public inbox for gentoo-commits@lists.gentoo.org
 help / color / mirror / Atom feed
From: "Sam James" <sam@gentoo.org>
To: gentoo-commits@lists.gentoo.org
Subject: [gentoo-commits] proj/gcc-patches:master commit in: 15.0.0/gentoo/
Date: Wed, 14 Aug 2024 02:57:32 +0000 (UTC)	[thread overview]
Message-ID: <1723604222.70f5adef33e0620d934fc7fb0822e592e3ff04a1.sam@gentoo> (raw)

commit:     70f5adef33e0620d934fc7fb0822e592e3ff04a1
Author:     Sam James <sam <AT> gentoo <DOT> org>
AuthorDate: Wed Aug 14 02:57:02 2024 +0000
Commit:     Sam James <sam <AT> gentoo <DOT> org>
CommitDate: Wed Aug 14 02:57:02 2024 +0000
URL:        https://gitweb.gentoo.org/proj/gcc-patches.git/commit/?id=70f5adef

15.0.0: add 32_all_genoutput-speedup.patch

Link: https://inbox.sourceware.org/gcc-patches/20240814021909.37082-1-cooper.qu <AT> linux.alibaba.com/
Signed-off-by: Sam James <sam <AT> gentoo.org>

 15.0.0/gentoo/32_all_genoutput-speedup.patch | 247 +++++++++++++++++++++++++++
 15.0.0/gentoo/README.history                 |   4 +
 2 files changed, 251 insertions(+)

