* [PATCH GCC8][31/33]Set range information for niter bound of vectorized loop @ 2017-04-18 10:54 Bin Cheng 2017-05-11 11:06 ` Richard Biener 0 siblings, 1 reply; 7+ messages in thread From: Bin Cheng @ 2017-04-18 10:54 UTC (permalink / raw) To: gcc-patches; +Cc: nd [-- Attachment #1: Type: text/plain, Size: 453 bytes --] Hi, Based on vect_peeling algorithm, we know for sure that vectorized loop must iterates at least once. This patch sets range information for niter bounds of vectorized loop. This helps niter analysis, so iv elimination too. Is it OK? Thanks, bin 2017-04-11 Bin Cheng <bin.cheng@arm.com> * tree-vect-loop-manip.c (vect_gen_vector_loop_niters): Refactor. Set range information for vector loop bound variable. (vect_do_peeling): Ditto. [-- Attachment #2: 0031-range_info-for-vect_loop-niters-20170224.txt --] [-- Type: text/plain, Size: 4006 bytes --] From 0962735526bf474591e410104d6c9576691e0a1f Mon Sep 17 00:00:00 2001 From: Bin Cheng <binche01@e108451-lin.cambridge.arm.com> Date: Tue, 14 Mar 2017 17:46:55 +0000 Subject: [PATCH 31/33] range_info-for-vect_loop-niters-20170224.txt --- gcc/tree-vect-loop-manip.c | 39 +++++++++++++++++++++++---------------- 1 file changed, 23 insertions(+), 16 deletions(-) diff --git a/gcc/tree-vect-loop-manip.c b/gcc/tree-vect-loop-manip.c index faeaa6d..0fc8cd3 100644 --- a/gcc/tree-vect-loop-manip.c +++ b/gcc/tree-vect-loop-manip.c @@ -1177,22 +1177,21 @@ vect_gen_vector_loop_niters (loop_vec_info loop_vinfo, tree niters, tree *niters_vector_ptr, bool niters_no_overflow) { tree ni_minus_gap, var; - tree niters_vector; + tree niters_vector, type = TREE_TYPE (niters); int vf = LOOP_VINFO_VECT_FACTOR (loop_vinfo); edge pe = loop_preheader_edge (LOOP_VINFO_LOOP (loop_vinfo)); - tree log_vf = build_int_cst (TREE_TYPE (niters), exact_log2 (vf)); + tree log_vf = build_int_cst (type, exact_log2 (vf)); /* If epilogue loop is required because of data accesses with gaps, we subtract one iteration from the total number of iterations here for correct calculation of RATIO. */ if (LOOP_VINFO_PEELING_FOR_GAPS (loop_vinfo)) { - ni_minus_gap = fold_build2 (MINUS_EXPR, TREE_TYPE (niters), - niters, - build_one_cst (TREE_TYPE (niters))); + ni_minus_gap = fold_build2 (MINUS_EXPR, type, niters, + build_one_cst (type)); if (!is_gimple_val (ni_minus_gap)) { - var = create_tmp_var (TREE_TYPE (niters), "ni_gap"); + var = create_tmp_var (type, "ni_gap"); gimple *stmts = NULL; ni_minus_gap = force_gimple_operand (ni_minus_gap, &stmts, true, var); @@ -1208,25 +1207,28 @@ vect_gen_vector_loop_niters (loop_vec_info loop_vinfo, tree niters, (niters - vf) >> log2(vf) + 1 by using the fact that we know ratio will be at least one. */ if (niters_no_overflow) - niters_vector = fold_build2 (RSHIFT_EXPR, TREE_TYPE (niters), - ni_minus_gap, log_vf); + niters_vector = fold_build2 (RSHIFT_EXPR, type, ni_minus_gap, log_vf); else niters_vector - = fold_build2 (PLUS_EXPR, TREE_TYPE (niters), - fold_build2 (RSHIFT_EXPR, TREE_TYPE (niters), - fold_build2 (MINUS_EXPR, TREE_TYPE (niters), - ni_minus_gap, - build_int_cst - (TREE_TYPE (niters), vf)), + = fold_build2 (PLUS_EXPR, type, + fold_build2 (RSHIFT_EXPR, type, + fold_build2 (MINUS_EXPR, type, ni_minus_gap, + build_int_cst (type, vf)), log_vf), - build_int_cst (TREE_TYPE (niters), 1)); + build_int_cst (type, 1)); if (!is_gimple_val (niters_vector)) { - var = create_tmp_var (TREE_TYPE (niters), "bnd"); + var = create_tmp_var (type, "bnd"); gimple *stmts = NULL; niters_vector = force_gimple_operand (niters_vector, &stmts, true, var); gsi_insert_seq_on_edge_immediate (pe, stmts); + /* Peeling algorithm guarantees that vector loop bound is at least ONE, + we set range information to make niters analyzer's life easier. */ + if (TREE_CODE (niters_vector) == SSA_NAME) + set_range_info (niters_vector, VR_RANGE, build_int_cst (type, 1), + fold_build2 (RSHIFT_EXPR, type, + TYPE_MAX_VALUE (type), log_vf)); } *niters_vector_ptr = niters_vector; @@ -1773,6 +1775,11 @@ vect_do_peeling (loop_vec_info loop_vinfo, tree niters, tree nitersm1, = fold_build2 (MINUS_EXPR, type, LOOP_VINFO_NITERSM1 (loop_vinfo), niters_prolog); niters = vect_build_loop_niters (loop_vinfo); + /* It's guaranteed that vector loop bound before vectorization is at + least VF, so set range information. */ + if (TREE_CODE (niters) == SSA_NAME) + set_range_info (niters, VR_RANGE, + build_int_cst (type, vf), TYPE_MAX_VALUE (type)); /* Prolog iterates at most bound_prolog times, latch iterates at most bound_prolog - 1 times. */ -- 1.9.1 ^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH GCC8][31/33]Set range information for niter bound of vectorized loop 2017-04-18 10:54 [PATCH GCC8][31/33]Set range information for niter bound of vectorized loop Bin Cheng @ 2017-05-11 11:06 ` Richard Biener 2017-05-15 16:10 ` Bin.Cheng 0 siblings, 1 reply; 7+ messages in thread From: Richard Biener @ 2017-05-11 11:06 UTC (permalink / raw) To: Bin Cheng; +Cc: gcc-patches, nd On Tue, Apr 18, 2017 at 12:54 PM, Bin Cheng <Bin.Cheng@arm.com> wrote > Hi, > Based on vect_peeling algorithm, we know for sure that vectorized loop must iterates at least once. > This patch sets range information for niter bounds of vectorized loop. This helps niter analysis, > so iv elimination too. > Is it OK? niters_vector = force_gimple_operand (niters_vector, &stmts, true, var); gsi_insert_seq_on_edge_immediate (pe, stmts); + /* Peeling algorithm guarantees that vector loop bound is at least ONE, + we set range information to make niters analyzer's life easier. */ + if (TREE_CODE (niters_vector) == SSA_NAME) + set_range_info (niters_vector, VR_RANGE, build_int_cst (type, 1), + fold_build2 (RSHIFT_EXPR, type, + TYPE_MAX_VALUE (type), log_vf)); if all of niters_vector folds to an original SSA name then niters_vector after gimplification is not a new SSA name and thus you can't set range-info on it. Likewise for the other case where LOOP_VINFO_NITERS is just an SSA name. Richard. > Thanks, > bin > 2017-04-11 Bin Cheng <bin.cheng@arm.com> > > * tree-vect-loop-manip.c (vect_gen_vector_loop_niters): Refactor. > Set range information for vector loop bound variable. > (vect_do_peeling): Ditto. ^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH GCC8][31/33]Set range information for niter bound of vectorized loop 2017-05-11 11:06 ` Richard Biener @ 2017-05-15 16:10 ` Bin.Cheng 2017-05-19 12:57 ` Richard Biener 0 siblings, 1 reply; 7+ messages in thread From: Bin.Cheng @ 2017-05-15 16:10 UTC (permalink / raw) To: Richard Biener; +Cc: gcc-patches [-- Attachment #1: Type: text/plain, Size: 2003 bytes --] On Thu, May 11, 2017 at 12:02 PM, Richard Biener <richard.guenther@gmail.com> wrote: > On Tue, Apr 18, 2017 at 12:54 PM, Bin Cheng <Bin.Cheng@arm.com> wrote >> Hi, >> Based on vect_peeling algorithm, we know for sure that vectorized loop must iterates at least once. >> This patch sets range information for niter bounds of vectorized loop. This helps niter analysis, >> so iv elimination too. >> Is it OK? > > niters_vector = force_gimple_operand (niters_vector, &stmts, true, var); > gsi_insert_seq_on_edge_immediate (pe, stmts); > + /* Peeling algorithm guarantees that vector loop bound is at least ONE, > + we set range information to make niters analyzer's life easier. */ > + if (TREE_CODE (niters_vector) == SSA_NAME) > + set_range_info (niters_vector, VR_RANGE, build_int_cst (type, 1), > + fold_build2 (RSHIFT_EXPR, type, > + TYPE_MAX_VALUE (type), log_vf)); > > if all of niters_vector folds to an original SSA name then > niters_vector after gimplification > is not a new SSA name and thus you can't set range-info on it. > > Likewise for the other case where LOOP_VINFO_NITERS is just an SSA name. Hi, This is updated patch. It checks whether the result ssa name is newly created tmp and only