* [OpenACC] internal fn folding
@ 2015-10-28 18:41 Nathan Sidwell
2015-11-02 13:56 ` Nathan Sidwell
0 siblings, 1 reply; 6+ messages in thread
From: Nathan Sidwell @ 2015-10-28 18:41 UTC (permalink / raw)
To: GCC Patches, Richard Guenther
[-- Attachment #1: Type: text/plain, Size: 324 bytes --]
Richard,
this patch adds folding for the new GOACC_DIM_POS and GOACC_DIM_SIZE internal
functions. IIUC gimple_fold_call is the right place to add this.
The size of a compute dimension is very often a compile-time constant. On the
host, in particular it's 1, which means we can deduce the POS must be zero.
ok?
nathan
[-- Attachment #2: trunk-dim-fold.patch --]
[-- Type: text/x-patch, Size: 2040 bytes --]
2015-10-28 Nathan Sidwell <nathan@codesourcery.com>
* gimple-fold.c: Include omp-low.h.
(fold_internal_goacc_dim): New.
(gimple_fold_call): Call it.
Index: gcc/gimple-fold.c
===================================================================
--- gcc/gimple-fold.c (revision 229488)
+++ gcc/gimple-fold.c (working copy)
@@ -64,6 +64,7 @@ along with GCC; see the file COPYING3.
#include "gimple-match.h"
#include "gomp-constants.h"
#include "optabs-query.h"
+#include "omp-low.h"
/* Return true when DECL can be referenced from current unit.
@@ -2925,6 +2926,40 @@ gimple_fold_builtin (gimple_stmt_iterato
return false;
}
+/* Transform IFN_GOACC_DIM_SIZE and IFN_GOACC_DIM_POS internal
+ function calls to constants, where possible. */
+
+static tree
+fold_internal_goacc_dim (gimple *call)
+{
+ tree size = integer_one_node;
+
+ if (tree attrs = get_oacc_fn_attrib (current_function_decl))
+ {
+ tree arg = gimple_call_arg (call, 0);
+ tree pos = TREE_VALUE (attrs);
+
+ for (unsigned axis = (unsigned) TREE_INT_CST_LOW (arg); axis--;)
+ pos = TREE_CHAIN (pos);
+ size = TREE_VALUE (pos);
+ }
+
+ tree result;
+ if (integer_zerop (size))
+ /* Dimension size is dynamic. */
+ result = NULL_TREE;
+ else if (gimple_call_internal_fn (call) == IFN_GOACC_DIM_SIZE)
+ result = size;
+ else if (integer_onep (size))
+ /* Size is one, so pos must be zero. */
+ result = integer_zero_node;
+ else
+ /* Size is more than 1, so POS might be non-zero. */
+ result = NULL_TREE;
+
+ return result;
+}
+
/* Return true if ARG0 CODE ARG1 in infinite signed precision operation
doesn't fit into TYPE. The test for overflow should be regardless of
-fwrapv, and even for unsigned types. */
@@ -3125,6 +3160,10 @@ gimple_fold_call (gimple_stmt_iterator *
return true;
}
break;
+ case IFN_GOACC_DIM_SIZE:
+ case IFN_GOACC_DIM_POS:
+ result = fold_internal_goacc_dim (stmt);
+ break;
case IFN_UBSAN_CHECK_ADD:
subcode = PLUS_EXPR;
break;
^ permalink raw reply [flat|nested] 6+ messages in thread* Re: [OpenACC] internal fn folding 2015-10-28 18:41 [OpenACC] internal fn folding Nathan Sidwell @ 2015-11-02 13:56 ` Nathan Sidwell 2015-11-04 10:02 ` Bernd Schmidt 0 siblings, 1 reply; 6+ messages in thread From: Nathan Sidwell @ 2015-11-02 13:56 UTC (permalink / raw) To: GCC Patches, Richard Guenther On 10/28/15 14:40, Nathan Sidwell wrote: > Richard, > this patch adds folding for the new GOACC_DIM_POS and GOACC_DIM_SIZE internal > functions. IIUC gimple_fold_call is the right place to add this. > > The size of a compute dimension is very often a compile-time constant. On the > host, in particular it's 1, which means we can deduce the