+2015-08-27 Sebastian Pop <s.pop@samsung.com>
+
+ * Makefile.in (OBJS): Remove graphite-blocking.o and
+ graphite-interchange.o.
+ * common.opt (floop-strip-mine, floop-interchange, floop-block):
+ Alias of floop-nest-optimize.
+ * doc/invoke.texi (floop-strip-mine, floop-interchange, floop-block):
+ Document as alias of -floop-nest-optimize.
+ * graphite-blocking.c: Remove.
+ * graphite-interchange.c: Remove.
+ * graphite-optimize-isl.c: Include dumpfile.h.
+ (getScheduleForBand): Add dump for tiled loops. Use
+ PARAM_LOOP_BLOCK_TILE_SIZE instead of hard coded constant.
+ * graphite-poly.c (scop_max_loop_depth): Remove.
+ (print_scattering_function_1): Remove.
+ (print_scattering_function): Remove.
+ (print_scattering_functions): Remove.
+ (debug_scattering_function): Remove.
+ (debug_scattering_functions): Remove.
+ (apply_poly_transforms): Remove use of flag_loop_block,
+ flag_loop_strip_mine, and flag_loop_interchange.
+ (new_poly_bb): Remove use of PBB_TRANSFORMED, PBB_SAVED, and
+ PBB_ORIGINAL.
+ (print_pdr_access_layout): Remove.
+ (print_pdr): Print ISL representation.
+ (new_scop): Remove use of SCOP_ORIGINAL_SCHEDULE,
+ SCOP_TRANSFORMED_SCHEDULE, and SCOP_SAVED_SCHEDULE.
+ (free_scop): Same.
+ (openscop_print_pbb_domain): Remove.
+ (print_pbb): Remove call to print_scattering_function.
+ (openscop_print_scop_context): Remove.
+ (print_scop_context): Do not print matrices anymore.
+ (print_scop): Do not print SCOP_ORIGINAL_SCHEDULE and
+ SCOP_TRANSFORMED_SCHEDULE.
+ (print_isl_set): Add printing of a new line.
+ (print_isl_map): Same.
+ (print_isl_aff): Same.
+ (print_isl_constraint): Same.
+ (loop_to_lst): Remove.
+ (scop_to_lst): Remove.
+ (lst_indent_to): Remove.
+ (print_lst): Remove.
+ (debug_lst): Remove.
+ (dot_lst_1): Remove.
+ (dot_lst): Remove.
+ (reverse_loop_at_level): Remove.
+ (reverse_loop_for_pbbs): Remove.
+ * graphite-poly.h (pdr_dim_iter_domain): Remove.
+ (pdr_nb_params): Remove.
+ (pdr_alias_set_dim): Remove.
+ (pdr_subscript_dim): Remove.
+ (pdr_iterator_dim): Remove.
+ (pdr_parameter_dim): Remove.
+ (same_pdr_p): Remove.
+ (struct poly_scattering): Remove.
+ (struct poly_bb): Remove _original, _transformed, _saved.
+ (PBB_DOMAIN, PBB_ORIGINAL, PBB_ORIGINAL_SCATTERING): Remove.
+ (PBB_TRANSFORMED, PBB_TRANSFORMED_SCATTERING, PBB_SAVED): Remove.
+ (PBB_NB_LOCAL_VARIABLES): Remove.
+ (PBB_NB_SCATTERING_TRANSFORM): Remove.
+ (schedule_to_scattering): Remove.
+ (number_of_write_pdrs): Remove.
+ (pbb_dim_iter_domain): Remove.
+ (pbb_nb_params): Remove.
+ (pbb_nb_scattering_orig): Remove.
+ (pbb_nb_scattering_transform): Remove.
+ (pbb_nb_dynamic_scattering_transform): Remove.
+ (pbb_nb_local_vars): Remove.
+ (pbb_iterator_dim): Remove.
+ (pbb_parameter_dim): Remove.
+ (psco_scattering_dim): Remove.
+ (psct_scattering_dim): Remove.
+ (psct_local_var_dim): Remove.
+ (psco_iterator_dim): Remove.
+ (psct_iterator_dim): Remove.
+ (psco_parameter_dim): Remove.
+ (psct_parameter_dim): Remove.
+ (psct_dynamic_dim): Remove.
+ (psct_static_dim): Remove.
+ (psct_add_local_variable): Remove.
+ (new_lst_loop): Remove.
+ (new_lst_stmt): Remove.
+ (free_lst): Remove.
+ (copy_lst): Remove.
+ (lst_add_loop_under_loop): Remove.
+ (lst_depth): Remove.
+ (lst_dewey_number): Remove.
+ (lst_dewey_number_at_depth): Remove.
+ (lst_pred): Remove.
+ (lst_succ): Remove.
+ (lst_find_pbb): Remove.
+ (find_lst_loop): Remove.
+ (lst_find_first_pbb): Remove.
+ (lst_empty_p): Remove.
+ (lst_find_last_pbb): Remove.
+ (lst_contains_p): Remove.
+ (lst_contains_pbb): Remove.
+ (lst_create_nest): Remove.
+ (lst_remove_from_sequence): Remove.
+ (lst_remove_loop_and_inline_stmts_in_loop_father): Remove.
+ (lst_niter_for_loop): Remove.
+ (pbb_update_scattering): Remove.
+ (lst_update_scattering_under): Remove.
+ (lst_update_scattering): Remove.
+ (lst_insert_in_sequence): Remove.
+ (lst_replace): Remove.
+ (lst_substitute_3): Remove.
+ (lst_distribute_lst): Remove.
+ (lst_remove_all_before_including_pbb): Remove.
+ (lst_remove_all_before_excluding_pbb): Remove.
+ (struct scop): Remove original_schedule, transformed_schedule, and
+ saved_schedule.
+ (SCOP_ORIGINAL_SCHEDULE, SCOP_TRANSFORMED_SCHEDULE): Remove.
+ (SCOP_SAVED_SCHEDULE): Remove.
+ (poly_scattering_new): Remove.
+ (poly_scattering_free): Remove.
+ (poly_scattering_copy): Remove.
+ (store_scattering_pbb): Remove.
+ (store_lst_schedule): Remove.
+ (restore_lst_schedule): Remove.
+ (store_scattering): Remove.
+ (restore_scattering_pbb): Remove.
+ (restore_scattering): Remove.
+ * graphite-sese-to-poly.c (build_pbb_scattering_polyhedrons):
+ Remove scattering_dimensions. Do not use pbb_dim_iter_domain:
+ compute the scattering polyhedron dimension from the dimension of
+ pbb->domain.
+ (build_scop_scattering): Update call to
+ build_pbb_scattering_polyhedrons.
+ (build_poly_scop): Remove call to scop_to_lst.
+ * graphite.c (graphite_transform_loops): Add call to print_scop.
+ (gate_graphite_transforms): Remove use of flag_loop_block,
+ flag_loop_interchange, and flag_loop_strip_mine.
+
2015-08-27 Sebastian Pop <s.pop@samsung.com>
* common.opt (floop-unroll-and-jam): Make alias of floop-nest-optimize.
graph.o \
graphds.o \
graphite.o \
- graphite-blocking.o \
graphite-isl-ast-to-gimple.o \
graphite-dependences.o \
- graphite-interchange.o \
graphite-optimize-isl.o \
graphite-poly.o \
graphite-scop-detection.o \
Mark all loops as parallel
floop-strip-mine
-Common Report Var(flag_loop_strip_mine) Optimization
-Enable Loop Strip Mining transformation
+Common Alias(floop-nest-optimize)
+Enable loop nest transforms. Same as -floop-nest-optimize
floop-interchange
-Common Report Var(flag_loop_interchange) Optimization
-Enable Loop Interchange transformation
+Common Alias(floop-nest-optimize)
+Enable loop nest transforms. Same as -floop-nest-optimize
floop-block
-Common Report Var(flag_loop_block) Optimization
-Enable Loop Blocking transformation
+Common Alias(floop-nest-optimize)
+Enable loop nest transforms. Same as -floop-nest-optimize
floop-unroll-and-jam
Common Alias(floop-nest-optimize)
Enable loop invariant motion on trees
ftree-loop-linear
-Common Alias(floop-interchange)
-Enable loop interchange transforms. Same as -floop-interchange
+Common Alias(floop-nest-optimize)
+Enable loop nest transforms. Same as -floop-nest-optimize
ftree-loop-ivcanon
Common Report Var(flag_tree_loop_ivcanon) Init(1) Optimization
at @option{-O} and higher.
@item -ftree-loop-linear
+@itemx -floop-interchange
+@itemx -floop-strip-mine
+@itemx -floop-block
+@itemx -floop-unroll-and-jam
@opindex ftree-loop-linear
-Perform loop interchange transformations on tree. Same as
-@option{-floop-interchange}. To use this code transformation, GCC has
-to be configured with @option{--with-isl} to enable the Graphite loop
-transformation infrastructure.
-
-@item -floop-interchange
@opindex floop-interchange
-Perform loop interchange transformations on loops. Interchanging two
-nested loops switches the inner and outer loops. For example, given a
-loop like:
-@smallexample
-DO J = 1, M
- DO I = 1, N
- A(J, I) = A(J, I) * C
- ENDDO
-ENDDO
-@end smallexample
-@noindent
-loop interchange transforms the loop as if it were written:
-@smallexample
-DO I = 1, N
- DO J = 1, M
- A(J, I) = A(J, I) * C
- ENDDO
-ENDDO
-@end smallexample
-which can be beneficial when @code{N} is larger than the caches,
-because in Fortran, the elements of an array are stored in memory
-contiguously by column, and the original loop iterates over rows,
-potentially creating at each access a cache miss. This optimization
-applies to all the languages supported by GCC and is not limited to
-Fortran. To use this code transformation, GCC has to be configured
-with @option{--with-isl} to enable the Graphite loop transformation
-infrastructure.
-
-@item -floop-strip-mine
@opindex floop-strip-mine
-Perform loop strip mining transformations on loops. Strip mining
-splits a loop into two nested loops. The outer loop has strides
-equal to the strip size and the inner loop has strides of the
-original loop within a strip. The strip length can be changed
-using the @option{loop-block-tile-size} parameter. For example,
-given a loop like:
-@smallexample
-DO I = 1, N
- A(I) = A(I) + C
-ENDDO
-@end smallexample
-@noindent
-loop strip mining transforms the loop as if it were written:
-@smallexample
-DO II = 1, N, 51
- DO I = II, min (II + 50, N)
- A(I) = A(I) + C
- ENDDO
-ENDDO
-@end smallexample
-This optimization applies to all the languages supported by GCC and is
-not limited to Fortran. To use this code transformation, GCC has to
-be configured with @option{--with-isl} to enable the Graphite loop
-transformation infrastructure.
-
-@item -floop-block
@opindex floop-block
-Perform loop blocking transformations on loops. Blocking strip mines
-each loop in the loop nest such that the memory accesses of the
-element loops fit inside caches. The strip length can be changed
-using the @option{loop-block-tile-size} parameter. For example, given
-a loop like:
-@smallexample
-DO I = 1, N
- DO J = 1, M
- A(J, I) = B(I) + C(J)
- ENDDO
-ENDDO
-@end smallexample
-@noindent
-loop blocking transforms the loop as if it were written:
-@smallexample
-DO II = 1, N, 51
- DO JJ = 1, M, 51
- DO I = II, min (II + 50, N)
- DO J = JJ, min (JJ + 50, M)
- A(J, I) = B(I) + C(J)
- ENDDO
- ENDDO
- ENDDO
-ENDDO
-@end smallexample
-which can be beneficial when @code{M} is larger than the caches,
-because the innermost loop iterates over a smaller amount of data
-which can be kept in the caches. This optimization applies to all the
-languages supported by GCC and is not limited to Fortran. To use this
-code transformation, GCC has to be configured with @option{--with-isl}
-to enable the Graphite loop transformation infrastructure.
+@opindex floop-unroll-and-jam
+Perform loop nest optimizations. Same as
+@option{-floop-nest-optimize}. To use this code transformation, GCC has
+to be configured with @option{--with-isl} to enable the Graphite loop
+transformation infrastructure.
@item -fgraphite-identity
@opindex fgraphite-identity
structure optimized for data-locality and parallelism. This option
is experimental.
-@item -floop-unroll-and-jam
-@opindex floop-unroll-and-jam
-Perform loop nest transformations. Same as
-@option{-floop-nest-optimize}. To use this code transformation, GCC has
-to be configured with @option{--with-isl} to enable the Graphite loop
-transformation infrastructure.
-
@item -floop-parallelize-all
@opindex floop-parallelize-all
Use the Graphite data dependence analysis to identify loops that can
+++ /dev/null
-/* Heuristics and transform for loop blocking and strip mining on
- polyhedral representation.
-
- Copyright (C) 2009-2015 Free Software Foundation, Inc.
- Contributed by Sebastian Pop <sebastian.pop@amd.com> and
- Pranav Garg <pranav.garg2107@gmail.com>.
-
-This file is part of GCC.
-
-GCC is free software; you can redistribute it and/or modify
-it under the terms of the GNU General Public License as published by
-the Free Software Foundation; either version 3, or (at your option)
-any later version.
-
-GCC is distributed in the hope that it will be useful,
-but WITHOUT ANY WARRANTY; without even the implied warranty of
-MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
-GNU General Public License for more details.
-
-You should have received a copy of the GNU General Public License
-along with GCC; see the file COPYING3. If not see
-<http://www.gnu.org/licenses/>. */
-
-#include "config.h"
-
-#ifdef HAVE_isl
-/* Workaround for GMP 5.1.3 bug, see PR56019. */
-#include <stddef.h>
-
-#include <isl/constraint.h>
-#include <isl/set.h>
-#include <isl/map.h>
-#include <isl/union_map.h>
-#include <isl/constraint.h>
-
-#include "system.h"
-#include "coretypes.h"
-#include "backend.h"
-#include "cfghooks.h"
-#include "tree.h"
-#include "gimple.h"
-#include "params.h"
-#include "fold-const.h"
-#include "gimple-iterator.h"
-#include "tree-ssa-loop.h"
-#include "dumpfile.h"
-#include "cfgloop.h"
-#include "tree-data-ref.h"
-#include "graphite-poly.h"
-
-/* Strip mines with a factor STRIDE the scattering (time) dimension
- around PBB at depth TIME_DEPTH.