diff --git a/15.0.0/gentoo/32_all_genoutput-speedup.patch b/15.0.0/gentoo/32_all_genoutput-speedup.patch
new file mode 100644
index 0000000..a379bf8
--- /dev/null
+++ b/15.0.0/gentoo/32_all_genoutput-speedup.patch
@@ -0,0 +1,247 @@
+https://inbox.sourceware.org/gcc-patches/20240814021909.37082-1-cooper.qu@linux.alibaba.com/
+
+From 23ea354ab6c1faf858120b65a0114c5d0bbeaf6e Mon Sep 17 00:00:00 2001
+Message-ID: <23ea354ab6c1faf858120b65a0114c5d0bbeaf6e.1723604026.git.sam@gentoo.org>
+From: Xianmiao Qu <cooper.qu@linux.alibaba.com>
+Date: Wed, 14 Aug 2024 10:19:09 +0800
+Subject: [PATCH] genoutput: Accelerate the place_operands function.
+
+With the increase in the number of modes and patterns for some
+backend architectures, the place_operands function becomes a
+bottleneck int the speed of genoutput, and may even become a
+bottleneck int the overall speed of building the GCC project.
+This patch aims to accelerate the place_operands function,
+the optimizations it includes are:
+1. Use a hash table to store operand information,
+   improving the lookup time for the first operand.
+2. Move mode comparison to the beginning to avoid the scenarios of most strcmp.
+
+I tested the speed improvements for the following backends,
+	Improvement Ratio
+x86_64	197.9%
+aarch64	954.5%
+riscv	2578.6%
+If the build machine is slow, then this improvement can save a lot of time.
+
+I tested the genoutput output for x86_64/aarch64/riscv backends,
+and there was no difference compared to before the optimization,
+so this shouldn't introduce any functional issues.
+
+gcc/
+	* genoutput.cc (struct operand_data): Add member 'eq_next' to
+	point to the next member with the same hash value in the
+	hash table.
+	(compare_operands): Move the comparison of the mode to the very
+	beginning to accelerate the comparison of the two operands.
+	(struct operand_data_hasher): New, a class that takes into account
+	the necessary elements for comparing the equality of two operands
+	in its hash value.
+	(operand_data_hasher::hash): New.
+	(operand_data_hasher::equal): New.
+	(operand_datas): New, hash table of konwn pattern operands.
+	(place_operands): Use a hash table instead of traversing the array
+	to find the same operand.
+	(main): Add initialization of the hash table 'operand_datas'.
+---
+ gcc/genoutput.cc | 111 +++++++++++++++++++++++++++++++++++++----------
+ 1 file changed, 88 insertions(+), 23 deletions(-)
+
+diff --git a/gcc/genoutput.cc b/gcc/genoutput.cc
+index efd81766bb5b..16fd811b5dd5 100644
+--- a/gcc/genoutput.cc
++++ b/gcc/genoutput.cc
+@@ -91,6 +91,7 @@ along with GCC; see the file COPYING3.  If not see
+ #include "errors.h"
+ #include "read-md.h"
+ #include "gensupport.h"
++#include "hash-table.h"
+ 
+ /* No instruction can have more operands than this.  Sorry for this
+    arbitrary limit, but what machine will have an instruction with
+@@ -112,6 +113,8 @@ static int next_operand_number = 1;
+ struct operand_data
+ {
+   struct operand_data *next;
++  /* Point to the next member with the same hash value in the hash table.  */
++  struct operand_data *eq_next;
+   int index;
+   const char *predicate;
+   const char *constraint;
+@@ -127,7 +130,7 @@ struct operand_data
+ 
+ static struct operand_data null_operand =
+ {
+-  0, 0, "", "", E_VOIDmode, 0, 0, 0, 0, 0
++  0, 0, 0, "", "", E_VOIDmode, 0, 0, 0, 0, 0
+ };
+ 
+ static struct operand_data *odata = &null_operand;
+@@ -174,8 +177,8 @@ static void output_operand_data (void);
+ static void output_insn_data (void);
+ static void output_get_insn_name (void);
+ static void scan_operands (class data *, rtx, int, int);
+-static int compare_operands (struct operand_data *,
+-			     struct operand_data *);
++static int compare_operands (const struct operand_data *,
++			     const struct operand_data *);
+ static void place_operands (class data *);
+ static void process_template (class data *, const char *);
+ static void validate_insn_alternatives (class data *);
+@@ -528,10 +531,18 @@ scan_operands (class data *d, rtx part, int this_address_p,
+ /* Compare two operands for content equality.  */
+ 
+ static int
+-compare_operands (struct operand_data *d0, struct operand_data *d1)
++compare_operands (const struct operand_data *d0,