POS must be zero. > > ok? https://gcc.gnu.org/ml/gcc-patches/2015-10/threads.html#03095 Ping? nathan ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [OpenACC] internal fn folding 2015-11-02 13:56 ` Nathan Sidwell @ 2015-11-04 10:02 ` Bernd Schmidt 2015-11-05 15:48 ` Nathan Sidwell 0 siblings, 1 reply; 6+ messages in thread From: Bernd Schmidt @ 2015-11-04 10:02 UTC (permalink / raw) To: Nathan Sidwell, GCC Patches, Richard Guenther On 11/02/2015 02:56 PM, Nathan Sidwell wrote: > On 10/28/15 14:40, Nathan Sidwell wrote: >> Richard, >> this patch adds folding for the new GOACC_DIM_POS and GOACC_DIM_SIZE >> internal >> functions. IIUC gimple_fold_call is the right place to add this. >> >> The size of a compute dimension is very often a compile-time >> constant. On the >> host, in particular it's 1, which means we can deduce the POS must be >> zero. >> >> ok? > > https://gcc.gnu.org/ml/gcc-patches/2015-10/threads.html#03095 > > Ping? Ok. Bernd ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [OpenACC] internal fn folding 2015-11-04 10:02 ` Bernd Schmidt @ 2015-11-05 15:48 ` Nathan Sidwell 2015-11-08 14:04 ` Thomas Schwinge 0 siblings, 1 reply; 6+ messages in thread From: Nathan Sidwell @ 2015-11-05 15:48 UTC (permalink / raw) To: Bernd Schmidt, GCC Patches, Richard Guenther [-- Attachment #1: Type: text/plain, Size: 644 bytes --] On 11/04/15 05:02, Bernd Schmidt wrote: > On 11/02/2015 02:56 PM, Nathan Sidwell wrote: >> On 10/28/15 14:40, Nathan Sidwell wrote: >>> Richard, >>> this patch adds folding for the new GOACC_DIM_POS and GOACC_DIM_SIZE >>> internal >>> functions. IIUC gimple_fold_call is the right place to add this. >>> >>> The size of a compute dimension is very often a compile-time >>> constant. On the >>> host, in particular it's 1, which means we can deduce the POS must be >>> zero. >>> >>> ok? This is what I committed, using the helpers I recently added. (I realized we can only get here for functions with the oacc attribute already set) nathan [-- Attachment #2: trunk-dim-fold-1105.patch --] [-- Type: text/x-patch, Size: 1777 bytes --] 2015-11-05 Nathan Sidwell <nathan@codesourcery.com> * gimple-fold.c: Include omp-low.h. (fold_internal_goacc_dim): New. (gimple_fold_call): Call it. Index: gimple-fold.c =================================================================== --- gimple-fold.c (revision 229809) +++ gimple-fold.c (working copy) @@ -52,6 +52,7 @@ along with GCC; see the file COPYING3. #include "gimple-match.h" #include "gomp-constants.h" #include "optabs-query.h" +#include "omp-low.h" /* Return true when DECL can be referenced from current unit. @@ -2906,6 +2907,28 @@ gimple_fold_builtin (gimple_stmt_iterato return false; } +/* Transform IFN_GOACC_DIM_SIZE and IFN_GOACC_DIM_POS internal + function calls to constants, where possible. */ + +static tree +fold_internal_goacc_dim (const gimple *call) +{ + int axis = get_oacc_ifn_dim_arg (call); + int size = get_oacc_fn_dim_size (current_function_decl, axis); + bool is_pos = gimple_call_internal_fn (call) == IFN_GOACC_DIM_POS; + tree result = NULL_TREE; + + /* If the size is 1, or we only want the size and it is not dynamic, + we know the answer. */ + if (size == 1 || (!is_pos && size)) + { + tree type = TREE_TYPE (gimple_call_lhs (call)); + result = build_int_cst (type, size - is_pos); + } + + return