-
- The following example comes from the wiki page:
- http://gcc.gnu.org/wiki/Graphite/Strip_mine
-
- The strip mine of a loop with a tile of 64 can be obtained with a
- scattering function as follows:
-
- $ cat ./albert_strip_mine.cloog
- # language: C
- c
-
- # parameter {n | n >= 0}
- 1 3
- # n 1
- 1 1 0
- 1
- n
-
- 1 # Number of statements:
-
- 1
- # {i | 0 <= i <= n}
- 2 4
- # i n 1
- 1 1 0 0
- 1 -1 1 0
-
- 0 0 0
- 1
- i
-
- 1 # Scattering functions
-
- 3 6
- # NEW OLD i n 1
- 1 -64 0 1 0 0
- 1 64 0 -1 0 63
- 0 0 1 -1 0 0
-
- 1
- NEW OLD
-
- #the output of CLooG is like this:
- #$ cloog ./albert_strip_mine.cloog
- # for (NEW=0;NEW<=floord(n,64);NEW++) {
- # for (OLD=max(64*NEW,0);OLD<=min(64*NEW+63,n);OLD++) {
- # S1(i = OLD) ;
- # }
- # }
-*/
-
-static void
-pbb_strip_mine_time_depth (poly_bb_p pbb, int time_depth, int stride)
-{
- isl_space *d;
- isl_constraint *c;
- int iter, strip;
- /* STRIP is the dimension that iterates with stride STRIDE. */
- /* ITER is the dimension that enumerates single iterations inside
- one strip that has at most STRIDE iterations. */
- strip = time_depth;
- iter = strip + 2;
-
- pbb->transformed = isl_map_insert_dims (pbb->transformed, isl_dim_out,
- strip, 2);
-
- /* Lower bound of the striped loop. */
- d = isl_map_get_space (pbb->transformed);
- c = isl_inequality_alloc (isl_local_space_from_space (d));
- c = isl_constraint_set_coefficient_si (c, isl_dim_out, strip, -stride);
- c = isl_constraint_set_coefficient_si (c, isl_dim_out, iter, 1);
- pbb->transformed = isl_map_add_constraint (pbb->transformed, c);
-
- /* Upper bound of the striped loop. */
- d = isl_map_get_space (pbb->transformed);
- c = isl_inequality_alloc (isl_local_space_from_space (d));
- c = isl_constraint_set_coefficient_si (c, isl_dim_out, strip, stride);
- c = isl_constraint_set_coefficient_si (c, isl_dim_out, iter, -1);
- c = isl_constraint_set_constant_si (c, stride - 1);
- pbb->transformed = isl_map_add_constraint (pbb->transformed, c);
-
- /* Static scheduling for ITER level.
- This is mandatory to keep the 2d + 1 canonical scheduling format. */
- d = isl_map_get_space (pbb->transformed);
- c = isl_equality_alloc (isl_local_space_from_space (d));
- c = isl_constraint_set_coefficient_si (c, isl_dim_out, strip + 1, 1);
- pbb->transformed = isl_map_add_constraint (pbb->transformed, c);
-}
-
-/* Returns true when strip mining with STRIDE of the loop LST is
- profitable. */
-
-static bool
-lst_strip_mine_profitable_p (lst_p lst, int stride)
-{
- mpz_t niter, strip_stride;
- bool res;
-
- gcc_assert (LST_LOOP_P (lst));
- mpz_init (strip_stride);
- mpz_init (niter);
-
- mpz_set_si (strip_stride, stride);
- lst_niter_for_loop (lst, niter);
- res = (mpz_cmp (niter, strip_stride) > 0);
-
- mpz_clear (strip_stride);
- mpz_clear (niter);
- return res;
-}
-
-/* Strip-mines all the loops of LST with STRIDE. Return the number of
- loops strip-mined. */
-
-static int
-lst_do_strip_mine_loop (lst_p lst, int depth, int stride)
-{
- int i;
- lst_p l;
- poly_bb_p pbb;
-
- if (!lst)
- return 0;
-
- if (LST_LOOP_P (lst))
- {
- int res = 0;
-
- FOR_EACH_VEC_ELT (LST_SEQ (lst), i, l)
- res += lst_do_strip_mine_loop (l, depth, stride);
-
- return res;
- }
-
- pbb = LST_PBB (lst);
- pbb_strip_mine_time_depth (pbb, psct_dynamic_dim (pbb, depth), stride);
- return 1;
-}
-
-/* Strip-mines all the loops of LST with STRIDE. When STRIDE is zero,
- read the stride from the PARAM_LOOP_BLOCK_TILE_SIZE. Return the
- number of strip-mined loops.
-
- Strip mining transforms a loop
-
- | for (i = 0; i < N; i++)
- | S (i);
-
- into the following loop nest:
-
- | for (k = 0; k < N; k += STRIDE)
- | for (j = 0; j < STRIDE; j++)
- | S (i = k + j);
-*/
-
-static int
-lst_do_strip_mine (lst_p lst, int stride)
-{
- int i;
- lst_p l;
- int res = 0;
- int depth;
-
- if (!stride)
- stride = PARAM_VALUE (PARAM_LOOP_BLOCK_TILE_SIZE);
-
- if (!lst
- || !LST_LOOP_P (lst))
- return false;
-
- FOR_EACH_VEC_ELT (LST_SEQ (lst), i, l)
- res += lst_do_strip_mine (l, stride);
-
- depth = lst_depth (lst);
- if (depth >= 0
- && lst_strip_mine_profitable_p (lst, stride))
- {
- res += lst_do_strip_mine_loop (lst, lst_depth (lst), stride);
- lst_add_loop_under_loop (lst);
- }
-
- return res;
-}
-
-/* Strip mines all the loops in SCOP. Returns the number of
- strip-mined loops. */
-
-int
-scop_do_strip_mine (scop_p scop, int stride)
-{
- return lst_do_strip_mine (SCOP_TRANSFORMED_SCHEDULE (scop), stride);
-}
-
-/* Loop blocks all the loops in SCOP. Returns true when we manage to
- block some loops. */
-
-bool
-scop_do_block (scop_p scop)
-{
- store_scattering (scop);
-
- /* If we don't strip mine at least two loops, or not interchange
- loops, the strip mine alone will not be profitable, and the
- transform is not a loop blocking: so revert the transform. */
- if (lst_do_strip_mine (SCOP_TRANSFORMED_SCHEDULE (scop), 0) < 2
- || scop_do_interchange (scop) == 0)
- {
- restore_scattering (scop);
- return false;
- }
-
- if (dump_file && (dump_flags & TDF_DETAILS))
- fprintf (dump_file, "SCoP will be loop blocked.\n");
-
- return true;
-}
-
-#endif
+++ /dev/null
-/* Interchange heuristics and transform for loop interchange on
- polyhedral representation.
-
- Copyright (C) 2009-2015 Free Software Foundation, Inc.
- Contributed by Sebastian Pop <sebastian.pop@amd.com> and
- Harsha Jagasia <harsha.jagasia@amd.com>.
-
-This file is part of GCC.
-
-GCC is free software; you can redistribute it and/or modify
-it under the terms of the GNU General Public License as published by
-the Free Software Foundation; either version 3, or (at your option)
-any later version.
-
-GCC is distributed in the hope that it will be useful,
-but WITHOUT ANY WARRANTY; without even the implied warranty of
-MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
-GNU General Public License for more details.
-
-You should have received a copy of the GNU General Public License
-along with GCC; see the file COPYING3. If not see
-<http://www.gnu.org/licenses/>. */
-
-#include "config.h"
-
-#ifdef HAVE_isl
-/* Workaround for GMP 5.1.3 bug, see PR56019. */
-#include <stddef.h>
-
-#include <isl/constraint.h>
-#include <isl/aff.h>
-#include <isl/set.h>
-#include <isl/map.h>
-#include <isl/union_map.h>
-#include <isl/ilp.h>
-#include <isl/val.h>
-
-/* Since ISL-0.13, the extern is in val_gmp.h. */
-#if !defined(HAVE_ISL_SCHED_CONSTRAINTS_COMPUTE_SCHEDULE) && defined(__cplusplus)
-extern "C" {
-#endif
-#include <isl/val_gmp.h>
-#if !defined(HAVE_ISL_SCHED_CONSTRAINTS_COMPUTE_SCHEDULE) && defined(__cplusplus)
-}
-#endif
-
-#include "system.h"
-#include "coretypes.h"
-#include "backend.h"
-#include "cfghooks.h"
-#include "tree.h"
-#include "gimple.h"
-#include "fold-const.h"
-#include "gimple-iterator.h"
-#include "tree-ssa-loop.h"
-#include "dumpfile.h"
-#include "cfgloop.h"
-#include "tree-data-ref.h"
-#include "graphite-poly.h"
-
-
-/* XXX isl rewrite following comment */
-/* Builds a linear expression, of dimension DIM, representing PDR's
- memory access:
-
- L = r_{n}*r_{n-1}*...*r_{1}*s_{0} + ... + r_{n}*s_{n-1} + s_{n}.
-
- For an array A[10][20] with two subscript locations s0 and s1, the
- linear memory access is 20 * s0 + s1: a stride of 1 in subscript s0
- corresponds to a memory stride of 20.
-
- OFFSET is a number of dimensions to prepend before the
- subscript dimensions: s_0, s_1, ..., s_n.
-
- Thus, the final linear expression has the following format:
- 0 .. 0_{offset} | 0 .. 0_{nit} | 0 .. 0_{gd} | 0 | c_0 c_1 ... c_n
- where the expression itself is:
- c_0 * s_0 + c_1 * s_1 + ... c_n * s_n. */
-
-static isl_constraint *
-build_linearized_memory_access (isl_map *map, poly_dr_p pdr)
-{
- isl_local_space *ls = isl_local_space_from_space (isl_map_get_space (map));
- isl_constraint *res = isl_equality_alloc (ls);
- isl_val *size = isl_val_int_from_ui (isl_map_get_ctx (map), 1);
-
- unsigned nsubs = isl_set_dim (pdr->subscript_sizes, isl_dim_set);
- /* -1 for the already included L dimension. */
- unsigned offset = isl_map_dim (map, isl_dim_out) - 1 - nsubs;
- res = isl_constraint_set_coefficient_si (res, isl_dim_out, offset + nsubs, -1);
- /* Go through all subscripts from last to first. The dimension "i=0"
- is the alias set, ignore it. */
- for (int i = nsubs - 1; i >= 1; i--)
- {
- isl_aff *extract_dim;
- res = isl_constraint_set_coefficient_val (res, isl_dim_out, offset + i,
- isl_val_copy (size));
- isl_space *dc = isl_set_get_space (pdr->subscript_sizes);
- extract_dim = isl_aff_zero_on_domain (isl_local_space_from_space (dc));
- extract_dim = isl_aff_set_coefficient_si (extract_dim, isl_dim_in, i, 1);
- isl_val *max = isl_set_max_val (pdr->subscript_sizes, extract_dim);
- isl_aff_free (extract_dim);
-
- /* The result is NULL in case of an error, the optimal value in case there
- is one, negative infinity or infinity if the problem is unbounded and
- NaN if the problem is empty. */
- gcc_assert (max);
-
- /* When one of the dimensions cannot be computed, we cannot build the size
- of the array for any outer dimensions. */
- if (!isl_val_is_int (max))
- {
- isl_val_free (max);
- break;
- }
- size = isl_val_mul (size, max);
- }
-
- isl_val_free (size);
-
- return res;
-}
-
-/* Set STRIDE to the stride of PDR in memory by advancing by one in
- the loop at DEPTH. */
-
-static void
-pdr_stride_in_loop (mpz_t stride, graphite_dim_t depth, poly_dr_p pdr)
-{
- poly_bb_p pbb = PDR_PBB (pdr);
- isl_map *map;
- isl_set *set;
- isl_aff *aff;
- isl_space *dc;
- isl_constraint *lma, *c;
- isl_val *islstride;
- graphite_dim_t time_depth;
- unsigned offset, nt;
- unsigned i;
- /* XXX isl rewrite following comments. */
- /* Builds a partial difference equations and inserts them
- into pointset powerset polyhedron P. Polyhedron is assumed
- to have the format: T|I|T'|I'|G|S|S'|l1|l2.
-
- TIME_DEPTH is the time dimension w.r.t. which we are
- differentiating.
- OFFSET represents the number of dimensions between
- columns t_{time_depth} and t'_{time_depth}.
- DIM_SCTR is the number of scattering dimensions. It is
- essentially the dimensionality of the T vector.
-
- The following equations are inserted into the polyhedron P:
- | t_1 = t_1'
- | ...
- | t_{time_depth-1} = t'_{time_depth-1}
- | t_{time_depth} = t'_{time_depth} + 1
- | t_{time_depth+1} = t'_{time_depth + 1}
- | ...
- | t_{dim_sctr} = t'_{dim_sctr}. */
-
- /* Add the equality: t_{time_depth} = t'_{time_depth} + 1.
- This is the core part of this alogrithm, since this
- constraint asks for the memory access stride (difference)
- between two consecutive points in time dimensions. */
-
- /* Add equalities:
- | t1 = t1'
- | ...
- | t_{time_depth-1} = t'_{time_depth-1}
- | t_{time_depth+1} = t'_{time_depth+1}
- | ...
- | t_{dim_sctr} = t'_{dim_sctr}
-
- This means that all the time dimensions are equal except for
- time_depth, where the constraint is t_{depth} = t'_{depth} + 1
- step. More to this: we should be careful not to add equalities
- to the 'coupled' dimensions, which happens when the one dimension
- is stripmined dimension, and the other dimension corresponds
- to the point loop inside stripmined dimension. */
-
- /* pdr->accesses: [P1..nb_param,I1..nb_domain]->[a,S1..nb_subscript]
- ??? [P] not used for PDRs?
