[gcc r15-3716] tree-optimization/116768 - wrong dependence analysis

https://gcc.gnu.org/g:5b5a36b122e1205449f1512bf39521b669e713ef commit r15-3716-g5b5a36b122e1205449f1512bf39521b669e713ef Author: Richard Biener Date: Thu Sep 19 14:58:18 2024 +0200 tree-optimization/116768 - wrong dependence analysis The following reverts a bogus fix done for PR10

[gcc r15-3818] tree-optimization/114855 - high update_ssa time

https://gcc.gnu.org/g:f9dfe8dea31bf5c56aa7798a0905707faf9e7ec4 commit r15-3818-gf9dfe8dea31bf5c56aa7798a0905707faf9e7ec4 Author: Richard Biener Date: Mon Sep 23 15:41:14 2024 +0200 tree-optimization/114855 - high update_ssa time Part of the problem in PR114855 is high update_ssa t

[gcc r15-3820] tree-optimization/115372 - failed store-lanes in some cases

https://gcc.gnu.org/g:f594008dcced0ebb86908f3d7602fcf943e05bc7 commit r15-3820-gf594008dcced0ebb86908f3d7602fcf943e05bc7 Author: Richard Biener Date: Fri Sep 20 15:07:24 2024 +0200 tree-optimization/115372 - failed store-lanes in some cases The gcc.target/riscv/rvv/autovec/struct/

[gcc r15-3853] Disable add_store_equivs when -fno-expensive-optimizations

https://gcc.gnu.org/g:0b2d3bfa38ccce0dda46aba023f64440cc638496 commit r15-3853-g0b2d3bfa38ccce0dda46aba023f64440cc638496 Author: Richard Biener Date: Tue Sep 24 14:21:33 2024 +0200 Disable add_store_equivs when -fno-expensive-optimizations IRAs add_store_equivs is quadratic in the

[gcc r15-3852] tree-optimization/114855 - slow VRP due to equiv oracle queries

https://gcc.gnu.org/g:caf3fe7880e62692da45489dc5bcae069c1555c8 commit r15-3852-gcaf3fe7880e62692da45489dc5bcae069c1555c8 Author: Richard Biener Date: Tue Sep 24 11:47:26 2024 +0200 tree-optimization/114855 - slow VRP due to equiv oracle queries For the testcase in PR114855 VRP tak

[gcc r15-3854] rtl-optimization/114855 - slow add_store_equivs in IRA

https://gcc.gnu.org/g:cc141b56b367b3d81c1b590e22ae174f1e013009 commit r15-3854-gcc141b56b367b3d81c1b590e22ae174f1e013009 Author: Richard Biener Date: Tue Sep 24 14:10:13 2024 +0200 rtl-optimization/114855 - slow add_store_equivs in IRA For the testcase in PR114855 at -O1 add_store

gcc-cvs@gcc.gnu.org

https://gcc.gnu.org/g:6efc770a71b7227cdbdc24c947ce1fef10794f4c commit r15-3864-g6efc770a71b7227cdbdc24c947ce1fef10794f4c Author: Richard Biener Date: Wed Sep 25 13:15:42 2024 +0200 Speed up wide_int_storage::operator=(wide_int_storage const&) wide_int_storage shows up high in the

[gcc r15-3865] Speed up get_bitmask_from_range

https://gcc.gnu.org/g:0e095df4b4c61bc57413cf3cbca56ff1a038de40 commit r15-3865-g0e095df4b4c61bc57413cf3cbca56ff1a038de40 Author: Richard Biener Date: Wed Sep 25 13:37:59 2024 +0200 Speed up get_bitmask_from_range When min != max we know min ^ max != 0. * value-ran

[gcc r15-3857] Use tree view for find_always_executed_bbs result

https://gcc.gnu.org/g:7ad17fe095772f71c2f76bc75f92dcb93941ad96 commit r15-3857-g7ad17fe095772f71c2f76bc75f92dcb93941ad96 Author: Richard Biener Date: Wed Sep 25 12:52:32 2024 +0200 Use tree view for find_always_executed_bbs result The following makes us use bitmap tree view for th

[gcc r15-3860] Remove recursion in simplify_control_stmt_condition_1 [PR114855].

