public inbox for gcc-patches@gcc.gnu.org
 help / color / mirror / Atom feed
From: Mikael Morin <mikael.morin@sfr.fr>
To: gfortran <fortran@gcc.gnu.org> , GCC patches <gcc-patches@gcc.gnu.org>
Subject: [Patch, fortran] [31/66] inline sum and product: Update the scalarizer: Split gfc_conv_loop_setup.
Date: Thu, 27 Oct 2011 23:34:00 -0000	[thread overview]
Message-ID: <20111027233151.18581.91359@gimli.local> (raw)
In-Reply-To: <20111027233144.18581.30688@gimli.local>

[-- Attachment #1: Type: text/plain, Size: 557 bytes --]

Loop-centered functions will be called recursively to handle nested loops.
This is the case of gfc_conv_loop_setup. However, gfc_conv_loop_setup has
a tempoary handling part in the middle which doesn't need to be called
recursively. This patch moves the beginning and ending of gfc_conv_loop_setup
into functions of their own, so that they can be called recursively (later).

To share gfc_conv_loop_setup's loopspec local variable in the three functions,
we store it into gfc_loopinfo's specloop field (which existed, but was
completely unused before).
OK?

[-- Attachment #2: pr43829-31.CL --]
[-- Type: text/plain, Size: 423 bytes --]

2011-10-19  Mikael Morin  <mikael@gcc.gnu.org>

	* trans-array.c (set_loop_bounds): Separate the beginning of
	gfc_conv_loop_setup into a function of its own.
	(set_delta): Separate the end of gfc_conv_loop_setup into a function
	of its own.
	(gfc_conv_loop_setup): Call set_loop_bounds and set delta.
	(set_loop_bounds, set_delta, gfc_conv_loop_setup): Make loopspec a
	pointer to the specloop field from the loop struct.

[-- Attachment #3: pr43829-31.patch --]
[-- Type: text/x-diff, Size: 2875 bytes --]

diff --git a/trans-array.c b/trans-array.c
index 045c426..302f937 100644
--- a/trans-array.c
+++ b/trans-array.c
@@ -3919,25 +3919,25 @@ temporary:
 }
 
 
-/* Initialize the scalarization loop.  Creates the loop variables.  Determines
-   the range of the loop variables.  Creates a temporary if required.
-   Calculates how to transform from loop variables to array indices for each
-   expression.  Also generates code for scalar expressions which have been
-   moved outside the loop.  */
+/* Browse through each array's information from the scalarizer and set the loop
+   bounds according to the "best" one (per dimension), i.e. the one which
+   provides the most information (constant bounds, shape, etc).  */
 
-void
-gfc_conv_loop_setup (gfc_loopinfo * loop, locus * where)
+static void
+set_loop_bounds (gfc_loopinfo *loop)
 {
   int n, dim, spec_dim;
   gfc_array_info *info;
   gfc_array_info *specinfo;
-  gfc_ss *ss, *tmp_ss;
+  gfc_ss *ss;
   tree tmp;
-  gfc_ss *loopspec[GFC_MAX_DIMENSIONS];
+  gfc_ss **loopspec;
   bool dynamic[GFC_MAX_DIMENSIONS];
   mpz_t *cshape;
   mpz_t i;
 
+  loopspec = loop->specloop;
+
   mpz_init (i);
   for (n = 0; n < loop->dimen; n++)
     {
@@ -4119,6 +4119,26 @@ gfc_conv_loop_setup (gfc_loopinfo * loop, locus * where)
 	  loop->from[n] = gfc_index_zero_node;
 	}
     }
+  mpz_clear (i);
+}
+
+
+static void set_delta (gfc_loopinfo *loop);
+
+
+/* Initialize the scalarization loop.  Creates the loop variables.  Determines
+   the range of the loop variables.  Creates a temporary if required.
+   Also generates code for scalar expressions which have been
+   moved outside the loop.  */
+
+void
+gfc_conv_loop_setup (gfc_loopinfo * loop, locus * where)
+{
+  gfc_ss *tmp_ss;
+  tree tmp;
+  int n;
+
+  set_loop_bounds (loop);
 
   /* Add all the scalar code that can be taken out of the loops.
      This may include calculating the loop bounds, so do it before
@@ -4153,15 +4173,31 @@ gfc_conv_loop_setup (gfc_loopinfo * loop, locus * where)
     }
 
   for (n = 0; n < loop->temp_dim; n++)
-    loopspec[loop->order[n]] = NULL;
-
-  mpz_clear (i);
+    loop->specloop[loop->order[n]] = NULL;
 
   /* For array parameters we don't have loop variables, so don't calculate the
      translations.  */
   if (loop->array_parameter)
     return;
 