- pdr->subscript_sizes: [a,S1..nb_subscript]
- pbb->domain: [P1..nb_param,I1..nb_domain]
- pbb->transformed: [P1..nb_param,I1..nb_domain]->[T1..Tnb_sctr]
- [T] includes local vars (currently unused)
-
- First we create [P,I] -> [T,a,S]. */
-
- map = isl_map_flat_range_product (isl_map_copy (pbb->transformed),
- isl_map_copy (pdr->accesses));
- /* Add a dimension for L: [P,I] -> [T,a,S,L].*/
- map = isl_map_add_dims (map, isl_dim_out, 1);
- /* Build a constraint for "lma[S] - L == 0", effectively calculating
- L in terms of subscripts. */
- lma = build_linearized_memory_access (map, pdr);
- /* And add it to the map, so we now have:
- [P,I] -> [T,a,S,L] : lma([S]) == L. */
- map = isl_map_add_constraint (map, lma);
-
- /* Then we create [P,I,P',I'] -> [T,a,S,L,T',a',S',L']. */
- map = isl_map_flat_product (map, isl_map_copy (map));
-
- /* Now add the equality T[time_depth] == T'[time_depth]+1. This will
- force L' to be the linear address at T[time_depth] + 1. */
- time_depth = psct_dynamic_dim (pbb, depth);
- /* Length of [a,S] plus [L] ... */
- offset = 1 + isl_map_dim (pdr->accesses, isl_dim_out);
- /* ... plus [T]. */
- offset += isl_map_dim (pbb->transformed, isl_dim_out);
-
- c = isl_equality_alloc (isl_local_space_from_space (isl_map_get_space (map)));
- c = isl_constraint_set_coefficient_si (c, isl_dim_out, time_depth, 1);
- c = isl_constraint_set_coefficient_si (c, isl_dim_out,
- offset + time_depth, -1);
- c = isl_constraint_set_constant_si (c, 1);
- map = isl_map_add_constraint (map, c);
-
- /* Now we equate most of the T/T' elements (making PITaSL nearly
- the same is (PITaSL)', except for one dimension, namely for 'depth'
- (an index into [I]), after translating to index into [T]. Take care
- to not produce an empty map, which indicates we wanted to equate
- two dimensions that are already coupled via the above time_depth
- dimension. Happens with strip mining where several scatter dimension
- are interdependend. */
- /* Length of [T]. */
- nt = pbb_nb_scattering_transform (pbb) + pbb_nb_local_vars (pbb);
- for (i = 0; i < nt; i++)
- if (i != time_depth)
- {
- isl_map *temp = isl_map_equate (isl_map_copy (map),
- isl_dim_out, i,
- isl_dim_out, offset + i);
- if (isl_map_is_empty (temp))
- isl_map_free (temp);
- else
- {
- isl_map_free (map);
- map = temp;
- }
- }
-
- /* Now maximize the expression L' - L. */
- set = isl_map_range (map);
- dc = isl_set_get_space (set);
- aff = isl_aff_zero_on_domain (isl_local_space_from_space (dc));
- aff = isl_aff_set_coefficient_si (aff, isl_dim_in, offset - 1, -1);
- aff = isl_aff_set_coefficient_si (aff, isl_dim_in, offset + offset - 1, 1);
- islstride = isl_set_max_val (set, aff);
- isl_val_get_num_gmp (islstride, stride);
- isl_val_free (islstride);
- isl_aff_free (aff);
- isl_set_free (set);
-
- if (dump_file && (dump_flags & TDF_DETAILS))
- {
- gmp_fprintf (dump_file, "\nStride in BB_%d, DR_%d, depth %d: %Zd ",
- pbb_index (pbb), PDR_ID (pdr), (int) depth, stride);
- }
-}
-
-/* Sets STRIDES to the sum of all the strides of the data references
- accessed in LOOP at DEPTH. */
-
-static void
-memory_strides_in_loop_1 (lst_p loop, graphite_dim_t depth, mpz_t strides)
-{
- int i, j;
- lst_p l;
- poly_dr_p pdr;
- mpz_t s, n;
-
- mpz_init (s);
- mpz_init (n);
-
- FOR_EACH_VEC_ELT (LST_SEQ (loop), j, l)
- if (LST_LOOP_P (l))
- memory_strides_in_loop_1 (l, depth, strides);
- else
- FOR_EACH_VEC_ELT (PBB_DRS (LST_PBB (l)), i, pdr)
- {
- pdr_stride_in_loop (s, depth, pdr);
- mpz_set_si (n, PDR_NB_REFS (pdr));
- mpz_mul (s, s, n);
- mpz_add (strides, strides, s);
- }
-
- mpz_clear (s);
- mpz_clear (n);
-}
-
-/* Sets STRIDES to the sum of all the strides of the data references
- accessed in LOOP at DEPTH. */
-
-static void
-memory_strides_in_loop (lst_p loop, graphite_dim_t depth, mpz_t strides)
-{
- if (mpz_cmp_si (loop->memory_strides, -1) == 0)
- {
- mpz_set_si (strides, 0);
- memory_strides_in_loop_1 (loop, depth, strides);
- }
- else
- mpz_set (strides, loop->memory_strides);
-}
-
-/* Return true when the interchange of loops LOOP1 and LOOP2 is
- profitable.
-
- Example:
-
- | int a[100][100];
- |
- | int
- | foo (int N)
- | {
- | int j;
- | int i;
- |
- | for (i = 0; i < N; i++)
- | for (j = 0; j < N; j++)
- | a[j][2 * i] += 1;
- |
- | return a[N][12];
- | }
-
- The data access A[j][i] is described like this:
-
- | i j N a s0 s1 1
- | 0 0 0 1 0 0 -5 = 0
- | 0 -1 0 0 1 0 0 = 0
- |-2 0 0 0 0 1 0 = 0
- | 0 0 0 0 1 0 0 >= 0
- | 0 0 0 0 0 1 0 >= 0
- | 0 0 0 0 -1 0 100 >= 0
- | 0 0 0 0 0 -1 100 >= 0
-
- The linearized memory access L to A[100][100] is:
-
- | i j N a s0 s1 1
- | 0 0 0 0 100 1 0
-
- TODO: the shown format is not valid as it does not show the fact
- that the iteration domain "i j" is transformed using the scattering.
-
- Next, to measure the impact of iterating once in loop "i", we build
- a maximization problem: first, we add to DR accesses the dimensions
- k, s2, s3, L1 = 100 * s0 + s1, L2, and D1: this is the polyhedron P1.
- L1 and L2 are the linearized memory access functions.
-
- | i j N a s0 s1 k s2 s3 L1 L2 D1 1
- | 0 0 0 1 0 0 0 0 0 0 0 0 -5 = 0 alias = 5
- | 0 -1 0 0 1 0 0 0 0 0 0 0 0 = 0 s0 = j
- |-2 0 0 0 0 1 0 0 0 0 0 0 0 = 0 s1 = 2 * i
- | 0 0 0 0 1 0 0 0 0 0 0 0 0 >= 0
- | 0 0 0 0 0 1 0 0 0 0 0 0 0 >= 0
- | 0 0 0 0 -1 0 0 0 0 0 0 0 100 >= 0
- | 0 0 0 0 0 -1 0 0 0 0 0 0 100 >= 0
- | 0 0 0 0 100 1 0 0 0 -1 0 0 0 = 0 L1 = 100 * s0 + s1
-
- Then, we generate the polyhedron P2 by interchanging the dimensions
- (s0, s2), (s1, s3), (L1, L2), (k, i)
-
- | i j N a s0 s1 k s2 s3 L1 L2 D1 1
- | 0 0 0 1 0 0 0 0 0 0 0 0 -5 = 0 alias = 5
- | 0 -1 0 0 0 0 0 1 0 0 0 0 0 = 0 s2 = j
- | 0 0 0 0 0 0 -2 0 1 0 0 0 0 = 0 s3 = 2 * k
- | 0 0 0 0 0 0 0 1 0 0 0 0 0 >= 0
- | 0 0 0 0 0 0 0 0 1 0 0 0 0 >= 0
- | 0 0 0 0 0 0 0 -1 0 0 0 0 100 >= 0
- | 0 0 0 0 0 0 0 0 -1 0 0 0 100 >= 0
- | 0 0 0 0 0 0 0 100 1 0 -1 0 0 = 0 L2 = 100 * s2 + s3
-
- then we add to P2 the equality k = i + 1:
-
- |-1 0 0 0 0 0 1 0 0 0 0 0 -1 = 0 k = i + 1
-
- and finally we maximize the expression "D1 = max (P1 inter P2, L2 - L1)".
-
- Similarly, to determine the impact of one iteration on loop "j", we
- interchange (k, j), we add "k = j + 1", and we compute D2 the
- maximal value of the difference.
-
- Finally, the profitability test is D1 < D2: if in the outer loop
- the strides are smaller than in the inner loop, then it is
- profitable to interchange the loops at DEPTH1 and DEPTH2. */
-
-static bool
-lst_interchange_profitable_p (lst_p nest, int depth1, int depth2)
-{
- mpz_t d1, d2;
- bool res;
-
- gcc_assert (depth1 < depth2);
-
- mpz_init (d1);
- mpz_init (d2);
-
- memory_strides_in_loop (nest, depth1, d1);
- memory_strides_in_loop (nest, depth2, d2);
-
- res = mpz_cmp (d1, d2) < 0;
-
- mpz_clear (d1);
- mpz_clear (d2);
-
- return res;
-}
-
-/* Interchanges the loops at DEPTH1 and DEPTH2 of the original
- scattering and assigns the resulting polyhedron to the transformed
- scattering. */
-
-static void
-pbb_interchange_loop_depths (graphite_dim_t depth1, graphite_dim_t depth2,
- poly_bb_p pbb)
-{
- unsigned i;
- unsigned dim1 = psct_dynamic_dim (pbb, depth1);
- unsigned dim2 = psct_dynamic_dim (pbb, depth2);
- isl_space *d = isl_map_get_space (pbb->transformed);
- isl_space *d1 = isl_space_range (d);
- unsigned n = isl_space_dim (d1, isl_dim_out);
- isl_space *d2 = isl_space_add_dims (d1, isl_dim_in, n);
- isl_map *x = isl_map_universe (d2);
-
- x = isl_map_equate (x, isl_dim_in, dim1, isl_dim_out, dim2);
- x = isl_map_equate (x, isl_dim_in, dim2, isl_dim_out, dim1);
-
- for (i = 0; i < n; i++)
- if (i != dim1 && i != dim2)
- x = isl_map_equate (x, isl_dim_in, i, isl_dim_out, i);
-
- pbb->transformed = isl_map_apply_range (pbb->transformed, x);
-}
-
-/* Apply the interchange of loops at depths DEPTH1 and DEPTH2 to all
- the statements below LST. */
-
-static void
-lst_apply_interchange (lst_p lst, int depth1, int depth2)
-{
- if (!lst)
- return;
-
- if (LST_LOOP_P (lst))
- {
- int i;
- lst_p l;
-
- FOR_EACH_VEC_ELT (LST_SEQ (lst), i, l)
- lst_apply_interchange (l, depth1, depth2);
- }
- else
- pbb_interchange_loop_depths (depth1, depth2, LST_PBB (lst));
-}
-
-/* Return true when the nest starting at LOOP1 and ending on LOOP2 is
- perfect: i.e. there are no sequence of statements. */
-
-static bool
-lst_perfectly_nested_p (lst_p loop1, lst_p loop2)
-{
- if (loop1 == loop2)
- return true;
-
- if (!LST_LOOP_P (loop1))
- return false;
-
- return LST_SEQ (loop1).length () == 1
- && lst_perfectly_nested_p (LST_SEQ (loop1)[0], loop2);
-}
-
-/* Transform the loop nest between LOOP1 and LOOP2 into a perfect
- nest. To continue the naming tradition, this function is called
- after perfect_nestify. NEST is set to the perfectly nested loop
- that is created. BEFORE/AFTER are set to the loops distributed
- before/after the loop NEST. */
-
-static void
-lst_perfect_nestify (lst_p loop1, lst_p loop2, lst_p *before,
- lst_p *nest, lst_p *after)
-{
- poly_bb_p first, last;
-
- gcc_assert (loop1 && loop2
- && loop1 != loop2
- && LST_LOOP_P (loop1) && LST_LOOP_P (loop2));
-
- first = LST_PBB (lst_find_first_pbb (loop2));
- last = LST_PBB (lst_find_last_pbb (loop2));
-
- *before = copy_lst (loop1);
- *nest = copy_lst (loop1);
- *after = copy_lst (loop1);
-
- lst_remove_all_before_including_pbb (*before, first, false);
- lst_remove_all_before_including_pbb (*after, last, true);
-
- lst_remove_all_before_excluding_pbb (*nest, first, true);
- lst_remove_all_before_excluding_pbb (*nest, last, false);
-
- if (lst_empty_p (*before))
- {
- free_lst (*before);
- *before = NULL;
- }
- if (lst_empty_p (*after))
- {
- free_lst (*after);
- *after = NULL;
- }
- if (lst_empty_p (*nest))
- {
- free_lst (*nest);
- *nest = NULL;
- }
-}
-
-/* Try to interchange LOOP1 with LOOP2 for all the statements of the
- body of LOOP2. LOOP1 contains LOOP2. Return true if it did the
- interchange. */
-
-static bool
-lst_try_interchange_loops (scop_p scop, lst_p loop1, lst_p loop2)
-{
- int depth1 = lst_depth (loop1);
- int depth2 = lst_depth (loop2);
- lst_p transformed;
-
- lst_p before = NULL, nest = NULL, after = NULL;
-
- if (!lst_perfectly_nested_p (loop1, loop2))
- lst_perfect_nestify (loop1, loop2, &before, &nest, &after);
-
- if (!lst_interchange_profitable_p (loop2, depth1, depth2))
- return false;
-
- lst_apply_interchange (loop2, depth1, depth2);
-
- /* Sync the transformed LST information and the PBB scatterings
- before using the scatterings in the data dependence analysis. */
- if (before || nest || after)
- {
- transformed = lst_substitute_3 (SCOP_TRANSFORMED_SCHEDULE (scop), loop1,
- before, nest, after);
- lst_update_scattering (transformed);
- free_lst (transformed);
- }
-
- if (graphite_legal_transform (scop))
- {
- if (dump_file && (dump_flags & TDF_DETAILS))
- fprintf (dump_file,
- "Loops at depths %d and %d will be interchanged.\n",
- depth1, depth2);
-
- /* Transform the SCOP_TRANSFORMED_SCHEDULE of the SCOP. */
- lst_insert_in_sequence (before, loop1, true);
- lst_insert_in_sequence (after, loop1, false);
-
- if (nest)
- {
- lst_replace (loop1, nest);
- free_lst (loop1);
- }
-
- return true;
- }
-
- /* Undo the transform. */
- free_lst (before);
- free_lst (nest);
- free_lst (after);
- lst_apply_interchange (loop2, depth2, depth1);
- return false;
-}
-
-/* Selects the inner loop in LST_SEQ (INNER_FATHER) to be interchanged
- with the loop OUTER in LST_SEQ (OUTER_FATHER). */
-
-static bool
-lst_interchange_select_inner (scop_p scop, lst_p outer_father, int outer,
- lst_p inner_father)
-{
- int inner;
- lst_p loop1, loop2;
-
- gcc_assert (outer_father
- && LST_LOOP_P (outer_father)
- && LST_LOOP_P (LST_SEQ (outer_father)[outer])
- && inner_father
- && LST_LOOP_P (inner_father));
-
- loop1 = LST_SEQ (outer_father)[outer];
-
- FOR_EACH_VEC_ELT (LST_SEQ (inner_father), inner, loop2)
- if (LST_LOOP_P (loop2)
- && (lst_try_interchange_loops (scop, loop1, loop2)
- || lst_interchange_select_inner (scop, outer_father, outer, loop2)))