https://gcc.gnu.org/g:9b7626383822799d60ea3c62e62e700f16337cd6 commit r15-3860-g9b7626383822799d60ea3c62e62e700f16337cd6 Author: Aldy Hernandez Date: Mon Aug 26 07:33:38 2024 +0200 Remove recursion in simplify_control_stmt_condition_1 [PR114855]. Remove some ad-hoc simplification

[gcc r15-3861] remove dominator recursion from reassoc

https://gcc.gnu.org/g:af8ff0047e45b95c8b7b9dd291b4978fcdf9001d commit r15-3861-gaf8ff0047e45b95c8b7b9dd291b4978fcdf9001d Author: Richard Biener Date: Wed Sep 25 12:46:28 2024 +0200 remove dominator recursion from reassoc The reassoc pass currently walks dominators in a recursive w

[gcc r15-4005] testsuite/116654 - adjust gcc.dg/vect/costmodel/ppc/costmodel-slp-12.c

https://gcc.gnu.org/g:b56dc0a9ac403891ddad8bce1d697ed7f5c365f7 commit r15-4005-gb56dc0a9ac403891ddad8bce1d697ed7f5c365f7 Author: Richard Biener Date: Wed Oct 2 08:25:00 2024 +0200 testsuite/116654 - adjust gcc.dg/vect/costmodel/ppc/costmodel-slp-12.c As we now SLP non-grouped stor

[gcc r15-4006] testsuite/116654 - adjust gcc.target/powerpc/p9-vec-length-full-8.c

https://gcc.gnu.org/g:eb0698ae62fedae90e5c5cc071c8aa5c1cb2f74d commit r15-4006-geb0698ae62fedae90e5c5cc071c8aa5c1cb2f74d Author: Richard Biener Date: Wed Oct 2 08:41:30 2024 +0200 testsuite/116654 - adjust gcc.target/powerpc/p9-vec-length-full-8.c gcc.target/powerpc/p9-vec-length-

[gcc r15-4007] Fix gcc.dg/pr116905.c

https://gcc.gnu.org/g:9175d08f756c1664ecb397fdacca150409dcca49 commit r15-4007-g9175d08f756c1664ecb397fdacca150409dcca49 Author: Richard Biener Date: Wed Oct 2 08:49:06 2024 +0200 Fix gcc.dg/pr116905.c I missed { dg-add-options float16 }. * gcc.dg/pr116905.c: Add

[gcc r15-4004] tree-optimization/116654 - missed dr_explicit_realign[_optimized] with SLP

https://gcc.gnu.org/g:e592ea8ddffb44bab52d36d9862bb39ffae4d144 commit r15-4004-ge592ea8ddffb44bab52d36d9862bb39ffae4d144 Author: Richard Biener Date: Tue Oct 1 15:17:18 2024 +0200 tree-optimization/116654 - missed dr_explicit_realign[_optimized] with SLP With single-lane SLP we mi

[gcc r15-4012] tree-optimization/116566 - single lane SLP for VLA inductions

https://gcc.gnu.org/g:ba7632674a2a9ba8193f082c8ca9614c642de3b7 commit r15-4012-gba7632674a2a9ba8193f082c8ca9614c642de3b7 Author: Richard Biener Date: Mon Sep 30 17:06:24 2024 +0200 tree-optimization/116566 - single lane SLP for VLA inductions The following adds SLP support for vec

[gcc r15-4013] testsuite/116660 - adjust testcases unexpectedly failing on 32bit sparc

https://gcc.gnu.org/g:79ea0aab75732c26c38d4b64f1d97acedf80155a commit r15-4013-g79ea0aab75732c26c38d4b64f1d97acedf80155a Author: Richard Biener Date: Wed Oct 2 11:27:09 2024 +0200 testsuite/116660 - adjust testcases unexpectedly failing on 32bit sparc Both testcases miss some effe

[gcc r15-4025] Replace another missed iterative_hash_object

https://gcc.gnu.org/g:dafbfdb01559092aca6be0e31afc2b39d3fdb2dd commit r15-4025-gdafbfdb01559092aca6be0e31afc2b39d3fdb2dd Author: Richard Biener Date: Wed Oct 2 14:18:05 2024 +0200 Replace another missed iterative_hash_object I missed one that's actually hit quite a lot, hashing of