++		  const struct operand_data *d1)
+ {
+   const char *p0, *p1;
+ 
++  /* On one hand, comparing strings for predicate and constraint
++     is time-consuming, and on the other hand, the probability of
++     different modes is relatively high. Therefore, checking the mode
++     first can speed up the execution of the program.  */
++  if (d0->mode != d1->mode)
++    return 0;
++
+   p0 = d0->predicate;
+   if (!p0)
+     p0 = "";
+@@ -550,9 +561,6 @@ compare_operands (struct operand_data *d0, struct operand_data *d1)
+   if (strcmp (p0, p1) != 0)
+     return 0;
+ 
+-  if (d0->mode != d1->mode)
+-    return 0;
+-
+   if (d0->strict_low != d1->strict_low)
+     return 0;
+ 
+@@ -562,6 +570,46 @@ compare_operands (struct operand_data *d0, struct operand_data *d1)
+   return 1;
+ }
+ 
++/* This is a class that takes into account the necessary elements for
++   comparing the equality of two operands in its hash value.  */
++struct operand_data_hasher : nofree_ptr_hash <operand_data>
++{
++  static inline hashval_t hash (const operand_data *);
++  static inline bool equal (const operand_data *, const operand_data *);
++};
++
++hashval_t
++operand_data_hasher::hash (const operand_data * op_info)
++{
++  inchash::hash h;
++  const char *pred, *cons;
++
++  pred = op_info->predicate;
++  if (!pred)
++    pred = "";
++  h.add (pred, strlen (pred) + 1);
++
++  cons = op_info->constraint;
++  if (!cons)
++    cons = "";
++  h.add (cons, strlen (cons) + 1);
++
++  h.add_object (op_info->mode);
++  h.add_object (op_info->strict_low);
++  h.add_object (op_info->eliminable);
++  return h.end ();
++}
++
++bool
++operand_data_hasher::equal (const operand_data * op_info1,
++			    const operand_data * op_info2)
++{
++  return compare_operands (op_info1, op_info2);
++}
++
++/* Hashtable of konwn pattern operands.  */
++static hash_table<operand_data_hasher> *operand_datas;
++
+ /* Scan the list of operands we've already committed to output and either
+    find a subsequence that is the same, or allocate a new one at the end.  */
+ 
+@@ -569,6 +617,7 @@ static void
+ place_operands (class data *d)
+ {
+   struct operand_data *od, *od2;
++  struct operand_data **slot;
+   int i;
+ 
+   if (d->n_operands == 0)
+@@ -577,23 +626,24 @@ place_operands (class data *d)
+       return;
+     }
+ 
++  od = operand_datas->find (&d->operand[0]);
+   /* Brute force substring search.  */
+-  for (od = odata, i = 0; od; od = od->next, i = 0)
+-    if (compare_operands (od, &d->operand[0]))
+-      {
+-	od2 = od->next;
+-	i = 1;
+-	while (1)
+-	  {
+-	    if (i == d->n_operands)
+-	      goto full_match;
+-	    if (od2 == NULL)
+-	      goto partial_match;
+-	    if (! compare_operands (od2, &d->operand[i]))
+-	      break;
+-	    ++i, od2 = od2->next;
+-	  }
+-      }
++  for (; od; od = od->eq_next)
++    {
++      od2 = od->next;
++      i = 1;
++      while (1)
++	{
++	  if (i == d->n_operands)
++	    goto full_match;
++	  if (od2 == NULL)
++	    goto partial_match;
++	  if (! compare_operands (od2, &d->operand[i]))
++	    break;
++	  ++i, od2 = od2->next;
++	}
++    }
++  i = 0;
+ 
+   /* Either partial match at the end of the list, or no match.  In either
+      case, we tack on what operands are remaining to the end of the list.  */
+@@ -605,6 +655,20 @@ place_operands (class data *d)
+       *odata_end = od2;
+       odata_end = &od2->next;
+       od2->index = next_operand_number++;
++      /* Insert the operand_data variable OD2 into the hash table.
++	 If a variable with the same hash value already exists in
++	 the hash table, insert the element at the end of the
++	 linked list connected through the eq_next member.  */
++      slot = operand_datas->find_slot (od2, INSERT);
++      if (*slot)
++	{
++	  struct operand_data *last = (struct operand_data *) *slot;
++	  while (last->eq_next)
++	    last = last->eq_next;
++	  last->eq_next = od2;
++	}
++      else
++	*slot = od2;
+     }
+   *odata_end = NULL;
+   return;
+@@ -1049,6 +1113,7 @@ main (int argc, const char **argv)
+   progname = "genoutput";
+ 
+   init_insn_for_nothing ();
++  operand_datas = new hash_table<operand_data_hasher> (1024);
+ 
+   if (!init_rtx_reader_args (argc, argv))
+     return (FATAL_EXIT_CODE);
+-- 
+2.45.2
+