- return true;
-
- return false;
-}
-
-/* Interchanges all the loops of LOOP and the loops of its body that
- are considered profitable to interchange. Return the number of
- interchanged loops. OUTER is the index in LST_SEQ (LOOP) that
- points to the next outer loop to be considered for interchange. */
-
-static int
-lst_interchange_select_outer (scop_p scop, lst_p loop, int outer)
-{
- lst_p l;
- int res = 0;
- int i = 0;
- lst_p father;
-
- if (!loop || !LST_LOOP_P (loop))
- return 0;
-
- father = LST_LOOP_FATHER (loop);
- if (father)
- {
- while (lst_interchange_select_inner (scop, father, outer, loop))
- {
- res++;
- loop = LST_SEQ (father)[outer];
- }
- }
-
- if (LST_LOOP_P (loop))
- FOR_EACH_VEC_ELT (LST_SEQ (loop), i, l)
- if (LST_LOOP_P (l))
- res += lst_interchange_select_outer (scop, l, i);
-
- return res;
-}
-
-/* Interchanges all the loop depths that are considered profitable for
- SCOP. Return the number of interchanged loops. */
-
-int
-scop_do_interchange (scop_p scop)
-{
- int res = lst_interchange_select_outer
- (scop, SCOP_TRANSFORMED_SCHEDULE (scop), 0);
-
- lst_update_scattering (SCOP_TRANSFORMED_SCHEDULE (scop));
-
- return res;
-}
-
-
-#endif /* HAVE_isl */
-
#include "tree-data-ref.h"
#include "graphite-poly.h"
#include "params.h"
+#include "dumpfile.h"
static isl_union_set *
scop_get_domains (scop_p scop ATTRIBUTE_UNUSED)
/* It does not make any sense to tile a band with just one dimension. */
if (*Dimensions == 1)
- return PartialSchedule;
+ {
+ if (dump_file && dump_flags)
+ fprintf (dump_file, "not tiled\n");
+ return PartialSchedule;
+ }
+
+ if (dump_file && dump_flags)
+ fprintf (dump_file, "tiled by %d\n",
+ PARAM_VALUE (PARAM_LOOP_BLOCK_TILE_SIZE));
ctx = isl_union_map_get_ctx (PartialSchedule);
Space = isl_union_map_get_space (PartialSchedule);
- TileMap = getTileMap (ctx, *Dimensions, 32);
+ TileMap = getTileMap (ctx, *Dimensions,
+ PARAM_VALUE (PARAM_LOOP_BLOCK_TILE_SIZE));
TileUMap = isl_union_map_from_map (isl_map_from_basic_map (TileMap));
TileUMap = isl_union_map_align_params (TileUMap, Space);
*Dimensions = 2 * *Dimensions;
gmp_fprintf (stderr, "%Zd", val);
}
-/* Return the maximal loop depth in SCOP. */
-
-int
-scop_max_loop_depth (scop_p scop)
-{
- int i;
- poly_bb_p pbb;
- int max_nb_loops = 0;
-
- FOR_EACH_VEC_ELT (SCOP_BBS (scop), i, pbb)
- {
- int nb_loops = pbb_dim_iter_domain (pbb);
- if (max_nb_loops < nb_loops)
- max_nb_loops = nb_loops;
- }
-
- return max_nb_loops;
-}
-
-/* Prints to FILE the scattering function of PBB, at some VERBOSITY
- level. */
-
-static void
-print_scattering_function_1 (FILE *file, poly_bb_p pbb, int verbosity)
-{
- graphite_dim_t i;
-
- if (verbosity > 0)
- {
- fprintf (file, "# scattering bb_%d (\n", pbb_index (pbb));
- fprintf (file, "#eq");
-
- for (i = 0; i < pbb_nb_scattering_transform (pbb); i++)
- fprintf (file, " s%d", (int) i);
-
- for (i = 0; i < pbb_nb_local_vars (pbb); i++)
- fprintf (file, " lv%d", (int) i);
-
- for (i = 0; i < pbb_dim_iter_domain (pbb); i++)
- fprintf (file, " i%d", (int) i);
-
- for (i = 0; i < pbb_nb_params (pbb); i++)
- fprintf (file, " p%d", (int) i);
-
- fprintf (file, " cst\n");
- }
-
- fprintf (file, "isl\n");
- print_isl_map (file, pbb->transformed ? pbb->transformed : pbb->schedule);
-
- if (verbosity > 0)
- fprintf (file, "#)\n");
-}
-
-/* Prints to FILE the scattering function of PBB, at some VERBOSITY
- level. */
-
-void
-print_scattering_function (FILE *file, poly_bb_p pbb, int verbosity)
-{
- if (!PBB_TRANSFORMED (pbb))
- return;
-
- if (pbb->schedule || pbb->transformed)
- {
- if (verbosity > 0)
- fprintf (file, "# Scattering function is provided\n");
-
- fprintf (file, "1\n");
- }
- else
- {
- if (verbosity > 0)
- fprintf (file, "# Scattering function is not provided\n");
-
- fprintf (file, "0\n");
- return;
- }
-
- print_scattering_function_1 (file, pbb, verbosity);
-
- if (verbosity > 0)
- fprintf (file, "# Scattering names are not provided\n");
-
- fprintf (file, "0\n");
-
-}
-
/* Prints to FILE the iteration domain of PBB, at some VERBOSITY
level. */
print_pbb_domain (file, pbb, verbosity);
}
-/* Prints to FILE the scattering functions of every PBB of SCOP. */
-
-void
-print_scattering_functions (FILE *file, scop_p scop, int verbosity)
-{
- int i;
- poly_bb_p pbb;
-
- FOR_EACH_VEC_ELT (SCOP_BBS (scop), i, pbb)
- print_scattering_function (file, pbb, verbosity);
-}
-
/* Prints to FILE the iteration domains of every PBB of SCOP, at some
VERBOSITY level. */
print_iteration_domain (file, pbb, verbosity);
}
-/* Prints to STDERR the scattering function of PBB, at some VERBOSITY
- level. */
-
-DEBUG_FUNCTION void
-debug_scattering_function (poly_bb_p pbb, int verbosity)
-{
- print_scattering_function (stderr, pbb, verbosity);
-}
-
/* Prints to STDERR the iteration domain of PBB, at some VERBOSITY
level. */
print_iteration_domain (stderr, pbb, verbosity);
}
-/* Prints to STDERR the scattering functions of every PBB of SCOP, at
- some VERBOSITY level. */
-
-DEBUG_FUNCTION void
-debug_scattering_functions (scop_p scop, int verbosity)
-{
- print_scattering_functions (stderr, scop, verbosity);
-}
-
/* Prints to STDERR the iteration domains of every PBB of SCOP, at
some VERBOSITY level. */
/* Generate code even if we did not apply any real transformation.
This also allows to check the performance for the identity
- transformation: GIMPLE -> GRAPHITE -> GIMPLE
- Keep in mind that CLooG optimizes in control, so the loop structure
- may change, even if we only use -fgraphite-identity. */
+ transformation: GIMPLE -> GRAPHITE -> GIMPLE. */
if (flag_graphite_identity)
transform_done = true;
if (flag_loop_parallelize_all)
transform_done = true;
- if (flag_loop_block)
- transform_done |= scop_do_block (scop);
- else
- {
- if (flag_loop_strip_mine)
- transform_done |= scop_do_strip_mine (scop, 0);
-
- if (flag_loop_interchange)
- transform_done |= scop_do_interchange (scop);
- }
-
- /* This pass needs to be run at the final stage, as it does not
- update the lst. */
if (flag_loop_optimize_isl)
transform_done |= optimize_isl (scop);
pbb->saved = NULL;
PBB_SCOP (pbb) = scop;
pbb_set_black_box (pbb, black_box);
- PBB_TRANSFORMED (pbb) = NULL;
- PBB_SAVED (pbb) = NULL;
- PBB_ORIGINAL (pbb) = NULL;
PBB_DRS (pbb).create (3);
PBB_IS_REDUCTION (pbb) = false;
GBB_PBB ((gimple_bb_p) black_box) = pbb;
XDELETE (pbb);
}
-static void
-print_pdr_access_layout (FILE *file, poly_bb_p pbb, poly_dr_p pdr)
-{
- graphite_dim_t i;
-
- fprintf (file, "# eq");
-
- fprintf (file, " alias");
-
- for (i = 0; i < PDR_NB_SUBSCRIPTS (pdr); i++)
- fprintf (file, " sub%d", (int) i);
-
- for (i = 0; i < pbb_dim_iter_domain (pbb); i++)
- fprintf (file, " i%d", (int) i);
-
- for (i = 0; i < pbb_nb_params (pbb); i++)
- fprintf (file, " p%d", (int) i);
-
- fprintf (file, " cst\n");
-}
-
/* Prints to FILE the polyhedral data reference PDR, at some VERBOSITY
level. */
if (verbosity > 0)
{
fprintf (file, "# data accesses (\n");
- print_pdr_access_layout (file, PDR_PBB (pdr), pdr);
+ print_isl_map (file, pdr->accesses);
+ print_isl_set (file, pdr->subscript_sizes);
+ fprintf (file, "#)\n");
}
-
- /* XXX isl dump accesses/subscripts */
-
- if (verbosity > 0)
- fprintf (file, "#)\n");
-
if (verbosity > 1)
fprintf (file, "#)\n");
}
scop->may_waw_no_source = NULL;
scop_set_region (scop, region);
SCOP_BBS (scop).create (3);
- SCOP_ORIGINAL_SCHEDULE (scop) = NULL;
- SCOP_TRANSFORMED_SCHEDULE (scop) = NULL;
- SCOP_SAVED_SCHEDULE (scop) = NULL;
POLY_SCOP_P (scop) = false;
return scop;
isl_union_map_free (scop->may_waw);
isl_union_map_free (scop->must_waw_no_source);
isl_union_map_free (scop->may_waw_no_source);
- free_lst (SCOP_ORIGINAL_SCHEDULE (scop));
- free_lst (SCOP_TRANSFORMED_SCHEDULE (scop));
- free_lst (SCOP_SAVED_SCHEDULE (scop));
XDELETE (scop);
}
-/* Print to FILE the domain of PBB in OpenScop format, at some VERBOSITY
- level. */
-
-static void
-openscop_print_pbb_domain (FILE *file, poly_bb_p pbb, int verbosity)
-{
- graphite_dim_t i;
- gimple_bb_p gbb = PBB_BLACK_BOX (pbb);
-
- if (!pbb->domain)
- return;
-
- if (verbosity > 0)
- {
- fprintf (file, "\n# Iteration domain of bb_%d (\n", GBB_BB (gbb)->index);
- fprintf (file, "#eq");
-
- for (i = 0; i < pbb_dim_iter_domain (pbb); i++)
- fprintf (file, " i%d", (int) i);
-
- for (i = 0; i < pbb_nb_params (pbb); i++)
- fprintf (file, " p%d", (int) i);
-
- fprintf (file, " cst\n");
- }
-
- fprintf (file, "XXX isl\n");
-
- if (verbosity > 0)
- fprintf (file, "#)\n");
-}
-
/* Print to FILE the domain of PBB, at some VERBOSITY level. */
void
dump_gbb_cases (file, PBB_BLACK_BOX (pbb));
}
- openscop_print_pbb_domain (file, pbb, verbosity);
- print_scattering_function (file, pbb, verbosity);
+ print_pbb_domain (file, pbb, verbosity);
print_pdrs (file, pbb, verbosity);
print_pbb_body (file, pbb, verbosity, false);
fprintf (file, "#)\n");
}
-/* Print to FILE the context of SCoP in OpenScop format, at some VERBOSITY
- level. */
-
-static void
-openscop_print_scop_context (FILE *file, scop_p scop, int verbosity)
-{
- graphite_dim_t i;
-
- if (verbosity > 0)
- {
- fprintf (file, "# Context (\n");
- fprintf (file, "#eq");
-
- for (i = 0; i < scop_nb_params (scop); i++)
- fprintf (file, " p%d", (int) i);
-
- fprintf (file, " cst\n");
- }
-
- if (scop->context)
- /* XXX isl print context */
- fprintf (file, "XXX isl\n");
- else
- fprintf (file, "0 %d 0 0 0 %d\n", (int) scop_nb_params (scop) + 2,
- (int) scop_nb_params (scop));
-
- if (verbosity > 0)
- fprintf (file, "# )\n");
-}
-
/* Print to FILE the context of SCoP, at some VERBOSITY level. */
void
print_scop_context (FILE *file, scop_p scop, int verbosity)
{
- graphite_dim_t i;
-
if (verbosity > 0)
- {
- fprintf (file, "# Context (\n");
- fprintf (file, "#eq");
-
- for (i = 0; i < scop_nb_params (scop); i++)
- fprintf (file, " p%d", (int) i);
-
- fprintf (file, " cst\n");
- }
+ fprintf (file, "# Context (\n");
if (scop->context)
print_isl_set (file, scop->context);
- else
- fprintf (file, "no isl context %d\n", (int) scop_nb_params (scop) + 2);
if (verbosity > 0)
fprintf (file, "# )\n");
fprintf (file, "SCoP 1\n#(\n");
fprintf (file, "# Language\nGimple\n");
- openscop_print_scop_context (file, scop, verbosity);
+ print_scop_context (file, scop, verbosity);
print_scop_params (file, scop, verbosity);
if (verbosity > 0)
FOR_EACH_VEC_ELT (SCOP_BBS (scop), i, pbb)
print_pbb (file, pbb, verbosity);
- if (verbosity > 1)
- {
- fprintf (file, "# original_lst (\n");
- print_lst (file, SCOP_ORIGINAL_SCHEDULE (scop), 0);
- fprintf (file, "\n#)\n");
-
- fprintf (file, "# transformed_lst (\n");
- print_lst (file, SCOP_TRANSFORMED_SCHEDULE (scop), 0);
- fprintf (file, "\n#)\n");
- }
-
fprintf (file, "#)\n");
}
{
isl_printer *p = isl_printer_to_file (the_isl_ctx, f);
p = isl_printer_print_set (p, set);
+ p = isl_printer_print_str (p, "\n");
isl_printer_free (p);
}
{
isl_printer *p = isl_printer_to_file (the_isl_ctx, f);
p = isl_printer_print_map (p, map);
+ p = isl_printer_print_str (p, "\n");
isl_printer_free (p);
}
{
isl_printer *p = isl_printer_to_file (the_isl_ctx, f);
p = isl_printer_print_aff (p, aff);
+ p = isl_printer_print_str (p, "\n");
isl_printer_free (p);
}
{
isl_printer *p = isl_printer_to_file (the_isl_ctx, f);
p = isl_printer_print_constraint (p, c);
+ p = isl_printer_print_str (p, "\n");
isl_printer_free (p);
}
isl_set_free (transdomain);
}
-/* Translates LOOP to LST. */
-
-static lst_p
-loop_to_lst (loop_p loop, vec<poly_bb_p> bbs, int *i)
-{
- poly_bb_p pbb;
- vec<lst_p> seq;
- seq.create (5);
-
- for (; bbs.iterate (*i, &pbb); (*i)++)
- {
- lst_p stmt;
- basic_block bb = GBB_BB (PBB_BLACK_BOX (pbb));
-
- if (bb->loop_father == loop)
- stmt = new_lst_stmt (pbb);
- else if (flow_bb_inside_loop_p (loop, bb))
- {
- loop_p next = loop->inner;
-
- while (next && !flow_bb_inside_loop_p (next, bb))
- next = next->next;
-
- stmt = loop_to_lst (next, bbs, i);
- }
- else
- {
- (*i)--;
- return new_lst_loop (seq);
- }
-
- seq.safe_push (stmt);