[gcc r15-4038] Restore aarch64 bootstrap

https://gcc.gnu.org/g:34670dee56a6bd030094801b3da900b1d214b454 commit r15-4038-g34670dee56a6bd030094801b3da900b1d214b454 Author: Richard Biener Date: Thu Oct 3 13:26:21 2024 +0200 Restore aarch64 bootstrap This zero-initializes vec_init to avoid a bogus maybe-uninitialized dia

[gcc r15-5012] Check LOOP_VINFO_PEELING_FOR_GAPS on epilog is supported

https://gcc.gnu.org/g:76048bd0693e30a5abc67aa6dcce9f4973ea208e commit r15-5012-g76048bd0693e30a5abc67aa6dcce9f4973ea208e Author: Richard Biener Date: Mon Nov 4 13:03:33 2024 +0100 Check LOOP_VINFO_PEELING_FOR_GAPS on epilog is supported We need to check that an epilogue doesn't re

[gcc r15-5014] Add LOOP_VINFO_MAIN_LOOP_INFO

https://gcc.gnu.org/g:42d99f63cfccabe1d19177993abf4f1219d6f967 commit r15-5014-g42d99f63cfccabe1d19177993abf4f1219d6f967 Author: Richard Biener Date: Mon Nov 4 12:58:41 2024 +0100 Add LOOP_VINFO_MAIN_LOOP_INFO The following introduces LOOP_VINFO_MAIN_LOOP_INFO alongside LOOP_V

[gcc r15-5013] Add LOOP_VINFO_DRS_ADVANCED_BY

https://gcc.gnu.org/g:2c25af0e52a631e46a1731594301e5e63bc28992 commit r15-5013-g2c25af0e52a631e46a1731594301e5e63bc28992 Author: Richard Biener Date: Mon Nov 4 13:09:21 2024 +0100 Add LOOP_VINFO_DRS_ADVANCED_BY The following remembers how we advanced DRs when vectorizing an ep

[gcc r15-4669] Default expand_vec_cond_expr_p code to ERROR_MARK

https://gcc.gnu.org/g:35c3aac80e7a18c32436259143cba0f47bd0cb5a commit r15-4669-g35c3aac80e7a18c32436259143cba0f47bd0cb5a Author: Richard Biener Date: Fri Oct 25 12:38:24 2024 +0200 Default expand_vec_cond_expr_p code to ERROR_MARK As we want to transition to only vcond_mask expand

[gcc r15-5008] rtl-optimization/117467 - 33% compile-time in rest of compilation

https://gcc.gnu.org/g:7a07de2c60b3c513b6aef206e9b55b3ffefe8b39 commit r15-5008-g7a07de2c60b3c513b6aef206e9b55b3ffefe8b39 Author: Richard Biener Date: Thu Nov 7 09:23:03 2024 +0100 rtl-optimization/117467 - 33% compile-time in rest of compilation ext-dce uses TV_NONE, that's not OK

[gcc r15-5015] Disable gather/scatter for non-first vectorized epilogue

https://gcc.gnu.org/g:7cd064a31db3292eb14b1f105005e0399a50957f commit r15-5015-g7cd064a31db3292eb14b1f105005e0399a50957f Author: Richard Biener Date: Tue Nov 5 14:58:59 2024 +0100 Disable gather/scatter for non-first vectorized epilogue We currently make vect_check_gather_scatter

[gcc r15-5038] Enable gcc.dg/vect/vect-early-break_21.c on x86_64

https://gcc.gnu.org/g:a9c31c2e76d9b63225448375a72a80591b43c7a6 commit r15-5038-ga9c31c2e76d9b63225448375a72a80591b43c7a6 Author: Richard Biener Date: Fri Nov 8 12:44:47 2024 +0100 Enable gcc.dg/vect/vect-early-break_21.c on x86_64 The following also enables the testcase on x86 as

[gcc r15-5045] Fix gcc.dg/vect/bb-slp-77.c for x86

https://gcc.gnu.org/g:12383255fe4e82c31f5e42c72a8fbcb1b5dea35d commit r15-5045-g12383255fe4e82c31f5e42c72a8fbcb1b5dea35d Author: Richard Biener Date: Fri Nov 8 15:11:34 2024 +0100 Fix gcc.dg/vect/bb-slp-77.c for x86 x86 doesn't have .REDUC_PLUS for V2SImode - there's no effective