diff --git a/15.0.0/gentoo/README.history b/15.0.0/gentoo/README.history
index 468a873..1849089 100644
--- a/15.0.0/gentoo/README.history
+++ b/15.0.0/gentoo/README.history
@@ -1,3 +1,7 @@
+10	????
+
+	+ 32_all_genoutput-speedup.patch
+
 9	11 August 2024
 
 	U 04_all_nossp-on-nostdlib.patch


             reply	other threads:[~2024-08-14  2:57 UTC|newest]

Thread overview: 224+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2024-08-14  2:57 Sam James [this message]
  -- strict thread matches above, loose matches on Subject: below --
2025-04-18 19:21 [gentoo-commits] proj/gcc-patches:master commit in: 15.0.0/gentoo/ Sam James
2025-04-16 13:30 Sam James
2025-04-16 10:04 Sam James
2025-04-16  9:58 Sam James
2025-04-15 21:09 Sam James
2025-04-15 16:40 Sam James
2025-04-14 22:59 Sam James
2025-04-14 22:47 Sam James
2025-04-14 22:44 Sam James
2025-04-14 21:06 Sam James
2025-04-14 20:51 Sam James
2025-04-14 16:08 Sam James
2025-04-14 16:08 Sam James
2025-04-13 22:54 Sam James
2025-04-10 21:59 Sam James
2025-04-10  9:08 Sam James
2025-04-09 15:27 Sam James
2025-04-09 13:27 Sam James
2025-04-09 12:42 Sam James
2025-04-07 18:02 Sam James
2025-04-07  7:07 Sam James
2025-04-07  6:54 Sam James
2025-04-06 23:08 Sam James
2025-04-05 15:33 Sam James
2025-04-05  8:02 Sam James
2025-04-05  1:43 Sam James
2025-04-04 19:06 Sam James
2025-04-02 18:48 Sam James
2025-04-02 18:03 Sam James
2025-04-02 16:14 Sam James
2025-04-02 13:56 Sam James
2025-04-02  4:59 Sam James
2025-04-01 14:46 Sam James
2025-04-01 14:46 Sam James
2025-03-31 22:16 Sam James
2025-03-31 22:03 Sam James
2025-03-31  4:05 Sam James
2025-03-29 20:31 Sam James
2025-03-29 14:33 Sam James
2025-03-29 13:51 Sam James
2025-03-26  6:25 Sam James
2025-03-25 10:27 Sam James
2025-03-25  8:38 Sam James
2025-03-25  2:32 Sam James
2025-03-25  1:27 Sam James
2025-03-24  0:35 Sam James
2025-03-21 19:31 Sam James
2025-03-21 17:21 Sam James
2025-03-21 16:23 Sam James
2025-03-21 11:20 Sam James
2025-03-21  8:51 Sam James
2025-03-21  6:07 Sam James
2025-03-20 22:08 Sam James
2025-03-20  1:59 Sam James
2025-03-20  1:59 Sam James
2025-03-16 22:37 Sam James
2025-03-14 14:46 Sam James
2025-03-14 13:37 Sam James
2025-03-13 16:48 Sam James
2025-03-13 10:08 Sam James
2025-03-11 10:32 Sam James
2025-03-07 16:54 Sam James
2025-03-03 16:38 Sam James
2025-03-01 10:33 Sam James
2025-03-01  6:50 Sam James
2025-02-17  1:30 Sam James
2025-02-13  9:23 Sam James
2025-02-12 15:12 Sam James
2025-02-10 21:22 Sam James
2025-02-09 23:58 Sam James
2025-02-07 23:37 Sam James
2025-02-07 21:19 Sam James
2025-02-03 22:04 Sam James
2025-02-02 22:41 Sam James
2025-01-29 20:21 Sam James
2025-01-26 22:52 Sam James
2025-01-22 16:27 Sam James
2025-01-19 22:43 Sam James
2025-01-16 23:11 Sam James
2025-01-16 23:11 Sam James
2025-01-15 11:41 Sam James
2025-01-14 16:22 Sam James
2025-01-14 15:06 Sam James
2025-01-14 15:06 Sam James
2025-01-14 12:29 Sam James
2025-01-14  8:43 Sam James
2025-01-14  8:40 Sam James
2025-01-13 13:58 Sam James
2025-01-13  6:00 Sam James
2025-01-13  3:40 Sam James
2025-01-13  3:23 Sam James
2025-01-13  3:20 Sam James
2025-01-13  0:20 Sam James
2025-01-12 18:53 Sam James
2025-01-11 12:53 Sam James
2025-01-08 21:51 Sam James
2025-01-06 10:50 Sam James
2025-01-06 10:03 Sam James
2025-01-06  4:49 Sam James
2025-01-06  4:44 Sam James
2025-01-06  4:13 Sam James
2025-01-06  4:13 Sam James
2025-01-06  4:13 Sam James
2025-01-06  4:03 Sam James
2025-01-05 23:19 Sam James