- }
-
- return new_lst_loop (seq);
-}
-
-/* Reads the original scattering of the SCOP and returns an LST
- representing it. */
-
-void
-scop_to_lst (scop_p scop)
-{
- lst_p res;
- int i, n = SCOP_BBS (scop).length ();
- vec<lst_p> seq;
- seq.create (5);
- sese region = SCOP_REGION (scop);
-
- for (i = 0; i < n; i++)
- {
- poly_bb_p pbb = SCOP_BBS (scop)[i];
- loop_p loop = outermost_loop_in_sese (region, GBB_BB (PBB_BLACK_BOX (pbb)));
-
- if (loop_in_sese_p (loop, region))
- res = loop_to_lst (loop, SCOP_BBS (scop), &i);
- else
- res = new_lst_stmt (pbb);
-
- seq.safe_push (res);
- }
-
- res = new_lst_loop (seq);
- SCOP_ORIGINAL_SCHEDULE (scop) = res;
- SCOP_TRANSFORMED_SCHEDULE (scop) = copy_lst (res);
-}
-
-/* Print to FILE on a new line COLUMN white spaces. */
-
-static void
-lst_indent_to (FILE *file, int column)
-{
- int i;
-
- if (column > 0)
- fprintf (file, "\n#");
-
- for (i = 0; i < column; i++)
- fprintf (file, " ");
-}
-
-/* Print LST to FILE with INDENT spaces of indentation. */
-
-void
-print_lst (FILE *file, lst_p lst, int indent)
-{
- if (!lst)
- return;
-
- lst_indent_to (file, indent);
-
- if (LST_LOOP_P (lst))
- {
- int i;
- lst_p l;
-
- if (LST_LOOP_FATHER (lst))
- fprintf (file, "%d (loop", lst_dewey_number (lst));
- else
- fprintf (file, "#(root");
-
- FOR_EACH_VEC_ELT (LST_SEQ (lst), i, l)
- print_lst (file, l, indent + 2);
-
- fprintf (file, ")");
- }
- else
- fprintf (file, "%d stmt_%d", lst_dewey_number (lst), pbb_index (LST_PBB (lst)));
-}
-
-/* Print LST to STDERR. */
-
-DEBUG_FUNCTION void
-debug_lst (lst_p lst)
-{
- print_lst (stderr, lst, 0);
-}
-
-/* Pretty print to FILE the loop statement tree LST in DOT format. */
-
-static void
-dot_lst_1 (FILE *file, lst_p lst)
-{
- if (!lst)
- return;
-
- if (LST_LOOP_P (lst))
- {
- int i;
- lst_p l;
-
- if (!LST_LOOP_FATHER (lst))
- fprintf (file, "L -> L_%d_%d\n",
- lst_depth (lst),
- lst_dewey_number (lst));
- else
- fprintf (file, "L_%d_%d -> L_%d_%d\n",
- lst_depth (LST_LOOP_FATHER (lst)),
- lst_dewey_number (LST_LOOP_FATHER (lst)),
- lst_depth (lst),
- lst_dewey_number (lst));
-
- FOR_EACH_VEC_ELT (LST_SEQ (lst), i, l)
- dot_lst_1 (file, l);
- }
-
- else
- fprintf (file, "L_%d_%d -> S_%d\n",
- lst_depth (LST_LOOP_FATHER (lst)),
- lst_dewey_number (LST_LOOP_FATHER (lst)),
- pbb_index (LST_PBB (lst)));
-
-}
-
-/* Display the LST using dotty. */
-
-DEBUG_FUNCTION void
-dot_lst (lst_p lst)
-{
- /* When debugging, enable the following code. This cannot be used
- in production compilers because it calls "system". */
-#if 0
- FILE *stream = fopen ("/tmp/lst.dot", "w");
- gcc_assert (stream);
-
- fputs ("digraph all {\n", stream);
- dot_lst_1 (stream, lst);
- fputs ("}\n\n", stream);
- fclose (stream);
-
- system ("dotty /tmp/lst.dot &");
-#else
- fputs ("digraph all {\n", stderr);
- dot_lst_1 (stderr, lst);
- fputs ("}\n\n", stderr);
-
-#endif
-}
-
-/* Reverse the loop around PBB at level DEPTH. */
-
-isl_map *
-reverse_loop_at_level (poly_bb_p pbb, int depth)
-{
- unsigned i, depth_dim = psct_dynamic_dim (pbb, depth);
- isl_space *d = isl_map_get_space (pbb->transformed);
- isl_space *d1 = isl_space_range (d);
- unsigned n = isl_space_dim (d1, isl_dim_out);
- isl_space *d2 = isl_space_add_dims (d1, isl_dim_in, n);
- isl_map *x = isl_map_universe (isl_space_copy (d2));
- isl_constraint *c = isl_equality_alloc (isl_local_space_from_space (d2));
-
- for (i = 0; i < n; i++)
- if (i != depth_dim)
- x = isl_map_equate (x, isl_dim_in, i, isl_dim_out, i);
-
- c = isl_constraint_set_coefficient_si (c, isl_dim_in, depth_dim, 1);
- c = isl_constraint_set_coefficient_si (c, isl_dim_out, depth_dim, 1);
- x = isl_map_add_constraint (x, c);
- return x;
-}
-
-/* Reverse the loop at level DEPTH for all the PBBS. */
-
-isl_union_map *
-reverse_loop_for_pbbs (scop_p scop, vec<poly_bb_p> pbbs, int depth)
-{
- poly_bb_p pbb;
- int i;
- isl_space *space = isl_space_from_domain (isl_set_get_space (scop->context));
- isl_union_map *res = isl_union_map_empty (space);
-
- for (i = 0; pbbs.iterate (i, &pbb); i++)
- res = isl_union_map_add_map (res, reverse_loop_at_level (pbb, depth));
-
- return res;
-}
-
-
#endif /* HAVE_isl */
typedef unsigned graphite_dim_t;
-static inline graphite_dim_t pbb_dim_iter_domain (const struct poly_bb *);
-static inline graphite_dim_t pbb_nb_params (const struct poly_bb *);
static inline graphite_dim_t scop_nb_params (scop_p);
/* A data reference can write or read some memory or we
void free_poly_dr (poly_dr_p);
void debug_pdr (poly_dr_p, int);
void print_pdr (FILE *, poly_dr_p, int);
-static inline scop_p pdr_scop (poly_dr_p pdr);
-
-/* The dimension of the iteration domain of the scop of PDR. */
-
-static inline graphite_dim_t
-pdr_dim_iter_domain (poly_dr_p pdr)
-{
- return pbb_dim_iter_domain (PDR_PBB (pdr));
-}
-
-/* The number of parameters of the scop of PDR. */
-
-static inline graphite_dim_t
-pdr_nb_params (poly_dr_p pdr)
-{
- return scop_nb_params (pdr_scop (pdr));
-}
-
-/* The dimension of the alias set in PDR. */
-
-static inline graphite_dim_t
-pdr_alias_set_dim (poly_dr_p pdr)
-{
- poly_bb_p pbb = PDR_PBB (pdr);
-
- return pbb_dim_iter_domain (pbb) + pbb_nb_params (pbb);
-}
-
-/* The dimension in PDR containing subscript S. */
-
-static inline graphite_dim_t
-pdr_subscript_dim (poly_dr_p pdr, graphite_dim_t s)
-{
- poly_bb_p pbb = PDR_PBB (pdr);
-
- return pbb_dim_iter_domain (pbb) + pbb_nb_params (pbb) + 1 + s;
-}
-
-/* The dimension in PDR containing the loop iterator ITER. */
-
-static inline graphite_dim_t
-pdr_iterator_dim (poly_dr_p pdr ATTRIBUTE_UNUSED, graphite_dim_t iter)
-{
- return iter;
-}
-
-/* The dimension in PDR containing parameter PARAM. */
-
-static inline graphite_dim_t
-pdr_parameter_dim (poly_dr_p pdr, graphite_dim_t param)
-{
- poly_bb_p pbb = PDR_PBB (pdr);
-
- return pbb_dim_iter_domain (pbb) + param;
-}
-
-/* Returns true when PDR is a "read". */
static inline bool
pdr_read_p (poly_dr_p pdr)
return PDR_TYPE (pdr) == PDR_MAY_WRITE;
}
-/* Return true when PDR1 and PDR2 are similar data accesses: they have
- the same base array, and the same access functions. */
-
-static inline bool
-same_pdr_p (poly_dr_p pdr1, poly_dr_p pdr2)
-{
- return PDR_NB_SUBSCRIPTS (pdr1) == PDR_NB_SUBSCRIPTS (pdr2)
- && PDR_BASE_OBJECT_SET (pdr1) == PDR_BASE_OBJECT_SET (pdr2);
-}
-
-typedef struct poly_scattering *poly_scattering_p;
-
-struct poly_scattering
-{
- /* The number of local variables. */
- int nb_local_variables;
-
- /* The number of scattering dimensions. */
- int nb_scattering;
-};
-
/* POLY_BB represents a blackbox in the polyhedral model. */
struct poly_bb
vec<poly_dr_p> drs;
/* The original scattering. */
- poly_scattering_p _original;
isl_map *schedule;
/* The transformed scattering. */
- poly_scattering_p _transformed;
isl_map *transformed;
/* A copy of the transformed scattering. */
- poly_scattering_p _saved;
isl_map *saved;
/* True when this PBB contains only a reduction statement. */
#define PBB_BLACK_BOX(PBB) ((gimple_bb_p) PBB->black_box)
#define PBB_SCOP(PBB) (PBB->scop)
-#define PBB_DOMAIN(PBB) (NULL)
#define PBB_DRS(PBB) (PBB->drs)
-#define PBB_ORIGINAL(PBB) (PBB->_original)
-#define PBB_ORIGINAL_SCATTERING(PBB) (NULL)
-#define PBB_TRANSFORMED(PBB) (PBB->_transformed)
-#define PBB_TRANSFORMED_SCATTERING(PBB) (NULL)
-#define PBB_SAVED(PBB) (PBB->_saved)
-/* XXX isl if we ever need local vars in the scatter, we can't use the
- out dimension of transformed to count the scatterting transform dimension.
- */
-#define PBB_NB_LOCAL_VARIABLES(PBB) (0)
-#define PBB_NB_SCATTERING_TRANSFORM(PBB) (isl_map_n_out (PBB->transformed))
#define PBB_IS_REDUCTION(PBB) (PBB->is_reduction)
extern poly_bb_p new_poly_bb (scop_p, void *);
extern void free_poly_bb (poly_bb_p);
extern void debug_loop_vec (poly_bb_p);
-extern void schedule_to_scattering (poly_bb_p, int);
extern void print_pbb_domain (FILE *, poly_bb_p, int);
extern void print_pbb (FILE *, poly_bb_p, int);
extern void print_scop_context (FILE *, scop_p, int);
extern void pbb_number_of_iterations_at_time (poly_bb_p, graphite_dim_t, mpz_t);
extern void debug_gmp_value (mpz_t);
-/* Return the number of write data references in PBB. */
-
-static inline int
-number_of_write_pdrs (poly_bb_p pbb)
-{
- int res = 0;
- int i;
- poly_dr_p pdr;
-
- for (i = 0; PBB_DRS (pbb).iterate (i, &pdr); i++)
- if (PDR_TYPE (pdr) == PDR_WRITE)
- res++;
-
- return res;
-}
-
/* Returns a gimple_bb from BB. */
static inline gimple_bb_p
return GBB_BB (PBB_BLACK_BOX (pbb));
}
-/* The index of the PBB. */
-
static inline int
pbb_index (poly_bb_p pbb)
{
pbb->black_box = black_box;
}
-/* The number of loops around PBB: the dimension of the iteration
- domain. */
-
-static inline graphite_dim_t
-pbb_dim_iter_domain (const struct poly_bb *pbb)
-{
- return isl_set_dim (pbb->domain, isl_dim_set);
-}
-
-/* The number of params defined in PBB. */
-
-static inline graphite_dim_t
-pbb_nb_params (const struct poly_bb *pbb)
-{
- scop_p scop = PBB_SCOP (pbb);
-
- return scop_nb_params (scop);
-}
-
-/* The number of scattering dimensions in the SCATTERING polyhedron
- of a PBB for a given SCOP. */
-
-static inline graphite_dim_t
-pbb_nb_scattering_orig (const struct poly_bb *pbb)
-{
- return 2 * pbb_dim_iter_domain (pbb) + 1;
-}
-
-/* The number of scattering dimensions in PBB. */
-
-static inline graphite_dim_t
-pbb_nb_scattering_transform (const struct poly_bb *pbb)
-{
- return PBB_NB_SCATTERING_TRANSFORM (pbb);
-}
-
-/* The number of dynamic scattering dimensions in PBB. */
-
-static inline graphite_dim_t
-pbb_nb_dynamic_scattering_transform (const struct poly_bb *pbb)
-{
- /* This function requires the 2d + 1 scattering format to be
- invariant during all transformations. */
- gcc_assert (PBB_NB_SCATTERING_TRANSFORM (pbb) % 2);
- return PBB_NB_SCATTERING_TRANSFORM (pbb) / 2;
-}
-
-/* Returns the number of local variables used in the transformed
- scattering polyhedron of PBB. */
-
-static inline graphite_dim_t
-pbb_nb_local_vars (const struct poly_bb *pbb ATTRIBUTE_UNUSED)
-{
- /* For now we do not have any local variables, as we do not do strip
- mining for example. */
- return PBB_NB_LOCAL_VARIABLES (pbb);
-}
-
-/* The dimension in the domain of PBB containing the iterator ITER. */
-
-static inline graphite_dim_t
-pbb_iterator_dim (poly_bb_p pbb ATTRIBUTE_UNUSED, graphite_dim_t iter)
-{
- return iter;
-}
-
-/* The dimension in the domain of PBB containing the iterator ITER. */
-
-static inline graphite_dim_t
-pbb_parameter_dim (poly_bb_p pbb, graphite_dim_t param)
-{
- return param
- + pbb_dim_iter_domain (pbb);
-}
-
-/* The dimension in the original scattering polyhedron of PBB
- containing the scattering iterator SCATTER. */
-
-static inline graphite_dim_t
-psco_scattering_dim (poly_bb_p pbb ATTRIBUTE_UNUSED, graphite_dim_t scatter)
-{
- gcc_assert (scatter < pbb_nb_scattering_orig (pbb));
- return scatter;
-}
-
-/* The dimension in the transformed scattering polyhedron of PBB
- containing the scattering iterator SCATTER. */
-
-static inline graphite_dim_t
-psct_scattering_dim (poly_bb_p pbb ATTRIBUTE_UNUSED, graphite_dim_t scatter)
-{
- gcc_assert (scatter <= pbb_nb_scattering_transform (pbb));
- return scatter;
-}
-
-/* The dimension in the transformed scattering polyhedron of PBB of
- the local variable LV. */
-
-static inline graphite_dim_t
-psct_local_var_dim (poly_bb_p pbb, graphite_dim_t lv)
-{
- gcc_assert (lv <= pbb_nb_local_vars (pbb));
- return lv + pbb_nb_scattering_transform (pbb);
-}
-
-/* The dimension in the original scattering polyhedron of PBB
- containing the loop iterator ITER. */
-
-static inline graphite_dim_t
-psco_iterator_dim (poly_bb_p pbb, graphite_dim_t iter)
-{
- gcc_assert (iter < pbb_dim_iter_domain (pbb));
- return iter + pbb_nb_scattering_orig (pbb);
-}
-
-/* The dimension in the transformed scattering polyhedron of PBB
- containing the loop iterator ITER. */
-
-static inline graphite_dim_t
-psct_iterator_dim (poly_bb_p pbb, graphite_dim_t iter)