result; +} + /* Return true if ARG0 CODE ARG1 in infinite signed precision operation doesn't fit into TYPE. The test for overflow should be regardless of -fwrapv, and even for unsigned types. */ @@ -3106,6 +3129,10 @@ gimple_fold_call (gimple_stmt_iterator * return true; } break; + case IFN_GOACC_DIM_SIZE: + case IFN_GOACC_DIM_POS: + result = fold_internal_goacc_dim (stmt); + break; case IFN_UBSAN_CHECK_ADD: subcode = PLUS_EXPR; break; ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [OpenACC] internal fn folding 2015-11-05 15:48 ` Nathan Sidwell @ 2015-11-08 14:04 ` Thomas Schwinge 2015-11-08 15:44 ` Tom de Vries 0 siblings, 1 reply; 6+ messages in thread From: Thomas Schwinge @ 2015-11-08 14:04 UTC (permalink / raw) To: Nathan Sidwell, GCC Patches, Tom de Vries; +Cc: Bernd Schmidt, Richard Guenther [-- Attachment #1: Type: text/plain, Size: 5328 bytes --] Hi! On Thu, 5 Nov 2015 10:48:02 -0500, Nathan Sidwell <nathan@acm.org> wrote: > On 11/04/15 05:02, Bernd Schmidt wrote: > > On 11/02/2015 02:56 PM, Nathan Sidwell wrote: > >> On 10/28/15 14:40, Nathan Sidwell wrote: > >>> Richard, > >>> this patch adds folding for the new GOACC_DIM_POS and GOACC_DIM_SIZE > >>> internal > >>> functions. IIUC gimple_fold_call is the right place to add this. > >>> > >>> The size of a compute dimension is very often a compile-time > >>> constant. On the > >>> host, in particular it's 1, which means we can deduce the POS must be > >>> zero. > This is what I committed, using the helpers I recently added. (I realized we can > only get here for functions with the oacc attribute already set) > --- gimple-fold.c (revision 229809) > +++ gimple-fold.c (working copy) > +/* Transform IFN_GOACC_DIM_SIZE and IFN_GOACC_DIM_POS internal > + function calls to constants, where possible. */ > + > +static tree > +fold_internal_goacc_dim (const gimple *call) > +{ > + int axis = get_oacc_ifn_dim_arg (call); > + int size = get_oacc_fn_dim_size (current_function_decl, axis); > + bool is_pos = gimple_call_internal_fn (call) == IFN_GOACC_DIM_POS; > + tree result = NULL_TREE; > + > + /* If the size is 1, or we only want the size and it is not dynamic, > + we know the answer. */ > + if (size == 1 || (!is_pos && size)) > + { > + tree type = TREE_TYPE (gimple_call_lhs (call)); > + result = build_int_cst (type, size - is_pos); > + } > + > + return result; > +} > @@ -3106,6 +3129,10 @@ gimple_fold_call (gimple_stmt_iterator * > return true; > } > break; > + case IFN_GOACC_DIM_SIZE: > + case IFN_GOACC_DIM_POS: > + result = fold_internal_goacc_dim (stmt); > + break; Merging this into gomp-4_0-branch, we'd run into a lot of regressions (for OpenACC kernels construct), for example: [...]/source-gcc/gcc/testsuite/c-c++-common/goacc/kernels-loop-g.c: In function 'main._omp_fn.0': [...]/source-gcc/gcc/testsuite/c-c++-common/goacc/kernels-loop-g.c:8:0: internal compiler error: Segmentation fault 0xac387f crash_signal [...]/source-gcc/gcc/toplev.c:336 0x9b8399 tree_int_cst_elt_check [...]/source-gcc/gcc/tree.h:3129 0x9b8399 get_oacc_fn_dim_size(tree_node*, int) [...]/source-gcc/gcc/omp-low.c:12630 0x86c530 fold_internal_goacc_dim [...]/source-gcc/gcc/gimple-fold.c:2917 0x86c530 gimple_fold_call [...]/source-gcc/gcc/gimple-fold.c:3134 0x86dfe4 fold_stmt_1 [...]/source-gcc/gcc/gimple-fold.c:3702 0xbf9953 execute [...]