sets range information if so. Is it OK? Thanks, bin 2017-04-11 Bin Cheng <bin.cheng@arm.com> * tree-vectorizer.h (vect_build_loop_niters): New parameter. * tree-vect-loop-manip.c (vect_build_loop_niters): New parameter. Set true to new parameter if new ssa variable is defined. (vect_gen_vector_loop_niters): Refactor. Set range information for the new vector loop bound variable. (vect_do_peeling): Ditto. > > Richard. > >> Thanks, >> bin >> 2017-04-11 Bin Cheng <bin.cheng@arm.com> >> >> * tree-vect-loop-manip.c (vect_gen_vector_loop_niters): Refactor. >> Set range information for vector loop bound variable. >> (vect_do_peeling): Ditto. [-- Attachment #2: 0005-range_info-for-vect_loop-niters-20170225.txt.patch --] [-- Type: text/x-patch, Size: 5803 bytes --] From fbf6867a1f8d8fed94ce53861ccefb0d704e9f96 Mon Sep 17 00:00:00 2001 From: Bin Cheng <binche01@e108451-lin.cambridge.arm.com> Date: Thu, 11 May 2017 17:26:57 +0100 Subject: [PATCH 5/9] range_info-for-vect_loop-niters-20170225.txt --- gcc/tree-vect-loop-manip.c | 52 +++++++++++++++++++++++++++++----------------- gcc/tree-vectorizer.h | 2 +- 2 files changed, 34 insertions(+), 20 deletions(-) diff --git a/gcc/tree-vect-loop-manip.c b/gcc/tree-vect-loop-manip.c index f48336b..52cd2bb 100644 --- a/gcc/tree-vect-loop-manip.c +++ b/gcc/tree-vect-loop-manip.c @@ -1095,10 +1095,11 @@ vect_update_inits_of_drs (loop_vec_info loop_vinfo, tree niters) /* This function builds ni_name = number of iterations. Statements - are emitted on the loop preheader edge. */ + are emitted on the loop preheader edge. If NEW_VAR_P is not NULL, set + it to TRUE if new ssa_var is generated. */ tree -vect_build_loop_niters (loop_vec_info loop_vinfo) +vect_build_loop_niters (loop_vec_info loop_vinfo, bool *new_var_p) { tree ni = unshare_expr (LOOP_VINFO_NITERS (loop_vinfo)); if (TREE_CODE (ni) == INTEGER_CST) @@ -1114,6 +1115,10 @@ vect_build_loop_niters (loop_vec_info loop_vinfo) if (stmts) gsi_insert_seq_on_edge_immediate (pe, stmts); + if (new_var_p != NULL) + *new_var_p = (TREE_CODE (ni_name) == SSA_NAME + && SSA_NAME_VAR (ni_name) == var); + return ni_name; } } @@ -1177,22 +1182,21 @@ vect_gen_vector_loop_niters (loop_vec_info loop_vinfo, tree niters, tree *niters_vector_ptr, bool niters_no_overflow) { tree ni_minus_gap, var; - tree niters_vector; + tree niters_vector, type = TREE_TYPE (niters); int vf = LOOP_VINFO_VECT_FACTOR (loop_vinfo); edge pe = loop_preheader_edge (LOOP_VINFO_LOOP (loop_vinfo)); - tree log_vf = build_int_cst (TREE_TYPE (niters), exact_log2 (vf)); + tree log_vf = build_int_cst (type, exact_log2 (vf)); /* If epilogue loop is required because of data accesses with gaps, we subtract one iteration from the total number of iterations here for correct calculation of RATIO. */ if (LOOP_VINFO_PEELING_FOR_GAPS (loop_vinfo)) { - ni_minus_gap = fold_build2 (MINUS_EXPR, TREE_TYPE (niters), - niters, - build_one_cst (TREE_TYPE (niters))); + ni_minus_gap = fold_build2 (MINUS_EXPR, type, niters, + build_one_cst (type)); if (!is_gimple_val (ni_minus_gap)) { - var = create_tmp_var (TREE_TYPE (niters), "ni_gap"); + var = create_tmp_var (type, "ni_gap"); gimple *stmts = NULL; ni_minus_gap = force_gimple_operand (ni_minus_gap, &stmts, true, var); @@ -1208,25 +1212,29 @@ vect_gen_vector_loop_niters (loop_vec_info loop_vinfo, tree niters, (niters - vf) >> log2(vf) + 1 by using the fact that we know ratio will be at least one. */ if (niters_no_overflow) - niters_vector = fold_build2 (RSHIFT_EXPR, TREE_TYPE (niters), - ni_minus_gap, log_vf); + niters_vector = fold_build2 (RSHIFT_EXPR, type, ni_minus_gap, log_vf); else niters_vector - = fold_build2 (PLUS_EXPR, TREE_TYPE (niters), - fold_build2 (RSHIFT_EXPR, TREE_TYPE (niters), - fold_build2 (MINUS_EXPR, TREE_TYPE (niters), - ni_minus_gap, - build_int_cst - (TREE_TYPE (niters), vf)), + = fold_build2 (PLUS_EXPR, type, + fold_build2 (RSHIFT_EXPR, type, + fold_build2 (MINUS_EXPR, type, ni_minus_gap, + build_int_cst (type, vf)), log_vf), - build_int_cst (TREE_TYPE (niters), 1)); + build_int_cst (type, 1)); if (!is_gimple_val (niters_vector)) { - var = create_tmp_var (TREE_TYPE (niters), "bnd"); + var = create_tmp_var (type, "bnd"); gimple *stmts = NULL; niters_vector = force_gimple_operand (niters_vector, &stmts, true, var); gsi_insert_seq_on_edge_immediate (pe, stmts); + /* Peeling algorithm guarantees that vector loop bound is at least ONE, + we set range information to make niters analyzer's life easier. */ + if (TREE_CODE (niters_vector) == SSA_NAME + && SSA_NAME_VAR (niters_vector) == var) + set_range_info (niters_vector, VR_RANGE, build_int_cst (type, 1), + fold_build2 (RSHIFT_EXPR, type, + TYPE_MAX_VALUE (type), log_vf)); } *niters_vector_ptr = niters_vector; @@ -1771,7 +1779,13 @@ vect_do_peeling (loop_vec_info loop_vinfo, tree niters, tree nitersm1, LOOP_VINFO_NITERSM1 (loop_vinfo) = fold_build2 (MINUS_EXPR, type, LOOP_VINFO_NITERSM1 (loop_vinfo), niters_prolog); - niters = vect_build_loop_niters (loop_vinfo); + bool new_var_p = false; + niters = vect_build_loop_niters (loop_vinfo, &new_var_p); + /* It's guaranteed that vector loop bound before vectorization is at + least VF, so set range information for newly generated var. */ + if (TREE_CODE (niters) == SSA_NAME && new_var_p) + set_range_info (niters, VR_RANGE, + build_int_cst (type, vf), TYPE_MAX_VALUE (type)); /* Prolog iterates at most bound_prolog times, latch iterates at most bound_prolog - 1 times. */ diff --git a/gcc/tree-vectorizer.h b/gcc/tree-vectorizer.h index 12bb904..f72a1d5 100644 --- a/gcc/tree-vectorizer.h +++ b/gcc/tree-vectorizer.h @@ -1168,7 +1168,7 @@ extern gimple *vect_force_simple_reduction (loop_vec_info, gimple *, bool, bool *, bool); /* Drive for loop analysis stage. */ extern loop_vec_info vect_analyze_loop (struct loop *, loop_vec_info); -extern tree vect_build_loop_niters (loop_vec_info); +extern tree vect_build_loop_niters (loop_vec_info, bool * = NULL); extern void vect_gen_vector_loop_niters (loop_vec_info, tree, tree *, bool); /* Drive for loop transformation stage. */ extern struct loop *vect_transform_loop (loop_vec_info); -- 1.9.1 ^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH GCC8][31/33]Set range information for niter bound of vectorized loop 2017-05-15 16:10 ` Bin.Cheng @ 2017-05-19 12:57 ` Richard Biener 2017-05-22 17:42 ` Bin.Cheng 0 siblings, 1 reply; 7+ messages in thread From: Richard Biener @ 2017-05-19 12:57 UTC (permalink / raw) To: Bin.Cheng; +Cc: gcc-patches On Mon, May 15, 2017 at 5:58 PM, Bin.Cheng <amker.cheng@gmail.com> wrote: > On Thu, May 11, 2017 at 12:02 PM, Richard Biener > <richard.guenther@gmail.com> wrote: >> On Tue, Apr 18, 2017 at 12:54 PM, Bin Cheng <Bin.Cheng@arm.com> wrote >>> Hi, >>> Based on vect_peeling algorithm, we know for sure that vectorized loop must iterates at least once. >>> This patch sets range information for niter bounds of vectorized loop. This helps niter analysis, >>> so iv elimination too. >>> Is it OK? >> >> niters_vector = force_gimple_operand (niters_vector, &stmts, true, var); >> gsi_insert_seq_on_edge_immediate (pe, stmts); >> + /* Peeling algorithm guarantees that vector loop bound is at least ONE, >> + we set range information to make niters analyzer's life easier. */ >> + if (TREE_CODE (niters_vector) == SSA_NAME) >> + set_range_info (niters_vector, VR_RANGE, build_int_cst (type, 1), >> + fold_build2 (RSHIFT_EXPR, type, >> + TYPE_MAX_VALUE (type), log_vf)); >> >> if all of niters_vector folds to an original SSA name then >> niters_vector after gimplification >> is not a new SSA name and thus you can't set range-info on it. >> >> Likewise for the other case where LOOP_VINFO_NITERS is just an SSA name. > Hi, > This is updated patch. It checks whether the result ssa name is newly > created tmp and only sets range