[gcc r15-5649] tree-optimization/117767 - VMAT_STRIDED_SLP and alignment

https://gcc.gnu.org/g:0b47d7579e84daa4dfa16baef1133ab73adc502d commit r15-5649-g0b47d7579e84daa4dfa16baef1133ab73adc502d Author: Richard Biener Date: Mon Nov 25 09:46:28 2024 +0100 tree-optimization/117767 - VMAT_STRIDED_SLP and alignment This plugs another hole in alignment check

[gcc r15-5784] tree-optimization/115438 - SLP reduction vect vs. bwaves

https://gcc.gnu.org/g:a54aa75ab30eb1a176ceaded507113252df24878 commit r15-5784-ga54aa75ab30eb1a176ceaded507113252df24878 Author: Richard Biener Date: Fri Nov 29 12:14:24 2024 +0100 tree-optimization/115438 - SLP reduction vect vs. bwaves 503.bwaves_r shows a case where the non-SLP

[gcc r14-11000] tree-optimization/117307 - STMT_VINFO_SLP_VECT_ONLY mis-computation

https://gcc.gnu.org/g:bbe47a2d763d14de94c1e98104e2fa553783972a commit r14-11000-gbbe47a2d763d14de94c1e98104e2fa553783972a Author: Richard Biener Date: Mon Oct 28 09:52:08 2024 +0100 tree-optimization/117307 - STMT_VINFO_SLP_VECT_ONLY mis-computation STMT_VINFO_SLP_VECT_ONLY isn't

[gcc r14-10999] tree-optimization/117254 - ICE with access diangostics

https://gcc.gnu.org/g:6cfe4fd3fb44b660eaf52302235c4a9992a9ddb4 commit r14-10999-g6cfe4fd3fb44b660eaf52302235c4a9992a9ddb4 Author: Richard Biener Date: Tue Oct 22 11:46:47 2024 +0200 tree-optimization/117254 - ICE with access diangostics The diagnostics code fails to handle non-con

[gcc r14-11001] tree-optimization/117333 - ICE with NULL access size DR

https://gcc.gnu.org/g:8a13eef88742c09d91be090fa8cb9bc3040290e7 commit r14-11001-g8a13eef88742c09d91be090fa8cb9bc3040290e7 Author: Richard Biener Date: Tue Oct 29 09:42:12 2024 +0100 tree-optimization/117333 - ICE with NULL access size DR dr_may_alias_p ICEs when TYPE_SIZE of DR->r

[gcc r14-11004] tree-optimization/117574 - bougs niter lt-to-ne

https://gcc.gnu.org/g:8764b8543b93494861bc034bb82af38db1688c8f commit r14-11004-g8764b8543b93494861bc034bb82af38db1688c8f Author: Richard Biener Date: Fri Nov 15 11:56:14 2024 +0100 tree-optimization/117574 - bougs niter lt-to-ne When trying to change a IV from IV0 < IV1 to IV0' !

[gcc r14-11005] tree-optimization/117594 - fix live op vectorization for length masked case

https://gcc.gnu.org/g:b1e14fc244c6b4543eb0ec1857afcf125cfbd917 commit r14-11005-gb1e14fc244c6b4543eb0ec1857afcf125cfbd917 Author: Richard Biener Date: Mon Nov 18 13:46:52 2024 +0100 tree-optimization/117594 - fix live op vectorization for length masked case The code was passing fa

[gcc r14-11002] tree-optimization/117417 - ICE with complex load optimization

https://gcc.gnu.org/g:1fce67961689a59fb2ffad6f50fc4dab76740174 commit r14-11002-g1fce67961689a59fb2ffad6f50fc4dab76740174 Author: Richard Biener Date: Tue Nov 12 11:15:15 2024 +0100 tree-optimization/117417 - ICE with complex load optimization When we decompose a complex load only