2025-01-03  3:07 Sam James
2024-12-30  1:05 Sam James
2024-12-29 10:00 Sam James
2024-12-27 15:14 Sam James
2024-12-24 20:48 Sam James
2024-12-22 22:46 Sam James
2024-12-20 11:25 Sam James
2024-12-20  5:57 Sam James
2024-12-20  1:55 Sam James
2024-12-19 18:34 Sam James
2024-12-13 13:23 Sam James
2024-12-13 11:52 Sam James
2024-12-13  5:08 Sam James
2024-12-12 12:28 Sam James
2024-12-11  4:41 Sam James
2024-12-11  0:58 Sam James
2024-12-10 19:19 Sam James
2024-12-10 14:55 Sam James
2024-12-10  5:19 Sam James
2024-12-10  5:13 Sam James
2024-12-10  5:11 Sam James
2024-12-10  5:07 Sam James
2024-12-09  3:05 Sam James
2024-12-08 22:41 Sam James
2024-12-06 17:33 Sam James
2024-12-04 20:40 Sam James
2024-12-01 22:51 Sam James
2024-12-01 22:51 Sam James
2024-11-30 11:30 Sam James
2024-11-27 17:42 Sam James
2024-11-25 15:10 Sam James
2024-11-25  3:01 Sam James
2024-11-25  3:00 Sam James
2024-11-25  3:00 Sam James
2024-11-24 22:42 Sam James
2024-11-18 17:25 Sam James
2024-11-18 10:42 Sam James
2024-11-18 10:42 Sam James
2024-11-18  9:25 Sam James
2024-11-18  9:25 Sam James
2024-11-14 18:38 Sam James
2024-11-13  4:26 Sam James
2024-11-13  0:16 Sam James
2024-11-12  2:33 Sam James
2024-11-11 19:46 Sam James
2024-11-11 19:46 Sam James
2024-11-10 22:41 Sam James
2024-11-09 16:24 Sam James
2024-11-09  7:55 Sam James
2024-11-08  8:22 Sam James
2024-11-07 16:13 Sam James
2024-11-03 23:16 Sam James
2024-11-01  8:24 Sam James
2024-11-01  8:24 Sam James
2024-11-01  8:18 Sam James
2024-11-01  8:17 Sam James
2024-10-30 16:03 Sam James
2024-10-29 19:17 Sam James
2024-10-28 21:32 Sam James
2024-10-28  8:09 Sam James
2024-10-23 15:40 Sam James
2024-10-22 19:09 Sam James
2024-10-22 18:34 Sam James
2024-10-21 12:33 Sam James
2024-10-21 12:27 Sam James
2024-10-21 12:26 Sam James
2024-10-21 11:45 Sam James
2024-10-20 22:42 Sam James
2024-10-18 14:05 Sam James
2024-10-18 10:35 Sam James
2024-10-17 23:33 Sam James
2024-10-17 23:03 Sam James
2024-10-17  5:01 Sam James
2024-10-17  4:15 Sam James
2024-10-13 22:48 Sam James
2024-10-07  2:45 Sam James
2024-10-04 10:37 Sam James
2024-10-04  9:28 Sam James
2024-10-02 19:45 Sam James
2024-09-30 14:05 Sam James
2024-09-29 22:56 Sam James
2024-09-24  1:41 Sam James
2024-09-23 15:23 Sam James
2024-09-02  2:28 Sam James
2024-08-26 13:44 Sam James
2024-08-26  6:24 Sam James
2024-08-23 13:51 Sam James
2024-08-20 20:31 Sam James
2024-08-19 18:43 Sam James
2024-08-14  9:48 Sam James
2024-08-11 22:40 Sam James
2024-08-09 19:54 Sam James
2024-08-09 19:54 Sam James
2024-08-09 19:47 Sam James
2024-08-09 19:25 Sam James
2024-08-08 11:10 Sam James
2024-08-08 11:06 Sam James
2024-08-08 11:03 Sam James
2024-08-05  9:09 Sam James
2024-08-05  1:54 Sam James
2024-08-05  1:51 Sam James
2024-08-02 20:39 Sam James
2024-08-01 14:40 Sam James
2024-07-28 23:34 Sam James
2024-07-22  1:11 Sam James
2024-07-19 11:14 Sam James
2024-07-18  0:45 Sam James
2024-07-14 23:36 Sam James
2024-06-28 12:49 Sam James
2024-06-27  0:02 Sam James
2024-06-26 23:57 Sam James
2024-06-16 22:45 Sam James
2024-06-10 20:18 Sam James
2024-06-10 17:28 Sam James
2024-06-10 17:28 Sam James
2024-06-10  2:08 Sam James
2024-06-08 17:03 Sam James
2024-06-08 17:03 Sam James

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=1723604222.70f5adef33e0620d934fc7fb0822e592e3ff04a1.sam@gentoo \
    --to=sam@gentoo.org \
    --cc=gentoo-commits@lists.gentoo.org \
    --cc=gentoo-dev@lists.gentoo.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