsz; i += tysz) {
         tcg_gen_ld_vec(t0, cpu_env, aofs + i);
         tcg_gen_ld_vec(t1, cpu_env, bofs + i);
-        tcg_gen_cmp_vec(cond, vece, t0, t0, t1);
-        tcg_gen_st_vec(t0, cpu_env, dofs + i);
+        tcg_gen_cmp_vec(cond, vece, t2, t0, t1);
+        tcg_gen_st_vec(t2, cpu_env, dofs + i);
     }
-    tcg_temp_free_vec(t1);
-    tcg_temp_free_vec(t0);
 }
 
 void tcg_gen_gvec_cmp(TCGCond cond, unsigned vece, uint32_t dofs,
-- 
2.34.1



  reply	other threads:[~2023-08-31  2:58 UTC|newest]

Thread overview: 9+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2023-08-31  2:57 [PATCH 0/4] tcg: Optimize loads and stores to env Richard Henderson
2023-08-31  2:57 ` Richard Henderson [this message]
2023-08-31  2:57 ` [PATCH 2/4] tcg/optimize: Pipe OptContext into reset_ts Richard Henderson
2023-10-02 12:37   ` Philippe Mathieu-Daudé
2023-08-31  2:57 ` [PATCH 3/4] tcg: Optimize env memory operations Richard Henderson
2023-08-31  2:57 ` [PATCH 4/4] tcg: Eliminate duplicate env store operations Richard Henderson
2023-09-28 22:45 ` [PATCH 0/4] tcg: Optimize loads and stores to env Richard Henderson
2023-10-13 17:40   ` Richard Henderson
2023-10-16  3:01 ` gaosong

Reply instructions:

You may reply publicly to this message via plain-text email
using any one of the following methods:

* Save the following mbox file, import it into your mail client,
  and reply-to-all from there: mbox

  Avoid top-posting and favor interleaved quoting:
  https://en.wikipedia.org/wiki/Posting_style#Interleaved_style

* Reply using the --to, --cc, and --in-reply-to
  switches of git-send-email(1):

  git send-email \
    --in-reply-to=20230831025729.1194388-2-richard.henderson@linaro.org \
    --to=richard.henderson@linaro.org \
    --cc=qemu-devel@nongnu.org \
    /path/to/YOUR_REPLY

  https://kernel.org/pub/software/scm/git/docs/git-send-email.html

* If your mail client supports setting the In-Reply-To header
  via mailto: links, try the mailto: link
Be sure your reply has a Subject: header at the top and a blank line before the message body.
This is a public inbox, see mirroring instructions
for how to clone and mirror all data and code used for this inbox;
as well as URLs for NNTP newsgroup(s).