[gcc r14-11003] middle-end/117433 - ICE with gimple BLKmode reg copy

https://gcc.gnu.org/g:a8e01b735a1b7e8386ad7baab54802faf58bae02 commit r14-11003-ga8e01b735a1b7e8386ad7baab54802faf58bae02 Author: Richard Biener Date: Mon Nov 4 11:39:05 2024 +0100 middle-end/117433 - ICE with gimple BLKmode reg copy When we end up expanding a SSA name copy with B

[gcc r15-5413] tree-optimization/117594 - fix live op vectorization for length masked case

https://gcc.gnu.org/g:c108785c425b2042f63fa975c58c274d19a8d160 commit r15-5413-gc108785c425b2042f63fa975c58c274d19a8d160 Author: Richard Biener Date: Mon Nov 18 13:46:52 2024 +0100 tree-optimization/117594 - fix live op vectorization for length masked case The code was passing fac

[gcc r15-5524] tree-optimization/117709 - bogus offset for gather load

https://gcc.gnu.org/g:f5bd88b5e874751d17a2f18c1cfc5bd054da8308 commit r15-5524-gf5bd88b5e874751d17a2f18c1cfc5bd054da8308 Author: Richard Biener Date: Wed Nov 20 16:47:08 2024 +0100 tree-optimization/117709 - bogus offset for gather load When diverting to VMAT_GATHER_SCATTER we fai

[gcc r15-5496] tree-optimization/117574 - bougs niter lt-to-ne

https://gcc.gnu.org/g:ff5a14abeb31cd6bd0ca55e7043d05c8141a8c7f commit r15-5496-gff5a14abeb31cd6bd0ca55e7043d05c8141a8c7f Author: Richard Biener Date: Fri Nov 15 11:56:14 2024 +0100 tree-optimization/117574 - bougs niter lt-to-ne When trying to change a IV from IV0 < IV1 to IV0' !=

[gcc r15-5523] tree-optimization/117698 - SLP vectorization and alignment

https://gcc.gnu.org/g:2383ed144b54be6bae5ba5aec8859c6c2e215daf commit r15-5523-g2383ed144b54be6bae5ba5aec8859c6c2e215daf Author: Richard Biener Date: Wed Nov 20 13:32:48 2024 +0100 tree-optimization/117698 - SLP vectorization and alignment When SLP vectorizing we fail to mark the

[gcc r15-5318] Remove dead code related to VEC_COND_EXPR expansion from ISEL

https://gcc.gnu.org/g:5a2c3a69fbdcb56a1d21d340d3b9fb5a9b68b717 commit r15-5318-g5a2c3a69fbdcb56a1d21d340d3b9fb5a9b68b717 Author: Richard Biener Date: Tue Nov 12 14:48:23 2024 +0100 Remove dead code related to VEC_COND_EXPR expansion from ISEL ISEL was introduced to translate vecto

[gcc r15-5320] Remove unused vcond{,u,eq} expander infrastructure

https://gcc.gnu.org/g:c57b2f88c93a00f4248b4e5da4de90d142967b4c commit r15-5320-gc57b2f88c93a00f4248b4e5da4de90d142967b4c Author: Richard Biener Date: Fri Nov 15 13:59:02 2024 +0100 Remove unused vcond{,u,eq} expander infrastructure Now that we no longer exercise vcond{,u,eq} patte

[gcc r15-5317] Streamline vector lowering of VEC_COND_EXPR and vector comparisons

https://gcc.gnu.org/g:f40010c198f7cb811c1803834b621f4185803c18 commit r15-5317-gf40010c198f7cb811c1803834b621f4185803c18 Author: Richard Biener Date: Tue Nov 12 15:07:34 2024 +0100 Streamline vector lowering of VEC_COND_EXPR and vector comparisons The following makes sure to lower

[gcc r15-5559] tree-optimization/117720 - check alignment for VMAT_STRIDED_SLP

https://gcc.gnu.org/g:7e9b0d90d363311caaa5a5e6edbb5088afe0308f commit r15-5559-g7e9b0d90d363311caaa5a5e6edbb5088afe0308f Author: Richard Biener Date: Thu Nov 21 09:14:53 2024 +0100 tree-optimization/117720 - check alignment for VMAT_STRIDED_SLP While vectorizable_store was already