information if so. > > Is it OK? A better way to check whether the SSA name is new is to see if 'stmts' filled by force_gimple_operand is non-empty. Richard. > Thanks, > bin > > 2017-04-11 Bin Cheng <bin.cheng@arm.com> > > * tree-vectorizer.h (vect_build_loop_niters): New parameter. > * tree-vect-loop-manip.c (vect_build_loop_niters): New parameter. > Set true to new parameter if new ssa variable is defined. > (vect_gen_vector_loop_niters): Refactor. Set range information > for the new vector loop bound variable. > (vect_do_peeling): Ditto. > >> >> Richard. >> >>> Thanks, >>> bin >>> 2017-04-11 Bin Cheng <bin.cheng@arm.com> >>> >>> * tree-vect-loop-manip.c (vect_gen_vector_loop_niters): Refactor. >>> Set range information for vector loop bound variable. >>> (vect_do_peeling): Ditto. ^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH GCC8][31/33]Set range information for niter bound of vectorized loop 2017-05-19 12:57 ` Richard Biener @ 2017-05-22 17:42 ` Bin.Cheng 2017-05-24 13:51 ` Richard Biener 0 siblings, 1 reply; 7+ messages in thread From: Bin.Cheng @ 2017-05-22 17:42 UTC (permalink / raw) To: Richard Biener; +Cc: gcc-patches [-- Attachment #1: Type: text/plain, Size: 2491 bytes --] On Fri, May 19, 2017 at 1:51 PM, Richard Biener <richard.guenther@gmail.com> wrote: > On Mon, May 15, 2017 at 5:58 PM, Bin.Cheng <amker.cheng@gmail.com> wrote: >> On Thu, May 11, 2017 at 12:02 PM, Richard Biener >> <richard.guenther@gmail.com> wrote: >>> On Tue, Apr 18, 2017 at 12:54 PM, Bin Cheng <Bin.Cheng@arm.com> wrote >>>> Hi, >>>> Based on vect_peeling algorithm, we know for sure that vectorized loop must iterates at least once. >>>> This patch sets range information for niter bounds of vectorized loop. This helps niter analysis, >>>> so iv elimination too. >>>> Is it OK? >>> >>> niters_vector = force_gimple_operand (niters_vector, &stmts, true, var); >>> gsi_insert_seq_on_edge_immediate (pe, stmts); >>> + /* Peeling algorithm guarantees that vector loop bound is at least ONE, >>> + we set range information to make niters analyzer's life easier. */ >>> + if (TREE_CODE (niters_vector) == SSA_NAME) >>> + set_range_info (niters_vector, VR_RANGE, build_int_cst (type, 1), >>> + fold_build2 (RSHIFT_EXPR, type, >>> + TYPE_MAX_VALUE (type), log_vf)); >>> >>> if all of niters_vector folds to an original SSA name then >>> niters_vector after gimplification >>> is not a new SSA name and thus you can't set range-info on it. >>> >>> Likewise for the other case where LOOP_VINFO_NITERS is just an SSA name. >> Hi, >> This is updated patch. It checks whether the result ssa name is newly >> created tmp and only sets range information if so. >> >> Is it OK? > > A better way to check whether the SSA name is new is to see if 'stmts' > filled by force_gimple_operand is non-empty. Hi, Here is updated patch checking empty gimple_seq. Is it OK? Thanks, bin > > Richard. > >> Thanks, >> bin >> >> 2017-04-11 Bin Cheng <bin.cheng@arm.com> >> >> * tree-vectorizer.h (vect_build_loop_niters): New parameter. >> * tree-vect-loop-manip.c (vect_build_loop_niters): New parameter. >> Set true to new parameter if new ssa variable is defined. >> (vect_gen_vector_loop_niters): Refactor. Set range information >> for the new vector loop bound variable. >> (vect_do_peeling): Ditto. >> >>> >>> Richard. >>> >>>> Thanks, >>>> bin >>>> 2017-04-11 Bin Cheng <bin.cheng@arm.com> >>>> >>>> * tree-vect-loop-manip.c (vect_gen_vector_loop_niters): Refactor. >>>> Set range information for vector loop bound variable. >>>> (vect_do_peeling): Ditto. [-- Attachment #2: 0005-range_info-for-vect_loop-niters-20170515.txt.patch --] [-- Type: text/x-patch, Size: 5420 bytes --] diff --git a/gcc/tree-vect-loop-manip.c b/gcc/tree-vect-loop-manip.c index f48336b..93c1542 100644 --- a/gcc/tree-vect-loop-manip.c +++ b/gcc/tree-vect-loop-manip.c @@ -1095,10 +1095,11 @@ vect_update_inits_of_drs (loop_vec_info loop_vinfo, tree niters) /* This function builds ni_name = number of iterations. Statements - are emitted on the loop preheader edge. */ + are emitted on the loop preheader edge. If NEW_VAR_P is not NULL, set + it to TRUE if new ssa_var is generated. */ tree -vect_build_loop_niters (loop_vec_info loop_vinfo) +vect_build_loop_niters (loop_vec_info loop_vinfo, bool *new_var_p) { tree ni = unshare_expr (LOOP_VINFO_NITERS (loop_vinfo)); if (TREE_CODE (ni) == INTEGER_CST) @@ -1112,7 +1113,11 @@ vect_build_loop_niters (loop_vec_info loop_vinfo) var = create_tmp_var (TREE_TYPE (ni), "niters"); ni_name = force_gimple_operand (ni, &stmts, false, var); if (stmts) - gsi_insert_seq_on_edge_immediate (pe, stmts); + { + gsi_insert_seq_on_edge_immediate (pe, stmts); + if (new_var_p != NULL) + *new_var_p = true; + } return ni_name; } @@ -1177,22 +1182,21 @@ vect_gen_vector_loop_niters (loop_vec_info loop_vinfo, tree niters, tree *niters_vector_ptr, bool niters_no_overflow) { tree ni_minus_gap, var; - tree niters_vector; + tree niters_vector, type = TREE_TYPE (niters); int vf = LOOP_VINFO_VECT_FACTOR (loop_vinfo); edge pe = loop_preheader_edge (LOOP_VINFO_LOOP (loop_vinfo)); - tree log_vf = build_int_cst (TREE_TYPE (niters), exact_log2 (vf)); + tree log_vf = build_int_cst (type, exact_log2 (vf)); /* If epilogue loop is required because of data accesses with gaps, we subtract one iteration from the total number of iterations here for correct calculation of RATIO. */ if (LOOP_VINFO_PEELING_FOR_GAPS (loop_vinfo)) { - ni_minus_gap = fold_build2 (MINUS_EXPR, TREE_TYPE (niters), - niters, - build_one_cst (TREE_TYPE (niters))); + ni_minus_gap = fold_build2 (MINUS_EXPR, type, niters, + build_one_cst (type)); if (!is_gimple_val (ni_minus_gap)) { - var = create_tmp_var (TREE_TYPE (niters), "ni_gap"); + var = create_tmp_var (type, "ni_gap"); gimple *stmts = NULL; ni_minus_gap = force_gimple_operand (ni_minus_gap, &stmts, true, var); @@ -1208,25 +1212,28 @@ vect_gen_vector_loop_niters (loop_vec_info loop_vinfo, tree niters, (niters - vf) >> log2(vf) + 1 by using the fact that we know ratio will be at least one. */ if (niters_no_overflow) - niters_vector = fold_build2 (RSHIFT_EXPR, TREE_TYPE (niters), - ni_minus_gap, log_vf); + niters_vector = fold_build2 (RSHIFT_EXPR, type, ni_minus_gap, log_vf); else niters_vector - = fold_build2 (PLUS_EXPR, TREE_TYPE (niters), - fold_build2 (RSHIFT_EXPR, TREE_TYPE (niters), - fold_build2 (MINUS_EXPR, TREE_TYPE (niters), - ni_minus_gap, - build_int_cst - (TREE_TYPE (niters), vf)), + = fold_build2 (PLUS_EXPR, type, + fold_build2 (RSHIFT_EXPR, type, + fold_build2 (MINUS_EXPR, type, ni_minus_gap, + build_int_cst (type, vf)), log_vf), - build_int_cst (TREE_TYPE (niters), 1)); + build_int_cst (type, 1)); if (!is_gimple_val (niters_vector)) { - var = create_tmp_var (TREE_TYPE (niters), "bnd"); - gimple *stmts = NULL; + var = create_tmp_var (type, "bnd"); + gimple_seq stmts = NULL; niters_vector = force_gimple_operand (niters_vector, &stmts, true, var); gsi_insert_seq_on_edge_immediate (pe, stmts); + /* Peeling algorithm guarantees that vector loop bound is at least ONE, + we set range information to make niters analyzer's life easier. */ + if (stmts != NULL) + set_range_info (niters_vector, VR_RANGE, build_int_cst (type, 1), + fold_build2 (RSHIFT_EXPR, type, + TYPE_MAX_VALUE (type), log_vf)); } *niters_vector_ptr = niters_vector; @@ -1771,7 +1778,13 @@ vect_do_peeling (loop_vec_info loop_vinfo, tree niters, tree nitersm1, LOOP_VINFO_NITERSM1 (loop_vinfo) = fold_build2 (MINUS_EXPR, type, LOOP_VINFO_NITERSM1 (loop_vinfo), niters_prolog); - niters = vect_build_loop_niters (loop_vinfo); + bool new_var_p = false; + niters = vect_build_loop_niters (loop_vinfo, &new_var_p); + /* It's guaranteed that vector loop bound before vectorization is at + least VF, so set range information for newly generated var. */ + if (new_var_p) + set_range_info (niters, VR_RANGE, + build_int_cst (type, vf), TYPE_MAX_VALUE (type)); /* Prolog iterates at most bound_prolog times, latch iterates at most bound_prolog - 1 times. */ diff --git a/gcc/tree-vectorizer.h b/gcc/tree-vectorizer.h index 12bb904..f72a1d5 100644 --- a/gcc/tree-vectorizer.h +++ b/gcc/tree-vectorizer.h @@ -1168,7 +1168,7 @@ extern gimple *vect_force_simple_reduction (loop_vec_info, gimple *, bool, bool *, bool); /* Drive for loop analysis stage. */ extern loop_vec_info vect_analyze_loop (struct loop *, loop_vec_info); -extern tree vect_build_loop_niters (loop_vec_info); +extern tree vect_build_loop_niters (loop_vec_info, bool * = NULL); extern void vect_gen_vector_loop_niters (loop_vec_info, tree, tree *, bool); /* Drive for loop transformation stage. */ extern struct loop *vect_transform_loop (loop_vec_info); ^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH GCC8][31/33]Set range information for niter bound of vectorized loop 2017-05-22 17:42 ` Bin.Cheng @ 2017-05-24 13:51 ` Richard Biener 2017-06-07 11:09 ` Bin.Cheng 0 siblings, 1 reply; 7+ messages in thread From: Richard Biener @ 2017-05-24 13:51 UTC (permalink / raw) To: Bin.Cheng; +Cc: gcc-patches On Mon, May 22, 2017 at 7:13 PM, Bin.Cheng <amker.cheng@gmail.com> wrote: > On Fri, May 19, 2017 at 1:51 PM, Richard Biener > <richard.guenther@gmail.com> wrote: >> On Mon, May 15, 2017 at 5:58 PM, Bin.Cheng <amker.cheng@gmail.com> wrote: >>> On Thu, May 11, 2017 at 12:02 PM, Richard Biener >>> <richard.guenther@gmail.com> wrote: >>>> On Tue, Apr 18, 2017 at 12:54 PM, Bin Cheng <Bin.Cheng@arm.com> wrote >>>>> Hi, >>>>> Based on vect_peeling algorithm, we know for sure that vectorized loop must iterates at least once. >>>>> This patch sets range information for niter bounds of vectorized loop. This helps niter analysis, >>>>> so iv elimination too. >>>>> Is it OK? >>>> >>>> niters_vector = force_gimple_operand (niters_vector, &stmts, true, var); >>>> gsi_insert_seq_on_edge_immediate (pe, stmts); >>>> + /* Peeling algorithm guarantees that vector loop bound is at least ONE, >>>> + we set range information to make niters analyzer's life easier. */ >>>> + if (TREE_CODE (niters_vector) == SSA_NAME) >>>> + set_range_info (niters_vector, VR_RANGE, build_int_cst (type, 1), >>>> + fold_build2 (RSHIFT_EXPR, type, >>>> + TYPE_MAX_VALUE (type), log_vf)); >>>> >>>> if all of niters_vector folds to an original SSA name then >>>> niters_vector after gimplification >>>> is not a new SSA name and thus you can't set range-info on it. >>>> >>>> Likewise for the other case where LOOP_VINFO_NITERS is just an SSA name. >>> Hi, >>> This is updated patch. It checks whether the result ssa name is newly >>> created tmp and only sets range information if so. >>> >>> Is it OK? >> >> A better way to check whether the SSA name is new is to see if 'stmts' >> filled by force_gimple_operand is non-empty. > Hi, > Here is updated patch checking empty gimple_seq. Is it OK? Ok. As a bonus you could have used wide-ints to feed set_range_info (not sure if wide_int_ref & allows you to pass '1' literally...) Richard. > Thanks, > bin >> >> Richard. >> >>> Thanks, >>> bin >>> >>> 2017-04-11 Bin Cheng <bin.cheng@arm.com> >>> >>> * tree-vectorizer.h (vect_build_loop_niters): New parameter. >>> * tree-vect-loop-manip.c (vect_build_loop_niters): New parameter. >>> Set true to new parameter if new ssa variable is defined. >>> (vect_gen_vector_loop_niters): Refactor. Set range information >>> for the new vector loop bound variable. >>> (vect_do_peeling): Ditto. >>> >>>> >>>> Richard. >>>> >>>>> Thanks, >>>>> bin >>>>> 2017-04-11 Bin Cheng <bin.cheng@arm.com> >>>>> >>>>> * tree-vect-loop-manip.c (vect_gen_vector_loop_niters): Refactor. >>>>> Set range information for vector loop bound variable. >>>>> (vect_do_peeling): Ditto. ^ permalink raw reply [flat|nested] 7+ messages in thread