+  set_delta (loop);
+}
+
+
+/* Calculates how to transform from loop variables to array indices for each
+   array: once loop bounds are chosen, sets the difference (DELTA field) between
+   loop bounds and array reference bounds, for each array info.  */
+
+static void
+set_delta (gfc_loopinfo *loop)
+{
+  gfc_ss *ss, **loopspec;
+  gfc_array_info *info;
+  tree tmp;
+  int n, dim;
+
+  loopspec = loop->specloop;
+
   /* Calculate the translation from loop variables to array indices.  */
   for (ss = loop->ss; ss != gfc_ss_terminator; ss = ss->loop_chain)
     {

  reply	other threads:[~2011-10-27 23:32 UTC|newest]

Thread overview: 69+ messages / expand[flat|nested]  mbox.gz  Atom feed  top
2011-10-27 23:43 [Patch, fortran] [00/66] PR fortran/43829 Inline sum and product (AKA scalarization of reductions) Mikael Morin
2011-10-27 23:32 ` [Patch, fortran] [20..30/66] inline sum and product: Update core structs Mikael Morin
2011-10-27 23:32   ` [Patch, fortran] [25/66] inline sum and product: Update core structs: Move string_length Mikael Morin
2011-10-27 23:32   ` [Patch, fortran] [29/66] inline sum and product: Update core structs: Move useflags flag Mikael Morin
2011-10-27 23:32   ` [Patch, fortran] [20/66] inline sum and product: Update core structs: Rename gfc_ss_info Mikael Morin
2011-10-27 23:33   ` [Patch, fortran] [23/66] inline sum and product: Update core structs: Move type Mikael Morin
2011-10-27 23:33   ` [Patch, fortran] [22/66] inline sum and product: Update core structs: Move shape Mikael Morin
2011-10-27 23:33   ` [Patch, fortran] [26/66] inline sum and product: Update core structs: Move scalar struct Mikael Morin
2011-10-27 23:34   ` [Patch, fortran] [27/66] inline sum and product: Update core structs: Move temp struct Mikael Morin
2011-10-27 23:35   ` [Patch, fortran] [21/66] inline sum and product: Update core structs: Move dim and dimen Mikael Morin
2011-10-27 23:35   ` [Patch, fortran] [30/66] inline sum and product: Update core structs: Move where flag Mikael Morin
2011-10-27 23:35   ` [Patch, fortran] [24/66] inline sum and product: Update core structs: Move expr Mikael Morin
2011-10-27 23:43   ` [Patch, fortran] [28/66] inline sum and product: Update core structs: Move info struct Mikael Morin
2011-10-27 23:35 ` [Patch, fortran] [07..12/66] inline sum and product: Preliminary cleanups Mikael Morin
2011-10-27 23:31   ` [Patch, fortran] [10/66] inline sum and product: Preliminary cleanups: Use array's instead of loop's dimensions Mikael Morin
2011-10-27 23:31   ` [Patch, fortran] [07/66] inline sum and product: Preliminary cleanups: Useless coarray code removal Mikael Morin
2011-10-27 23:32   ` [Patch, fortran] [08/66] inline sum and product: Preliminary cleanups: Remove redundant condition Mikael Morin
2011-10-27 23:32   ` [Patch, fortran] [12/66] inline sum and product: Preliminary cleanups: Stop loop before end marker Mikael Morin
2011-10-27 23:36   ` [Patch, fortran] [11/66] inline sum and product: Preliminary cleanups: Skip temporary case Mikael Morin
2011-10-27 23:36   ` [Patch, fortran] [09/66] inline sum and product: Preliminary cleanups: Assertify condition Mikael Morin
2011-10-27 23:35 ` [Patch, fortran] [31..53/66] inline sum and product: Update the scalarizer Mikael Morin
2011-10-27 23:34   ` Mikael Morin [this message]
2011-10-27 23:34   ` [Patch, fortran] [32/66] inline sum and product: Update the scalarizer: clear specloop in gfc_trans_create_temp_arrays Mikael Morin
2011-10-27 23:35   ` [Patch, fortran] [45/66] inline sum and product: Update the scalarizer: Update dimension mapping inversion functions Mikael Morin
2011-10-27 23:35   ` [Patch, fortran] [33/66] inline sum and product: Update the scalarizer Mikael Morin
2011-10-27 23:35   ` [Patch, fortran] [44/66] inline sum and product: Update the scalarizer: New gfc_ss::nested_ss field Mikael Morin
2011-10-27 23:35   ` [Patch, fortran] [47..48/66] inline sum and product: Update the scalarizer: New gfc_loopinfo::nested_loop field Mikael Morin
2011-10-27 23:36   ` [Patch, fortran] [35..39/66] inline sum and product: Update the scalarizer: New gfc_ss::loop field Mikael Morin
2011-10-27 23:36   ` [Patch, fortran] [34/66] inline sum and product: Update the scalarizer: gfc_ss_info refcounting Mikael Morin
2011-10-27 23:36   ` [Patch, fortran] [40..43/66] inline sum and product: Update the scalarizer: New gfc_ss::parent field Mikael Morin