[gcc r15-5863] tree-optimization/116352 - SLP scheduling and stmt order

https://gcc.gnu.org/g:5ab3f091b3eb42795340d3c9cea8aaec2060693c commit r15-5863-g5ab3f091b3eb42795340d3c9cea8aaec2060693c Author: Richard Biener Date: Mon Dec 2 11:07:46 2024 +0100 tree-optimization/116352 - SLP scheduling and stmt order The PR uncovers unchecked constraints on the

[gcc r15-5227] tree-optimization/117556 - SLP of live stmts from load-lanes

https://gcc.gnu.org/g:4b4aa47ed296968507b2fde70d2e651129ff3b36 commit r15-5227-g4b4aa47ed296968507b2fde70d2e651129ff3b36 Author: Richard Biener Date: Wed Nov 13 14:43:27 2024 +0100 tree-optimization/117556 - SLP of live stmts from load-lanes The following fixes SLP live lane gener

[gcc r15-5229] tree-optimization/117554 - correct single-element interleaving check

https://gcc.gnu.org/g:72df175c93d81d4eea67d889d76fda7381fb6743 commit r15-5229-g72df175c93d81d4eea67d889d76fda7381fb6743 Author: Richard Biener Date: Wed Nov 13 15:05:00 2024 +0100 tree-optimization/117554 - correct single-element interleaving check In addition to a single DR we a

[gcc r15-5230] Do not consider overrun for VMAT_ELEMENTWISE

https://gcc.gnu.org/g:6d85a0bc2e09221bdb412bc47aefbcd10c546fd5 commit r15-5230-g6d85a0bc2e09221bdb412bc47aefbcd10c546fd5 Author: Richard Biener Date: Wed Nov 13 11:32:13 2024 +0100 Do not consider overrun for VMAT_ELEMENTWISE When we classify an SLP access as VMAT_ELEMENTWISE we s

[gcc r15-5228] tree-optimization/117559 - avoid hybrid SLP for masked load/store lanes

https://gcc.gnu.org/g:ba192895d17c4fc4676444fe1bd4a593d20e6def commit r15-5228-gba192895d17c4fc4676444fe1bd4a593d20e6def Author: Richard Biener Date: Wed Nov 13 13:56:13 2024 +0100 tree-optimization/117559 - avoid hybrid SLP for masked load/store lanes Hybrid analysis is confused

[gcc r15-5234] tree-optimization/117567 - make SLP reassoc resilent against NULL lanes

https://gcc.gnu.org/g:f56343c5ea9a11fa740e0330ec121b826e349879 commit r15-5234-gf56343c5ea9a11fa740e0330ec121b826e349879 Author: Richard Biener Date: Thu Nov 14 10:17:23 2024 +0100 tree-optimization/117567 - make SLP reassoc resilent against NULL lanes The following tries to make

[gcc r15-5235] Remove last comparison-code expand_vec_cond_expr_p call from vectorizer

https://gcc.gnu.org/g:86708a88fb53518f1a54105c1fd006b57f14e7ce commit r15-5235-g86708a88fb53518f1a54105c1fd006b57f14e7ce Author: Richard Biener Date: Tue Nov 12 13:55:14 2024 +0100 Remove last comparison-code expand_vec_cond_expr_p call from vectorizer The following refactors the

[gcc r15-5236] Avoid expand_vec_cond_expr_p with comparison code

https://gcc.gnu.org/g:9ede072ffafcde27d0e9fe76bb7ffacb4f48a2d6 commit r15-5236-g9ede072ffafcde27d0e9fe76bb7ffacb4f48a2d6 Author: Richard Biener Date: Tue Nov 12 14:45:02 2024 +0100 Avoid expand_vec_cond_expr_p with comparison code This removes the obsolete API use by vector divmod

[gcc r15-5337] tree-optimization/117558 - peeling for gaps and VL vectors

https://gcc.gnu.org/g:935aaface9469df644a3d926320da68cd796c067 commit r15-5337-g935aaface9469df644a3d926320da68cd796c067 Author: Richard Biener Date: Fri Nov 15 08:42:04 2024 +0100 tree-optimization/117558 - peeling for gaps and VL vectors The following ensures that peeling a sing