* Re: [PATCH GCC8][31/33]Set range information for niter bound of vectorized loop 2017-05-24 13:51 ` Richard Biener @ 2017-06-07 11:09 ` Bin.Cheng 0 siblings, 0 replies; 7+ messages in thread From: Bin.Cheng @ 2017-06-07 11:09 UTC (permalink / raw) To: Richard Biener; +Cc: gcc-patches On Wed, May 24, 2017 at 2:48 PM, Richard Biener <richard.guenther@gmail.com> wrote: > On Mon, May 22, 2017 at 7:13 PM, Bin.Cheng <amker.cheng@gmail.com> wrote: >> On Fri, May 19, 2017 at 1:51 PM, Richard Biener >> <richard.guenther@gmail.com> wrote: >>> On Mon, May 15, 2017 at 5:58 PM, Bin.Cheng <amker.cheng@gmail.com> wrote: >>>> On Thu, May 11, 2017 at 12:02 PM, Richard Biener >>>> <richard.guenther@gmail.com> wrote: >>>>> On Tue, Apr 18, 2017 at 12:54 PM, Bin Cheng <Bin.Cheng@arm.com> wrote >>>>>> Hi, >>>>>> Based on vect_peeling algorithm, we know for sure that vectorized loop must iterates at least once. >>>>>> This patch sets range information for niter bounds of vectorized loop. This helps niter analysis, >>>>>> so iv elimination too. >>>>>> Is it OK? >>>>> >>>>> niters_vector = force_gimple_operand (niters_vector, &stmts, true, var); >>>>> gsi_insert_seq_on_edge_immediate (pe, stmts); >>>>> + /* Peeling algorithm guarantees that vector loop bound is at least ONE, >>>>> + we set range information to make niters analyzer's life easier. */ >>>>> + if (TREE_CODE (niters_vector) == SSA_NAME) >>>>> + set_range_info (niters_vector, VR_RANGE, build_int_cst (type, 1), >>>>> + fold_build2 (RSHIFT_EXPR, type, >>>>> + TYPE_MAX_VALUE (type), log_vf)); >>>>> >>>>> if all of niters_vector folds to an original SSA name then >>>>> niters_vector after gimplification >>>>> is not a new SSA name and thus you can't set range-info on it. >>>>> >>>>> Likewise for the other case where LOOP_VINFO_NITERS is just an SSA name. >>>> Hi, >>>> This is updated patch. It checks whether the result ssa name is newly >>>> created tmp and only sets range information if so. >>>> >>>> Is it OK? >>> >>> A better way to check whether the SSA name is new is to see if 'stmts' >>> filled by force_gimple_operand is non-empty. >> Hi, >> Here is updated patch checking empty gimple_seq. Is it OK? > > Ok. As a bonus you could have used wide-ints to feed set_range_info > (not sure if wide_int_ref & allows you to pass '1' literally...) Patch applied @r248958. Being lazy not using wide_int here, I will pick up this after current work. Thanks, bin > > Richard. > >> Thanks, >> bin >>> >>> Richard. >>> >>>> Thanks, >>>> bin >>>> >>>> 2017-04-11 Bin Cheng <bin.cheng@arm.com> >>>> >>>> * tree-vectorizer.h (vect_build_loop_niters): New parameter. >>>> * tree-vect-loop-manip.c (vect_build_loop_niters): New parameter. >>>> Set true to new parameter if new ssa variable is defined. >>>> (vect_gen_vector_loop_niters): Refactor. Set range information >>>> for the new vector loop bound variable. >>>> (vect_do_peeling): Ditto. >>>> >>>>> >>>>> Richard. >>>>> >>>>>> Thanks, >>>>>> bin >>>>>> 2017-04-11 Bin Cheng <bin.cheng@arm.com> >>>>>> >>>>>> * tree-vect-loop-manip.c (vect_gen_vector_loop_niters): Refactor. >>>>>> Set range information for vector loop bound variable. >>>>>> (vect_do_peeling): Ditto. ^ permalink raw reply [flat|nested] 7+ messages in thread
end of thread, other threads:[~2017-06-07 11:09 UTC | newest] Thread overview: 7+ messages (download: mbox.gz / follow: Atom feed) -- links below jump to the message on this page -- 2017-04-18 10:54 [PATCH GCC8][31/33]Set range information for niter bound of vectorized loop Bin Cheng 2017-05-11 11:06 ` Richard Biener 2017-05-15 16:10 ` Bin.Cheng 2017-05-19 12:57 ` Richard Biener 2017-05-22 17:42 ` Bin.Cheng 2017-05-24 13:51 ` Richard Biener 2017-06-07 11:09 ` Bin.Cheng
This is a public inbox, see mirroring instructions for how to clone and mirror all data and code used for this inbox; as well as URLs for read-only IMAP folder(s) and NNTP newsgroup(s).