/source-gcc/gcc/tree-ssa-forwprop.c:2310 The dims in gcc/omp-low.c:get_oacc_fn_dim_size don't have values set, so the "TREE_INT_CST_LOW (TREE_VALUE (dims))" fails. I have not analyzed what exactly is going wrong; I just figured out that it's related to the IFN_GOACC_DIM_POS without LHS usage that Tom introduced in gomp-4_0-branch r228735 for OpenACC kernels to "neuter gang-single code in gang-redundant mode", <http://news.gmane.org/find-root.php?message_id=%3C561C1336.2050401%40mentor.com%3E>. So, in r229948 I merged Nathan's trunk r229816 into gomp-4_0-branch with an additional hack as indicated ("++" prefix) by the following three-way diff: commit b421a6415fc223866bc97f8248a1fbd0a524505e Merge: 7a6eb6b b0ccb4e Author: tschwinge <tschwinge@138bc75d-0d04-0410-961f-82ee72b054a4> Date: Sun Nov 8 13:49:46 2015 +0000 svn merge -r 229814:229816 svn+ssh://gcc.gnu.org/svn/gcc/trunk git-svn-id: svn+ssh://gcc.gnu.org/svn/gcc/branches/gomp-4_0-branch@229948 138bc75d-0d04-0410-961f-82ee72b054a4 gcc/ChangeLog | 6 ++++++ gcc/gimple-fold.c | 33 +++++++++++++++++++++++++++++++++ 2 files changed, 39 insertions(+) diff --cc gcc/gimple-fold.c index c9b9593,45840af..869c6c2 --- gcc/gimple-fold.c +++ gcc/gimple-fold.c @@@ -2906,6 -2907,28 +2907,34 @@@ gimple_fold_builtin (gimple_stmt_iterat return false; } + /* Transform IFN_GOACC_DIM_SIZE and IFN_GOACC_DIM_POS internal + function calls to constants, where possible. */ + + static tree + fold_internal_goacc_dim (const gimple *call) + { ++ /* TODO. There is something going wrong here, for the gang_single ++ IFN_GOACC_DIM_POS without LHS, generated in gcc/omp-low.c:lower_omp_target ++ for is_oacc_kernels (see gomp-4_0-branch r228735). */ ++ if (gimple_call_lhs (call) == NULL_TREE) ++ return NULL_TREE; ++ + int axis = get_oacc_ifn_dim_arg (call); + int size = get_oacc_fn_dim_size (current_function_decl, axis); + bool is_pos = gimple_call_internal_fn (call) == IFN_GOACC_DIM_POS; + tree result = NULL_TREE; + + /* If the size is 1, or we only want the size and it is not dynamic, + we know the answer. */ + if (size == 1 || (!is_pos && size)) + { + tree type = TREE_TYPE (gimple_call_lhs (call)); + result = build_int_cst (type, size - is_pos); + } + + return result; + } + /* Return true if ARG0 CODE ARG1 in infinite signed precision operation doesn't fit into TYPE. The test for overflow should be regardless of -fwrapv, and even for unsigned types. */ Grüße Thomas [-- Attachment #2: signature.asc --] [-- Type: application/pgp-signature, Size: 472 bytes --] ^ permalink raw reply [flat|nested] 6+ messages in thread
* Re: [OpenACC] internal fn folding 2015-11-08 14:04 ` Thomas Schwinge @ 2015-11-08 15:44 ` Tom de Vries 0 siblings, 0 replies; 6+ messages in thread From: Tom de Vries @ 2015-11-08 15:44 UTC (permalink / raw) To: Thomas Schwinge, Nathan Sidwell, GCC Patches, Tom de Vries Cc: Bernd Schmidt, Richard Guenther On 08/11/15 15:04, Thomas Schwinge wrote: > Hi! > > On Thu, 5 Nov 2015 10:48:02 -0500, Nathan Sidwell<nathan@acm.org> wrote: >> >On 11/04/15 05:02, Bernd Schmidt wrote: >>> > >On 11/02/2015 02:56 PM, Nathan Sidwell wrote: >>>> > >>On 10/28/15 14:40, Nathan Sidwell wrote: >>>>> > >>>Richard, >>>>> > >>>this patch adds folding for the new GOACC_DIM_POS and GOACC_DIM_SIZE >>>>> > >>>internal >>>>> > >>>functions. IIUC