-{
- gcc_assert (iter < pbb_dim_iter_domain (pbb));
- return iter
- + pbb_nb_scattering_transform (pbb)
- + pbb_nb_local_vars (pbb);
-}
-
-/* The dimension in the original scattering polyhedron of PBB
- containing parameter PARAM. */
-
-static inline graphite_dim_t
-psco_parameter_dim (poly_bb_p pbb, graphite_dim_t param)
-{
- gcc_assert (param < pbb_nb_params (pbb));
- return param
- + pbb_nb_scattering_orig (pbb)
- + pbb_dim_iter_domain (pbb);
-}
-
-/* The dimension in the transformed scattering polyhedron of PBB
- containing parameter PARAM. */
-
-static inline graphite_dim_t
-psct_parameter_dim (poly_bb_p pbb, graphite_dim_t param)
-{
- gcc_assert (param < pbb_nb_params (pbb));
- return param
- + pbb_nb_scattering_transform (pbb)
- + pbb_nb_local_vars (pbb)
- + pbb_dim_iter_domain (pbb);
-}
-
-/* The scattering dimension of PBB corresponding to the dynamic level
- LEVEL. */
-
-static inline graphite_dim_t
-psct_dynamic_dim (poly_bb_p pbb, graphite_dim_t level)
-{
- graphite_dim_t result = 1 + 2 * level;
-
- gcc_assert (result < pbb_nb_scattering_transform (pbb));
- return result;
-}
-
-/* The scattering dimension of PBB corresponding to the static
- sequence of the loop level LEVEL. */
-
-static inline graphite_dim_t
-psct_static_dim (poly_bb_p pbb, graphite_dim_t level)
-{
- graphite_dim_t result = 2 * level;
-
- gcc_assert (result < pbb_nb_scattering_transform (pbb));
- return result;
-}
-
-/* Adds to the transformed scattering polyhedron of PBB a new local
- variable and returns its index. */
-
-static inline graphite_dim_t
-psct_add_local_variable (poly_bb_p pbb ATTRIBUTE_UNUSED)
-{
- gcc_unreachable ();
- return 0;
-}
-
-typedef struct lst *lst_p;
-
-/* Loops and Statements Tree. */
-struct lst {
-
- /* LOOP_P is true when an LST node is a loop. */
- bool loop_p;
-
- /* A pointer to the loop that contains this node. */
- lst_p loop_father;
-
- /* The sum of all the memory strides for an LST loop. */
- mpz_t memory_strides;
-
- /* Loop nodes contain a sequence SEQ of LST nodes, statements
- contain a pointer to their polyhedral representation PBB. */
- union {
- poly_bb_p pbb;
- vec<lst_p> seq;
- } node;
-};
-
-#define LST_LOOP_P(LST) ((LST)->loop_p)
-#define LST_LOOP_FATHER(LST) ((LST)->loop_father)
-#define LST_PBB(LST) ((LST)->node.pbb)
-#define LST_SEQ(LST) ((LST)->node.seq)
-#define LST_LOOP_MEMORY_STRIDES(LST) ((LST)->memory_strides)
-
-void scop_to_lst (scop_p);
-void print_lst (FILE *, lst_p, int);
-void debug_lst (lst_p);
-void dot_lst (lst_p);
-
-/* Creates a new LST loop with SEQ. */
-
-static inline lst_p
-new_lst_loop (vec<lst_p> seq)
-{
- lst_p lst = XNEW (struct lst);
- int i;
- lst_p l;
-
- LST_LOOP_P (lst) = true;
- LST_SEQ (lst) = seq;
- LST_LOOP_FATHER (lst) = NULL;
- mpz_init (LST_LOOP_MEMORY_STRIDES (lst));
- mpz_set_si (LST_LOOP_MEMORY_STRIDES (lst), -1);
-
- for (i = 0; seq.iterate (i, &l); i++)
- LST_LOOP_FATHER (l) = lst;
-
- return lst;
-}
-
-/* Creates a new LST statement with PBB. */
-
-static inline lst_p
-new_lst_stmt (poly_bb_p pbb)
-{
- lst_p lst = XNEW (struct lst);
-
- LST_LOOP_P (lst) = false;
- LST_PBB (lst) = pbb;
- LST_LOOP_FATHER (lst) = NULL;
- return lst;
-}
-
-/* Frees the memory used by LST. */
-
-static inline void
-free_lst (lst_p lst)
-{
- if (!lst)
- return;
-
- if (LST_LOOP_P (lst))
- {
- int i;
- lst_p l;
-
- for (i = 0; LST_SEQ (lst).iterate (i, &l); i++)
- free_lst (l);
-
- mpz_clear (LST_LOOP_MEMORY_STRIDES (lst));
- LST_SEQ (lst).release ();
- }
-
- free (lst);
-}
-
-/* Returns a copy of LST. */
-
-static inline lst_p
-copy_lst (lst_p lst)
-{
- if (!lst)
- return NULL;
-
- if (LST_LOOP_P (lst))
- {
- int i;
- lst_p l;
- vec<lst_p> seq;
- seq.create (5);
-
- for (i = 0; LST_SEQ (lst).iterate (i, &l); i++)
- seq.safe_push (copy_lst (l));
-
- return new_lst_loop (seq);
- }
-
- return new_lst_stmt (LST_PBB (lst));
-}
-
-/* Adds a new loop under the loop LST. */
-
-static inline void
-lst_add_loop_under_loop (lst_p lst)
-{
- vec<lst_p> seq;
- seq.create (1);
- lst_p l = new_lst_loop (LST_SEQ (lst));
-
- gcc_assert (LST_LOOP_P (lst));
-
- LST_LOOP_FATHER (l) = lst;
- seq.quick_push (l);
- LST_SEQ (lst) = seq;
-}
-
-/* Returns the loop depth of LST. */
-
-static inline int
-lst_depth (lst_p lst)
-{
- if (!lst)
- return -2;
-
- /* The depth of the outermost "fake" loop is -1. This outermost
- loop does not have a loop father and it is just a container, as
- in the loop representation of GCC. */
- if (!LST_LOOP_FATHER (lst))
- return -1;
-
- return lst_depth (LST_LOOP_FATHER (lst)) + 1;
-}
-
-/* Returns the Dewey number for LST. */
-
-static inline int
-lst_dewey_number (lst_p lst)
-{
- int i;
- lst_p l;
-
- if (!lst)
- return -1;
-
- if (!LST_LOOP_FATHER (lst))
- return 0;
-
- FOR_EACH_VEC_ELT (LST_SEQ (LST_LOOP_FATHER (lst)), i, l)
- if (l == lst)
- return i;
-
- return -1;
-}
-
-/* Returns the Dewey number of LST at depth DEPTH. */
-
-static inline int
-lst_dewey_number_at_depth (lst_p lst, int depth)
-{
- gcc_assert (lst && depth >= 0 && lst_depth (lst) <= depth);
-
- if (lst_depth (lst) == depth)
- return lst_dewey_number (lst);
-
- return lst_dewey_number_at_depth (LST_LOOP_FATHER (lst), depth);
-}
-
-/* Returns the predecessor of LST in the sequence of its loop father.
- Returns NULL if LST is the first statement in the sequence. */
-
-static inline lst_p
-lst_pred (lst_p lst)
-{
- int dewey;
- lst_p father;
-
- if (!lst || !LST_LOOP_FATHER (lst))
- return NULL;
-
- dewey = lst_dewey_number (lst);
- if (dewey == 0)
- return NULL;
-
- father = LST_LOOP_FATHER (lst);
- return LST_SEQ (father)[dewey - 1];
-}
-
-/* Returns the successor of LST in the sequence of its loop father.
- Returns NULL if there is none. */
-
-static inline lst_p
-lst_succ (lst_p lst)
-{
- int dewey;
- lst_p father;
-
- if (!lst || !LST_LOOP_FATHER (lst))
- return NULL;
-
- dewey = lst_dewey_number (lst);
- father = LST_LOOP_FATHER (lst);
-
- if (LST_SEQ (father).length () == (unsigned) dewey + 1)
- return NULL;
-
- return LST_SEQ (father)[dewey + 1];
-}
-
-
-/* Return the LST node corresponding to PBB. */
-
-static inline lst_p
-lst_find_pbb (lst_p lst, poly_bb_p pbb)
-{
- int i;
- lst_p l;
-
- if (!lst)
- return NULL;
-
- if (!LST_LOOP_P (lst))
- return (pbb == LST_PBB (lst)) ? lst : NULL;
-
- for (i = 0; LST_SEQ (lst).iterate (i, &l); i++)
- {
- lst_p res = lst_find_pbb (l, pbb);
- if (res)
- return res;
- }
-
- return NULL;
-}
-
-/* Return the LST node corresponding to the loop around STMT at depth
- LOOP_DEPTH. */
-
-static inline lst_p
-find_lst_loop (lst_p stmt, int loop_depth)
-{
- lst_p loop = LST_LOOP_FATHER (stmt);
-
- gcc_assert (loop_depth >= 0);
-
- while (loop_depth < lst_depth (loop))
- loop = LST_LOOP_FATHER (loop);
-
- return loop;
-}
-
-/* Return the first LST representing a PBB statement in LST. */
-
-static inline lst_p
-lst_find_first_pbb (lst_p lst)
-{
- int i;
- lst_p l;
-
- if (!lst)
- return NULL;
-
- if (!LST_LOOP_P (lst))
- return lst;
-
- for (i = 0; LST_SEQ (lst).iterate (i, &l); i++)
- {
- lst_p res = lst_find_first_pbb (l);
- if (res)
- return res;
- }
-
- return NULL;
-}
-
-/* Returns true when LST is a loop that does not contain
- statements. */
-
-static inline bool
-lst_empty_p (lst_p lst)
-{
- return !lst_find_first_pbb (lst);
-}
-
-/* Return the last LST representing a PBB statement in LST. */
-
-static inline lst_p
-lst_find_last_pbb (lst_p lst)
-{
- int i;
- lst_p l, res = NULL;
-
- if (!lst)
- return NULL;
-
- if (!LST_LOOP_P (lst))
- return lst;
-
- for (i = 0; LST_SEQ (lst).iterate (i, &l); i++)
- {
- lst_p last = lst_find_last_pbb (l);
-
- if (last)
- res = last;
- }
-
- gcc_assert (res);
- return res;
-}
-
-/* Returns true if LOOP contains LST, in other words, if LST is nested
- in LOOP. */
-
-static inline bool
-lst_contains_p (lst_p loop, lst_p lst)
-{
- if (!loop || !lst || !LST_LOOP_P (loop))
- return false;
-
- if (loop == lst)
- return true;
-
- return lst_contains_p (loop, LST_LOOP_FATHER (lst));
-}
-
-/* Returns true if LOOP contains PBB, in other words, if PBB is nested
- in LOOP. */
-
-static inline bool
-lst_contains_pbb (lst_p loop, poly_bb_p pbb)
-{
- return lst_find_pbb (loop, pbb) ? true : false;
-}
-
-/* Creates a loop nest of depth NB_LOOPS containing LST. */
-
-static inline lst_p
-lst_create_nest (int nb_loops, lst_p lst)
-{
- lst_p res, loop;
- vec<lst_p> seq;
-
- if (nb_loops == 0)
- return lst;
-
- seq.create (1);
- loop = lst_create_nest (nb_loops - 1, lst);
- seq.quick_push (loop);
- res = new_lst_loop (seq);
- LST_LOOP_FATHER (loop) = res;
-
- return res;
-}
-
-/* Removes LST from the sequence of statements of its loop father. */
-
-static inline void
-lst_remove_from_sequence (lst_p lst)
-{
- lst_p father = LST_LOOP_FATHER (lst);
- int dewey = lst_dewey_number (lst);
-
- gcc_assert (lst && father && dewey >= 0);
-
- LST_SEQ (father).ordered_remove (dewey);
- LST_LOOP_FATHER (lst) = NULL;
-}
-
-/* Removes the loop LST and inline its body in the father loop. */
-
-static inline void
-lst_remove_loop_and_inline_stmts_in_loop_father (lst_p lst)
-{
- lst_p l, father = LST_LOOP_FATHER (lst);
- int i, dewey = lst_dewey_number (lst);
-
- gcc_assert (lst && father && dewey >= 0);
-
- LST_SEQ (father).ordered_remove (dewey);
- LST_LOOP_FATHER (lst) = NULL;
-
- FOR_EACH_VEC_ELT (LST_SEQ (lst), i, l)
- {
- LST_SEQ (father).safe_insert (dewey + i, l);
- LST_LOOP_FATHER (l) = father;
- }
-}
-
-/* Sets NITER to the upper bound approximation of the number of
- iterations of loop LST. */
-
-static inline void
-lst_niter_for_loop (lst_p lst, mpz_t niter)
-{
- int depth = lst_depth (lst);
- poly_bb_p pbb = LST_PBB (lst_find_first_pbb (lst));
-
- gcc_assert (LST_LOOP_P (lst));
- pbb_number_of_iterations_at_time (pbb, psct_dynamic_dim (pbb, depth), niter);
-}
-
-/* Updates the scattering of PBB to be at the DEWEY number in the loop
- at depth LEVEL. */
-
-static inline void
-pbb_update_scattering (poly_bb_p pbb, graphite_dim_t level, int dewey)
-{
- graphite_dim_t sched = psct_static_dim (pbb, level);
- isl_space *d = isl_map_get_space (pbb->transformed);
- isl_space *d1 = isl_space_range (d);
- unsigned i, n = isl_space_dim (d1, isl_dim_out);
- isl_space *d2 = isl_space_add_dims (d1, isl_dim_in, n);
- isl_map *x = isl_map_universe (d2);
-
- x = isl_map_fix_si (x, isl_dim_out, sched, dewey);
-
- for (i = 0; i < n; i++)
- if (i != sched)
- x = isl_map_equate (x, isl_dim_in, i, isl_dim_out, i);
-
- pbb->transformed = isl_map_apply_range (pbb->transformed, x);
-}
-
-/* Updates the scattering of all the PBBs under LST to be at the DEWEY
- number in the loop at depth LEVEL. */
-
-static inline void
-lst_update_scattering_under (lst_p lst, int level, int dewey)
-{
- int i;
- lst_p l;
-
- gcc_assert (lst && level >= 0 && dewey >= 0);
-
- if (LST_LOOP_P (lst))
- for (i = 0; LST_SEQ (lst).iterate (i, &l); i++)
- lst_update_scattering_under (l, level, dewey);
- else
- pbb_update_scattering (LST_PBB (lst), level, dewey);
-}
-
-/* Updates the all the scattering levels of all the PBBs under
- LST. */
-
-static inline void
-lst_update_scattering (lst_p lst)
-{
- int i;
- lst_p l;
-
- if (!lst)
- return;
-
- if (LST_LOOP_FATHER (lst))
- {
- lst_p father = LST_LOOP_FATHER (lst);
- int dewey = lst_dewey_number (lst);
- int level = lst_depth (lst);
-
- gcc_assert (lst && father && dewey >= 0 && level >= 0);
-
- for (i = dewey; LST_SEQ (father).iterate (i, &l); i++)
- lst_update_scattering_under (l, level, i);
- }
-
- if (LST_LOOP_P (lst))
- for (i = 0; LST_SEQ (lst).iterate (i, &l); i++)
- lst_update_scattering (l);
-}
-
-/* Inserts LST1 before LST2 if BEFORE is true; inserts LST1 after LST2
- if BEFORE is false. */
-
-static inline void
-lst_insert_in_sequence (lst_p lst1, lst_p lst2, bool before)
-{
- lst_p father;
- int dewey;
-
- /* Do not insert empty loops. */
- if (!lst1 || lst_empty_p (lst1))
- return;
-
- father = LST_LOOP_FATHER (lst2);
- dewey = lst_dewey_number (lst2);
-
- gcc_assert (lst2 && father && dewey >= 0);
-