2011-10-27 23:36   ` [Patch, fortran] [46/66] inline sum and product: Update the scalarizer: Update gfc_trans_create_temp_array Mikael Morin
2011-10-27 23:36   ` [Patch, fortran] [52/66] inline sum and product: Update the scalarizer: New outermost_loop function Mikael Morin
2011-10-27 23:36   ` [Patch, fortran] [53/66] inline sum and product: Update the scalarizer: Update gfc_trans_preloop_setup Mikael Morin
2011-10-27 23:36   ` [Patch, fortran] [49..51/66] inline sum and product: Update the scalarizer: New parent loop Mikael Morin
2011-10-27 23:36 ` [Patch, fortran] [01..06/66] inline sum and product: Prepare gfc_trans_preloop_setup Mikael Morin
2011-10-27 23:30   ` [Patch, fortran] [01/66] " Mikael Morin
2011-10-27 23:30   ` [Patch, fortran] [06/66] " Mikael Morin
2011-10-30  9:52     ` Paul Richard Thomas
2011-10-30 21:57       ` Mikael Morin
2011-10-27 23:31   ` [Patch, fortran] [03/66] " Mikael Morin
2011-10-27 23:36   ` [Patch, fortran] [05/66] " Mikael Morin
2011-10-27 23:36   ` [Patch, fortran] [02/66] " Mikael Morin
2011-10-27 23:38   ` [Patch, fortran] [04/66] " Mikael Morin
2011-10-27 23:36 ` [Patch, fortran] [62..66/66] inline sum and product: Inline sum Mikael Morin
2011-10-27 23:36   ` [Patch, fortran] [65/66] inline sum and product: Inline sum: Change se initialization Mikael Morin
2011-10-27 23:36   ` [Patch, fortran] [63/66] inline sum and product: Inline sum: Change argument handling Mikael Morin
2011-10-27 23:36   ` [Patch, fortran] [64/66] inline sum and product: Inline sum: Change loop use Mikael Morin
2011-10-27 23:36   ` [Patch, fortran] [66/66] inline sum and product: Inline sum: The end Mikael Morin
2011-10-28  0:29   ` [Patch, fortran] [62/66] inline sum and product: Inline sum: Change conditions Mikael Morin
2011-10-28  0:02 ` [Patch, fortran] [13..19/66] inline sum and product: Interfaces changes Mikael Morin
2011-10-27 23:31   ` [Patch, fortran] [14/66] inline sum and product: Interfaces changes: gfc_trans_array_bound_check, gfc_conv_array_index_offset Mikael Morin
2011-10-27 23:32   ` [Patch, fortran] [17/66] inline sum and product: Interfaces changes: gfc_set_vector_loop_bounds Mikael Morin
2011-10-27 23:32   ` [Patch, fortran] [15/66] inline sum and product: Interfaces changes: obtain name more simply Mikael Morin
2011-10-27 23:36   ` [Patch, fortran] [13/66] inline sum and product: Interfaces changes: gfc_trans_array_constructor Mikael Morin
2011-10-27 23:36   ` [Patch, fortran] [16/66] inline sum and product: Interfaces changes: gfc_trans_create_temp_array Mikael Morin
2011-10-27 23:43   ` [Patch, fortran] [18/66] inline sum and product: Interfaces changes: get_array_ref_dim Mikael Morin
2011-10-27 23:44   ` [Patch, fortran] [19/66] inline sum and product: Interfaces changes: dim_ok Mikael Morin
2011-10-28  0:22 ` [Patch, fortran] [54..61/66] inline sum and product: Prevent regressions Mikael Morin
2011-10-27 23:36   ` [Patch, fortran] [60/66] inline sum and product: Update the scalarizer: Fix error markers Mikael Morin
2011-10-27 23:36   ` [Patch, fortran] [57..59/66] inline sum and product: Prevent regressions: Fix {min, max}{loc, val} Mikael Morin
2011-10-27 23:36   ` [Patch, fortran] [55..56/66] inline sum and product: Prevent regressions: Fix gfc_conv_elemental_dependencies Mikael Morin
2011-10-27 23:43   ` [Patch, fortran] [54/66] inline sum and product: Prevent regressions: Add dependency checking Mikael Morin
2011-10-28  0:01   ` [Patch, fortran] [61/66] inline sum and product: Prevent regressions: Disable frontend optimizations Mikael Morin
2011-10-28 14:35 ` [Patch, fortran] [00/66] PR fortran/43829 Inline sum and product (AKA scalarization of reductions) Jack Howarth
2011-10-28 17:25   ` Mikael Morin
2011-10-29 16:04     ` [Patch, fortran] [00/66] PR fortran/43829 Inline sum and?product " Jack Howarth
2011-11-01 21:33 ` [Patch, fortran] [00/66] PR fortran/43829 Inline sum and product " Paul Richard Thomas
2011-11-04  3:51   ` Mikael Morin
2011-11-04  9:39     ` Richard Guenther

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=20111027233151.18581.91359@gimli.local \
    --to=mikael.morin@sfr.fr \
    --cc=fortran@gcc.gnu.org \
    --cc=gcc-patches@gcc.gnu.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 read-only IMAP folder(s) and NNTP newsgroup(s).