gimple_fold_call is the right place to add this. >>>>> > >>> >>>>> > >>>The size of a compute dimension is very often a compile-time >>>>> > >>>constant. On the >>>>> > >>>host, in particular it's 1, which means we can deduce the POS must be >>>>> > >>>zero. >> >This is what I committed, using the helpers I recently added. (I realized we can >> >only get here for functions with the oacc attribute already set) >> >--- gimple-fold.c (revision 229809) >> >+++ gimple-fold.c (working copy) >> >+/* Transform IFN_GOACC_DIM_SIZE and IFN_GOACC_DIM_POS internal >> >+ function calls to constants, where possible. */ >> >+ >> >+static tree >> >+fold_internal_goacc_dim (const gimple *call) >> >+{ >> >+ int axis = get_oacc_ifn_dim_arg (call); >> >+ int size = get_oacc_fn_dim_size (current_function_decl, axis); >> >+ bool is_pos = gimple_call_internal_fn (call) == IFN_GOACC_DIM_POS; >> >+ tree result = NULL_TREE; >> >+ >> >+ /* If the size is 1, or we only want the size and it is not dynamic, >> >+ we know the answer. */ >> >+ if (size == 1 || (!is_pos && size)) >> >+ { >> >+ tree type = TREE_TYPE (gimple_call_lhs (call)); >> >+ result = build_int_cst (type, size - is_pos); >> >+ } >> >+ >> >+ return result; >> >+} >> >@@ -3106,6 +3129,10 @@ gimple_fold_call (gimple_stmt_iterator * >> > return true; >> > } >> > break; >> >+ case IFN_GOACC_DIM_SIZE: >> >+ case IFN_GOACC_DIM_POS: >> >+ result = fold_internal_goacc_dim (stmt); >> >+ break; > Merging this into gomp-4_0-branch, we'd run into a lot of regressions > (for OpenACC kernels construct), for example: > > [...]/source-gcc/gcc/testsuite/c-c++-common/goacc/kernels-loop-g.c: In function 'main._omp_fn.0': > [...]/source-gcc/gcc/testsuite/c-c++-common/goacc/kernels-loop-g.c:8:0: internal compiler error: Segmentation fault > 0xac387f crash_signal > [...]/source-gcc/gcc/toplev.c:336 > 0x9b8399 tree_int_cst_elt_check > [...]/source-gcc/gcc/tree.h:3129 > 0x9b8399 get_oacc_fn_dim_size(tree_node*, int) > [...]/source-gcc/gcc/omp-low.c:12630 > 0x86c530 fold_internal_goacc_dim > [...]/source-gcc/gcc/gimple-fold.c:2917 > 0x86c530 gimple_fold_call > [...]/source-gcc/gcc/gimple-fold.c:3134 > 0x86dfe4 fold_stmt_1 > [...]/source-gcc/gcc/gimple-fold.c:3702 > 0xbf9953 execute > [...]/source-gcc/gcc/tree-ssa-forwprop.c:2310 > > The dims in gcc/omp-low.c:get_oacc_fn_dim_size don't have values set, so > the "TREE_INT_CST_LOW (TREE_VALUE (dims))" fails. I have not analyzed > what exactly is going wrong; I just figured out that it's related to the > IFN_GOACC_DIM_POS without LHS usage that Tom introduced in > gomp-4_0-branch r228735 for OpenACC kernels to "neuter gang-single code > in gang-redundant mode", > <http://news.gmane.org/find-root.php?message_id=%3C561C1336.2050401%40mentor.com%3E>. I've just removed that ( https://gcc.gnu.org/ml/gcc-patches/2015-11/msg00847.html ). Thanks, - Tom ^ permalink raw reply [flat|nested] 6+ messages in thread
end of thread, other threads:[~2015-11-08 15:44 UTC | newest] Thread overview: 6+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2015-10-28 18:41 [OpenACC] internal fn folding Nathan Sidwell 2015-11-02 13:56 ` Nathan Sidwell 2015-11-04 10:02 ` Bernd Schmidt 2015-11-05 15:48 ` Nathan Sidwell 2015-11-08 14:04 ` Thomas Schwinge 2015-11-08 15:44 ` Tom de Vries
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).