- LST_SEQ (father).safe_insert (before ? dewey : dewey + 1, lst1);
- LST_LOOP_FATHER (lst1) = father;
-}
-
-/* Replaces LST1 with LST2. */
-
-static inline void
-lst_replace (lst_p lst1, lst_p lst2)
-{
- lst_p father;
- int dewey;
-
- if (!lst2 || lst_empty_p (lst2))
- return;
-
- father = LST_LOOP_FATHER (lst1);
- dewey = lst_dewey_number (lst1);
- LST_LOOP_FATHER (lst2) = father;
- LST_SEQ (father)[dewey] = lst2;
-}
-
-/* Returns a copy of ROOT where LST has been replaced by a copy of the
- LSTs A B C in this sequence. */
-
-static inline lst_p
-lst_substitute_3 (lst_p root, lst_p lst, lst_p a, lst_p b, lst_p c)
-{
- int i;
- lst_p l;
- vec<lst_p> seq;
-
- if (!root)
- return NULL;
-
- gcc_assert (lst && root != lst);
-
- if (!LST_LOOP_P (root))
- return new_lst_stmt (LST_PBB (root));
-
- seq.create (5);
-
- for (i = 0; LST_SEQ (root).iterate (i, &l); i++)
- if (l != lst)
- seq.safe_push (lst_substitute_3 (l, lst, a, b, c));
- else
- {
- if (!lst_empty_p (a))
- seq.safe_push (copy_lst (a));
- if (!lst_empty_p (b))
- seq.safe_push (copy_lst (b));
- if (!lst_empty_p (c))
- seq.safe_push (copy_lst (c));
- }
-
- return new_lst_loop (seq);
-}
-
-/* Moves LST before LOOP if BEFORE is true, and after the LOOP if
- BEFORE is false. */
-
-static inline void
-lst_distribute_lst (lst_p loop, lst_p lst, bool before)
-{
- int loop_depth = lst_depth (loop);
- int depth = lst_depth (lst);
- int nb_loops = depth - loop_depth;
-
- gcc_assert (lst && loop && LST_LOOP_P (loop) && nb_loops > 0);
-
- lst_remove_from_sequence (lst);
- lst_insert_in_sequence (lst_create_nest (nb_loops, lst), loop, before);
-}
-
-/* Removes from LOOP all the statements before/after and including PBB
- if BEFORE is true/false. Returns the negation of BEFORE when the
- statement PBB has been found. */
-
-static inline bool
-lst_remove_all_before_including_pbb (lst_p loop, poly_bb_p pbb, bool before)
-{
- int i;
- lst_p l;
-
- if (!loop || !LST_LOOP_P (loop))
- return before;
-
- for (i = 0; LST_SEQ (loop).iterate (i, &l);)
- if (LST_LOOP_P (l))
- {
- before = lst_remove_all_before_including_pbb (l, pbb, before);
-
- if (LST_SEQ (l).length () == 0)
- {
- LST_SEQ (loop).ordered_remove (i);
- free_lst (l);
- }
- else
- i++;
- }
- else
- {
- if (before)
- {
- if (LST_PBB (l) == pbb)
- before = false;
-
- LST_SEQ (loop).ordered_remove (i);
- free_lst (l);
- }
- else if (LST_PBB (l) == pbb)
- {
- before = true;
- LST_SEQ (loop).ordered_remove (i);
- free_lst (l);
- }
- else
- i++;
- }
-
- return before;
-}
-
-/* Removes from LOOP all the statements before/after and excluding PBB
- if BEFORE is true/false; Returns the negation of BEFORE when the
- statement PBB has been found. */
-
-static inline bool
-lst_remove_all_before_excluding_pbb (lst_p loop, poly_bb_p pbb, bool before)
-{
- int i;
- lst_p l;
-
- if (!loop || !LST_LOOP_P (loop))
- return before;
-
- for (i = 0; LST_SEQ (loop).iterate (i, &l);)
- if (LST_LOOP_P (l))
- {
- before = lst_remove_all_before_excluding_pbb (l, pbb, before);
-
- if (LST_SEQ (l).length () == 0)
- {
- LST_SEQ (loop).ordered_remove (i);
- free_lst (l);
- continue;
- }
-
- i++;
- }
- else
- {
- if (before && LST_PBB (l) != pbb)
- {
- LST_SEQ (loop).ordered_remove (i);
- free_lst (l);
- continue;
- }
-
- i++;
-
- if (LST_PBB (l) == pbb)
- before = before ? false : true;
- }
-
- return before;
-}
-
/* A SCOP is a Static Control Part of the program, simple enough to be
represented in polyhedral form. */
struct scop
representation. */
vec<poly_bb_p> bbs;
- /* Original, transformed and saved schedules. */
- lst_p original_schedule, transformed_schedule, saved_schedule;
-
/* The context describes known restrictions concerning the parameters
and relations in between the parameters.
#define SCOP_BBS(S) (S->bbs)
#define SCOP_REGION(S) (S->region)
#define SCOP_CONTEXT(S) (NULL)
-#define SCOP_ORIGINAL_SCHEDULE(S) (S->original_schedule)
-#define SCOP_TRANSFORMED_SCHEDULE(S) (S->transformed_schedule)
-#define SCOP_SAVED_SCHEDULE(S) (S->saved_schedule)
#define POLY_SCOP_P(S) (S->poly_scop_p)
extern scop_p new_scop (sese);
extern void free_scops (vec<scop_p> );
extern void print_generated_program (FILE *, scop_p);
extern void debug_generated_program (scop_p);
-extern void print_scattering_function (FILE *, poly_bb_p, int);
-extern void print_scattering_functions (FILE *, scop_p, int);
-extern void debug_scattering_function (poly_bb_p, int);
-extern void debug_scattering_functions (scop_p, int);
-extern int scop_max_loop_depth (scop_p);
extern int unify_scattering_dimensions (scop_p);
extern bool apply_poly_transforms (scop_p);
extern bool graphite_legal_transform (scop_p);
scop->nb_params = nb_params;
}
-/* Allocates a new empty poly_scattering structure. */
-
-static inline poly_scattering_p
-poly_scattering_new (void)
-{
- poly_scattering_p res = XNEW (struct poly_scattering);
-
- res->nb_local_variables = 0;
- res->nb_scattering = 0;
- return res;
-}
-
-/* Free a poly_scattering structure. */
-
-static inline void
-poly_scattering_free (poly_scattering_p s)
-{
- free (s);
-}
-
-/* Copies S and return a new scattering. */
-
-static inline poly_scattering_p
-poly_scattering_copy (poly_scattering_p s)
-{
- poly_scattering_p res = poly_scattering_new ();
-
- res->nb_local_variables = s->nb_local_variables;
- res->nb_scattering = s->nb_scattering;
- return res;
-}
-
-/* Saves the transformed scattering of PBB. */
-
-static inline void
-store_scattering_pbb (poly_bb_p pbb)
-{
- isl_map_free (pbb->saved);
- pbb->saved = isl_map_copy (pbb->transformed);
-}
-
-/* Stores the SCOP_TRANSFORMED_SCHEDULE to SCOP_SAVED_SCHEDULE. */
-
-static inline void
-store_lst_schedule (scop_p scop)
-{
- if (SCOP_SAVED_SCHEDULE (scop))
- free_lst (SCOP_SAVED_SCHEDULE (scop));
-
- SCOP_SAVED_SCHEDULE (scop) = copy_lst (SCOP_TRANSFORMED_SCHEDULE (scop));
-}
-
-/* Restores the SCOP_TRANSFORMED_SCHEDULE from SCOP_SAVED_SCHEDULE. */
-
-static inline void
-restore_lst_schedule (scop_p scop)
-{
- if (SCOP_TRANSFORMED_SCHEDULE (scop))
- free_lst (SCOP_TRANSFORMED_SCHEDULE (scop));
-
- SCOP_TRANSFORMED_SCHEDULE (scop) = copy_lst (SCOP_SAVED_SCHEDULE (scop));
-}
-
-/* Saves the scattering for all the pbbs in the SCOP. */
-
-static inline void
-store_scattering (scop_p scop)
-{
- int i;
- poly_bb_p pbb;
-
- for (i = 0; SCOP_BBS (scop).iterate (i, &pbb); i++)
- store_scattering_pbb (pbb);
-
- store_lst_schedule (scop);
-}
-
-/* Restores the scattering of PBB. */
-
-static inline void
-restore_scattering_pbb (poly_bb_p pbb)
-{
- gcc_assert (pbb->saved);
-
- isl_map_free (pbb->transformed);
- pbb->transformed = isl_map_copy (pbb->saved);
-}
-
-/* Restores the scattering for all the pbbs in the SCOP. */
-
-static inline void
-restore_scattering (scop_p scop)
-{
- int i;
- poly_bb_p pbb;
-
- for (i = 0; SCOP_BBS (scop).iterate (i, &pbb); i++)
- restore_scattering_pbb (pbb);
-
- restore_lst_schedule (scop);
-}
-
bool graphite_legal_transform (scop_p);
-isl_map *reverse_loop_at_level (poly_bb_p, int);
-isl_union_map *reverse_loop_for_pbbs (scop_p, vec<poly_bb_p> , int);
__isl_give isl_union_map *extend_schedule (__isl_take isl_union_map *);
-
void
compute_deps (scop_p scop, vec<poly_bb_p> pbbs,
isl_union_map **must_raw,
/* Converts the STATIC_SCHEDULE of PBB into a scattering polyhedron.
We generate SCATTERING_DIMENSIONS scattering dimensions.
- CLooG 0.15.0 and previous versions require, that all
- scattering functions of one CloogProgram have the same number of
- scattering dimensions, therefore we allow to specify it. This
- should be removed in future versions of CLooG.
-
The scattering polyhedron consists of these dimensions: scattering,
loop_iterators, parameters.
Example:
| scattering_dimensions = 5
- | used_scattering_dimensions = 3
| nb_iterators = 1
| scop_nb_params = 2
|
static void
build_pbb_scattering_polyhedrons (isl_aff *static_sched,
- poly_bb_p pbb, int scattering_dimensions)
+ poly_bb_p pbb)
{
int i;
- int nb_iterators = pbb_dim_iter_domain (pbb);
- int used_scattering_dimensions = nb_iterators * 2 + 1;
isl_val *val;
isl_space *dc, *dm;
- gcc_assert (scattering_dimensions >= used_scattering_dimensions);
+ int scattering_dimensions = isl_set_dim (pbb->domain, isl_dim_set) * 2 + 1;
dc = isl_set_get_space (pbb->domain);
dm = isl_space_add_dims (isl_space_from_domain (dc),
{
gimple_bb_p gbb = PBB_BLACK_BOX (pbb);
int prefix;
- int nb_scat_dims = pbb_dim_iter_domain (pbb) * 2 + 1;
if (previous_gbb)
prefix = nb_common_loops (SCOP_REGION (scop), previous_gbb, gbb);
static_sched = isl_aff_add_coefficient_si (static_sched, isl_dim_in,
prefix, 1);
- build_pbb_scattering_polyhedrons (static_sched, pbb, nb_scat_dims);
+ build_pbb_scattering_polyhedrons (static_sched, pbb);
}
isl_aff_free (static_sched);
}
/* Can all ivs be represented by a signed integer?
- As CLooG might generate negative values in its expressions, signed loop ivs
+ As ISL might generate negative values in its expressions, signed loop ivs
are required in the backend. */
static bool
build_scop_bbs (scop);
- /* FIXME: This restriction is needed to avoid a problem in CLooG.
- Once CLooG is fixed, remove this guard. Anyways, it makes no
- sense to optimize a scop containing only PBBs that do not belong
+ /* Do not optimize a scop containing only PBBs that do not belong
to any loops. */
if (nb_pbbs_in_loops (scop) == 0)
return;
rewrite_cross_bb_scalar_deps_out_of_ssa (scop);
build_scop_drs (scop);
- scop_to_lst (scop);
build_scop_scattering (scop);
/* This SCoP has been translated to the polyhedral
An early description of this pass can be found in the GCC Summit'06
paper "GRAPHITE: Polyhedral Analyses and Optimizations for GCC".
The wiki page http://gcc.gnu.org/wiki/Graphite contains pointers to
- the related work.
-
- One important document to read is CLooG's internal manual:
- http://repo.or.cz/w/cloog-ppl.git?a=blob_plain;f=doc/cloog.texi;hb=HEAD
- that describes the data structure of loops used in this file, and
- the functions that are used for transforming the code. */
+ the related work. */
#include "config.h"
scop->ctx = ctx;
build_poly_scop (scop);
+ if (dump_file && dump_flags)
+ print_scop (dump_file, scop, 3);
+
if (POLY_SCOP_P (scop)
&& apply_poly_transforms (scop)
&& graphite_regenerate_ast_isl (scop))
{
/* Enable -fgraphite pass if any one of the graphite optimization flags
is turned on. */
- if (flag_loop_block
- || flag_loop_interchange
- || flag_loop_strip_mine
- || flag_graphite_identity
+ if (flag_graphite_identity
|| flag_loop_parallelize_all
|| flag_loop_optimize_isl)
flag_graphite = 1;
+2015-08-27 Sebastian Pop <s.pop@samsung.com>
+
+ * gcc.dg/graphite/graphite.exp (interchange_files, block_files):
+ Replaced by opt_files, compile with -floop-nest-optimize.
+ * g++.dg/graphite/graphite.exp: Same.
+ * gfortran.dg/graphite/graphite.exp: Same.
+ * gcc.dg/graphite/block-0.c: Adjust pattern.
+ * gcc.dg/graphite/block-1.c: Same.
+ * gcc.dg/graphite/block-3.c: Same.
+ * gcc.dg/graphite/block-4.c: Same.
+ * gcc.dg/graphite/block-5.c: Same.
+ * gcc.dg/graphite/block-6.c: Same.
+ * gcc.dg/graphite/block-7.c: Same.
+ * gcc.dg/graphite/block-8.c: Same.
+ * gcc.dg/graphite/block-pr47654.c: Same.
+ * gcc.dg/graphite/interchange-0.c: Same.
+ * gcc.dg/graphite/interchange-1.c: Same.
+ * gcc.dg/graphite/interchange-10.c: Same.
+ * gcc.dg/graphite/interchange-11.c: Same.
+ * gcc.dg/graphite/interchange-12.c: Same.
+ * gcc.dg/graphite/interchange-13.c: Same.
+ * gcc.dg/graphite/interchange-14.c: Same.
+ * gcc.dg/graphite/interchange-15.c: Same.
+ * gcc.dg/graphite/interchange-3.c: Same.
+ * gcc.dg/graphite/interchange-4.c: Same.
+ * gcc.dg/graphite/interchange-5.c: Same.
+ * gcc.dg/graphite/interchange-6.c: Same.
+ * gcc.dg/graphite/interchange-7.c: Same.
+ * gcc.dg/graphite/interchange-8.c: Same.
+ * gcc.dg/graphite/interchange-9.c: Same.
+ * gcc.dg/graphite/interchange-mvt.c: Same.
+ * gcc.dg/graphite/pr37485.c: Same.
+ * gcc.dg/graphite/uns-block-1.c: Same.
+ * gcc.dg/graphite/uns-interchange-12.c: Same.
+ * gcc.dg/graphite/uns-interchange-14.c: Same.
+ * gcc.dg/graphite/uns-interchange-15.c: Same.
+ * gcc.dg/graphite/uns-interchange-9.c: Same.
+ * gcc.dg/graphite/uns-interchange-mvt.c: Same.
+ * gfortran.dg/graphite/interchange-3.f90: Same.
+
2015-08-27 Uros Bizjak <ubizjak@gmail.com>
PR target/67317
set scop_files [lsort [glob -nocomplain $srcdir/$subdir/scop-*.C ] ]
set id_files [lsort [glob -nocomplain $srcdir/$subdir/id-*.C ] ]
set run_id_files [lsort [glob -nocomplain $srcdir/$subdir/run-id-*.C ] ]
-set interchange_files [lsort [glob -nocomplain $srcdir/$subdir/interchange-*.C ] ]
-set block_files [lsort [glob -nocomplain $srcdir/$subdir/block-*.C ] ]
+set opt_files [lsort [glob -nocomplain $srcdir/$subdir/interchange-*.C \
+ $srcdir/$subdir/block-*.C ] ]
# Tests to be compiled.
set dg-do-what-default compile
# Tests to be run.
set dg-do-what-default run
g++-dg-runtest $run_id_files "" "-O2 -fgraphite-identity"
-g++-dg-runtest $interchange_files "" "-O2 -floop-interchange -fno-loop-block -fno-loop-strip-mine -ffast-math -fdump-tree-graphite-all"
-g++-dg-runtest $block_files "" "-O2 -floop-block -fno-loop-strip-mine -fno-loop-interchange -ffast-math -fdump-tree-graphite-all"
+g++-dg-runtest $opt_files "" "-O2 -floop-nest-optimize -ffast-math -fdump-tree-graphite-all"
# The default action for the rest of the files is 'compile'.
set dg-do-what-default compile
foreach f $scop_files {lremove wait_to_run_files $f}
foreach f $id_files {lremove wait_to_run_files $f}
foreach f $run_id_files {lremove wait_to_run_files $f}
-foreach f $interchange_files {lremove wait_to_run_files $f}
-foreach f $block_files {lremove wait_to_run_files $f}
+foreach f $opt_files {lremove wait_to_run_files $f}
g++-dg-runtest $wait_to_run_files "" "-pedantic-errors"
# Clean up.
return 0;
}
-/* { dg-final { scan-tree-dump-not "will be loop blocked" "graphite" } } */
+/* { dg-final { scan-tree-dump-times "not tiled" 3 "graphite" } } */
return 0;
}
-/* { dg-final { scan-tree-dump-times "will be loop blocked" 3 "graphite" { xfail *-*-* } } } */
+/* { dg-final { scan-tree-dump-times "tiled by" 5 "graphite" } } */
{
int i, j, k;
- /* These loops contain too few iterations to be blocked by 64. */
+ /* These loops contain too few iterations to be blocked. */
for (i = 0; i < 24; i++)
for (j = 0; j < 24; j++)
for (k = 0; k < 24; k++)
return 0;
}
-/* { dg-final { scan-tree-dump-times "will be loop blocked" 1 "graphite" } } */
+/* { dg-final { scan-tree-dump-times "tiled by" 3 "graphite" } } */
{
int i, j, k;
- /* This should NOT be blocked: each loop iterates only 24 times. */
+ /* This should NOT be blocked. */
for (i = 0; i < 24; i++)
for (j = 0; j < 24; j++)
for (k = 0; k < 24; k++)
return 0;
}
-/* { dg-final { scan-tree-dump-times "will be loop blocked" 1 "graphite" { xfail *-*-* } } } */
+/* { dg-final { scan-tree-dump-times "tiled by" 7 "graphite" } } */
return 0;
}
-/* { dg-final { scan-tree-dump-times "will be loop blocked" 1 "graphite" } } */
+/* { dg-final { scan-tree-dump-times "tiled by" 3 "graphite" } } */
return 0;
}
-/* { dg-final { scan-tree-dump-times "will be loop blocked" 0 "graphite" } } */
+/* { dg-final { scan-tree-dump-times "tiled by" 3 "graphite" } } */
return 0;
}
-/* { dg-final { scan-tree-dump-times "will be loop blocked" 1 "graphite" } } */
+/* { dg-final { scan-tree-dump-times "tiled by" 6 "graphite" } } */
return 0;
}
-/* { dg-final { scan-tree-dump-times "will be loop blocked" 1 "graphite" } } */
+/* { dg-final { scan-tree-dump-times "tiled by" 7 "graphite" } } */
return 0;
}
-/* { dg-final { scan-tree-dump-not "will be loop blocked" "graphite" } } */
+/* { dg-final { scan-tree-dump-times "tiled by" 1 "graphite" } } */
set scop_files [lsort [glob -nocomplain $srcdir/$subdir/scop-*.c ] ]
set id_files [lsort [glob -nocomplain $srcdir/$subdir/id-*.c ] ]
set run_id_files [lsort [glob -nocomplain $srcdir/$subdir/run-id-*.c ] ]
-set interchange_files [lsort [glob -nocomplain $srcdir/$subdir/interchange-*.c \
- $srcdir/$subdir/uns-interchange-*.c ] ]
-set block_files [lsort [glob -nocomplain $srcdir/$subdir/block-*.c \
- $srcdir/$subdir/uns-block-*.c ] ]
+set opt_files [lsort [glob -nocomplain $srcdir/$subdir/interchange-*.c \
+ $srcdir/$subdir/uns-interchange-*.c \
+ $srcdir/$subdir/block-*.c \
+ $srcdir/$subdir/uns-block-*.c ] ]
set vect_files [lsort [glob -nocomplain $srcdir/$subdir/vect-*.c ] ]
# Tests to be compiled.
# Tests to be run.
set dg-do-what-default run
dg-runtest $run_id_files "" "-O2 -fgraphite-identity"
-dg-runtest $interchange_files "" "-O2 -floop-interchange -fno-loop-block -fno-loop-strip-mine -ffast-math -fdump-tree-graphite-all"
-dg-runtest $block_files "" "-O2 -floop-block -fno-loop-strip-mine -fno-loop-interchange -ffast-math -fdump-tree-graphite-all"
+dg-runtest $opt_files "" "-O2 -ffast-math -floop-nest-optimize -fdump-tree-graphite-all"
# Vectorizer tests, to be run or compiled, depending on target capabilities.
global DEFAULT_VECTCFLAGS
foreach f $scop_files {lremove wait_to_run_files $f}
foreach f $id_files {lremove wait_to_run_files $f}
foreach f $run_id_files {lremove wait_to_run_files $f}
-foreach f $interchange_files {lremove wait_to_run_files $f}
-foreach f $block_files {lremove wait_to_run_files $f}
+foreach f $opt_files {lremove wait_to_run_files $f}
foreach f $vect_files {lremove wait_to_run_files $f}
dg-runtest $wait_to_run_files "" "-ansi -pedantic-errors"
return 0;
}
-/* { dg-final { scan-tree-dump-times "will be interchanged" 1 "graphite" } } */
+/* { dg-final { scan-tree-dump-times "tiled by" 2 "graphite" } } */
return 0;
}
-/* { dg-final { scan-tree-dump-times "will be interchanged" 1 "graphite" } } */
+/* { dg-final { scan-tree-dump-times "tiled by" 2 "graphite" } } */
return 0;
}
-/* { dg-final { scan-tree-dump-times "will be interchanged" 2 "graphite" } } */
+/* { dg-final { scan-tree-dump-times "tiled by" 4 "graphite" } } */
return 0;
}
-/* { dg-final { scan-tree-dump-times "will be interchanged" 1 "graphite" } } */
+/* { dg-final { scan-tree-dump-times "tiled by" 2 "graphite" } } */
return 0;
}
-/* { dg-final { scan-tree-dump-times "will be interchanged" 1 "graphite" { xfail *-*-* } } } */
+/* { dg-final { scan-tree-dump-times "tiled by" 5 "graphite" } } */
}
-/* { dg-final { scan-tree-dump-times "will be interchanged" 1 "graphite" } } */
+/* { dg-final { scan-tree-dump-times "tiled by" 2 "graphite" } } */
return 0;
}
-/* PRE destroys the perfect nest and we can't cope with that yet. */
-/* { dg-final { scan-tree-dump-times "will be interchanged" 1 "graphite" { xfail *-*-* } } } */
+/* { dg-final { scan-tree-dump-times "tiled by" 7 "graphite" } } */
return 0;
}
-/* PRE destroys the perfect nest and we can't cope with that yet. */
-/* { dg-final { scan-tree-dump-times "will be interchanged" 1 "graphite" { xfail *-*-* } } } */
-
+/* { dg-final { scan-tree-dump-times "tiled by" 4 "graphite" } } */
return 0;
}
-/* { dg-final { scan-tree-dump-times "will be interchanged" 1 "graphite" } } */
+/* { dg-final { scan-tree-dump-times "tiled by" 2 "graphite" } } */
return 0;
}
-/* { dg-final { scan-tree-dump-times "will be interchanged" 1 "graphite" } } */
+/* { dg-final { scan-tree-dump-times "tiled by" 2 "graphite" } } */
return 0;
}
-/* { dg-final { scan-tree-dump-times "will be interchanged" 1 "graphite" } } */
+/* { dg-final { scan-tree-dump-times "tiled by" 2 "graphite" } } */
return 0;
}
-/* { dg-final { scan-tree-dump-times "will be interchanged" 1 "graphite" } } */
+/* { dg-final { scan-tree-dump-times "tiled by" 2 "graphite" } } */
return 0;
}
-/* { dg-final { scan-tree-dump-times "will be interchanged" 1 "graphite" } } */
+/* { dg-final { scan-tree-dump-times "tiled by" 2 "graphite" } } */
return 0;
}
-/* { dg-final { scan-tree-dump-times "will be interchanged" 2 "graphite" } } */
+/* { dg-final { scan-tree-dump-times "tiled by" 5 "graphite" } } */
return 0;
}
-/* { dg-final { scan-tree-dump-times "will be interchanged" 1 "graphite" { xfail *-*-* } } } */
+/* { dg-final { scan-tree-dump-times "tiled by" 3 "graphite" } } */
return 0;
}
-/* PRE destroys the perfect nest and we can't cope with that yet. */
-/* { dg-final { scan-tree-dump-times "will be interchanged" 1 "graphite" { xfail *-*-* } } } */
-
+/* { dg-final { scan-tree-dump-times "tiled by" 7 "graphite" } } */
-/* { dg-options "-O2 -floop-block -fno-loop-strip-mine -fno-loop-interchange -ffast-math -fdump-tree-graphite-all" } */
+/* { dg-options "-O2 -floop-block -ffast-math -fdump-tree-graphite-all" } */
typedef unsigned char UChar;
typedef int Int32;
}
AssertH ( j < 256, 1005 );
}
-/* { dg-final { scan-tree-dump-times "Loop blocked" 1 "graphite" { xfail *-*-* }} } */
+
+/* { dg-final { scan-tree-dump-times "tiled by" 1 "graphite" } } */
return 0;
}
-/* { dg-final { scan-tree-dump-times "will be loop blocked" 3 "graphite" } } */
+/* { dg-final { scan-tree-dump-times "tiled by" 4 "graphite" } } */
return 0;
}
-/* { dg-final { scan-tree-dump-times "will be interchanged" 1 "graphite" } } */
+/* { dg-final { scan-tree-dump-times "tiled by" 4 "graphite" } } */
return 0;
}
-/* PRE destroys the perfect nest and we can't cope with that yet. */
-/* { dg-final { scan-tree-dump-times "will be interchanged" 1 "graphite" } } */
+/* { dg-final { scan-tree-dump-times "tiled by" 6 "graphite" } } */
return 0;
}
-/* PRE destroys the perfect nest and we can't cope with that yet. */
-/* { dg-final { scan-tree-dump-times "will be interchanged" 1 "graphite" } } */
-
+/* { dg-final { scan-tree-dump-times "tiled by" 2 "graphite" } } */
return 0;
}
-/* { dg-final { scan-tree-dump-times "will be interchanged" 1 "graphite" } } */
+/* { dg-final { scan-tree-dump-times "tiled by" 2 "graphite" } } */
return 0;
}
-/* PRE destroys the perfect nest and we can't cope with that yet. */
-/* { dg-final { scan-tree-dump-times "will be interchanged" 1 "graphite" } } */
-
+/* { dg-final { scan-tree-dump-times "tiled by" 3 "graphite" } } */
dg-init
set wait_to_run_files [lsort [glob -nocomplain $srcdir/$subdir/*.\[fF\]{,90,95,03,08} ] ]
-set block_files [lsort [glob -nocomplain $srcdir/$subdir/block-*.\[fF\]{,90,95,03,08} ] ]
set id_files [lsort [glob -nocomplain $srcdir/$subdir/id-*.\[fF\]{,90,95,03,08} ] ]
-set interchange_files [lsort [glob -nocomplain $srcdir/$subdir/interchange-*.\[fF\]{,90,95,03,08} ] ]
+set opt_files [lsort [glob -nocomplain $srcdir/$subdir/interchange-*.\[fF\]{,90,95,03,08} \
+ $srcdir/$subdir/block-*.\[fF\]{,90,95,03,08} ] ]
set scop_files [lsort [glob -nocomplain $srcdir/$subdir/scop-*.\[fF\]{,90,95,03,08} ] ]
set run_id_files [lsort [glob -nocomplain $srcdir/$subdir/run-id-*.\[fF\]{,90,95,03,08} ] ]
set vect_files [lsort [glob -nocomplain $srcdir/$subdir/vect-*.\[fF\]{,90,95,03,08} ] ]
set dg-do-what-default compile
gfortran-dg-runtest $scop_files "" "-O2 -fgraphite -fdump-tree-graphite-all"
gfortran-dg-runtest $id_files "" "-O2 -fgraphite-identity -ffast-math"
-gfortran-dg-runtest $interchange_files "" "-O2 -floop-interchange -fno-loop-block -fno-loop-strip-mine -ffast-math -fdump-tree-graphite-all"
-gfortran-dg-runtest $block_files "" "-O2 -floop-block -fno-loop-strip-mine -fno-loop-interchange -ffast-math -fdump-tree-graphite-all"
+gfortran-dg-runtest $opt_files "" "-O2 -floop-nest-optimize -ffast-math -fdump-tree-graphite-all"
# Vectorizer tests, to be run or compiled, depending on target capabilities.
if [check_vect_support_and_set_flags] {
# The default action for the rest of the files is 'compile'.
set dg-do-what-default compile
-foreach f $block_files {lremove wait_to_run_files $f}
foreach f $id_files {lremove wait_to_run_files $f}
-foreach f $interchange_files {lremove wait_to_run_files $f}
+foreach f $opt_files {lremove wait_to_run_files $f}
foreach f $scop_files {lremove wait_to_run_files $f}
foreach f $run_id_files {lremove wait_to_run_files $f}
foreach f $vect_files {lremove wait_to_run_files $f}
end Program FOO
-! { dg-final { scan-tree-dump-times "will be interchanged" 1 "graphite" } }
+! { dg-final { scan-tree-dump-times "tiled by" 2 "graphite" } }
projects / gcc.git / commitdiff