[gcc r15-5339] tree-optimization/117606 - SLP and single element interleaving

https://gcc.gnu.org/g:39e763ca876fba4449c0781f61115e063f481add commit r15-5339-g39e763ca876fba4449c0781f61115e063f481add Author: Richard Biener Date: Fri Nov 15 10:04:23 2024 +0100 tree-optimization/117606 - SLP and single element interleaving The following tries to reduce the amo

[gcc r15-5340] Flip vectorization to forced SLP

https://gcc.gnu.org/g:4b8c5b337e007ea0d1b9b86c523570b2b7732d86 commit r15-5340-g4b8c5b337e007ea0d1b9b86c523570b2b7732d86 Author: Richard Biener Date: Fri Nov 15 10:08:11 2024 +0100 Flip vectorization to forced SLP The following flips the vectorizer to forced SLP, there is almost

[gcc r15-5338] tree-optimization/117605 - SLP with large negative single-element interleaving

https://gcc.gnu.org/g:4621b684aeb846a682de7e857d67b72eb94f458e commit r15-5338-g4621b684aeb846a682de7e857d67b72eb94f458e Author: Richard Biener Date: Fri Nov 15 09:22:37 2024 +0100 tree-optimization/117605 - SLP with large negative single-element interleaving We fail to demote th

[gcc r15-5249] Fix typo in peeling for gap compute of get_group_load_store_type

https://gcc.gnu.org/g:faf594ce210cddcd701c6c37326d54b0f91d50f1 commit r15-5249-gfaf594ce210cddcd701c6c37326d54b0f91d50f1 Author: Richard Biener Date: Thu Nov 14 13:28:48 2024 +0100 Fix typo in peeling for gap compute of get_group_load_store_type When fixing a maybe-uninit diagnost

[gcc r15-5250] Fix another thinko in peeling for gap compute of get_group_load_store_type

https://gcc.gnu.org/g:119c1b76e22ff40bbae7a5b1beeb63a1e30538fa commit r15-5250-g119c1b76e22ff40bbae7a5b1beeb63a1e30538fa Author: Richard Biener Date: Thu Nov 14 14:22:01 2024 +0100 Fix another thinko in peeling for gap compute of get_group_load_store_type There's inconsistent hand

[gcc r15-5651] target/116760 - 416.gamess slowdown with SLP

https://gcc.gnu.org/g:cd8db107b9bef73fd822ffb420f96ed2bc622a19 commit r15-5651-gcd8db107b9bef73fd822ffb420f96ed2bc622a19 Author: Richard Biener Date: Mon Nov 25 13:32:15 2024 +0100 target/116760 - 416.gamess slowdown with SLP For the TWOTFF loop vectorization the backend scales co

[gcc r15-5650] Add extra 64bit SSE vector epilogue in some cases

https://gcc.gnu.org/g:d9c908b75039653f2b7717b4b7cdffdc4f0fcc7d commit r15-5650-gd9c908b75039653f2b7717b4b7cdffdc4f0fcc7d Author: Richard Biener Date: Fri Nov 22 13:58:08 2024 +0100 Add extra 64bit SSE vector epilogue in some cases Similar to the X86_TUNE_AVX512_TWO_EPILOGUES tunin

[gcc r15-5646] tree-optimization/115825 - improve unroll estimates for volatile accesses

https://gcc.gnu.org/g:d1cf0d7a0f27fdd55302785f19f07d1c3f162ba5 commit r15-5646-gd1cf0d7a0f27fdd55302785f19f07d1c3f162ba5 Author: Richard Biener Date: Wed Jul 10 12:45:02 2024 +0200 tree-optimization/115825 - improve unroll estimates for volatile accesses The loop unrolling code as

[gcc r15-5876] tree-optimization/117874 - missed vectorization that's formerly hybrid

https://gcc.gnu.org/g:be8d1a358e3abc50c14a1d7b1cfee82fe6f6aa3c commit r15-5876-gbe8d1a358e3abc50c14a1d7b1cfee82fe6f6aa3c Author: Richard Biener Date: Mon Dec 2 14:59:00 2024 +0100 tree-optimization/117874 - missed vectorization that's formerly hybrid With SLP forced we fail to con

[gcc r15-5894] tree-optimization/117874 - optimize SLP discovery budget use

https://gcc.gnu.org/g:af9a3fe6a52974252516b3eea4c5ab5caae47b4b commit r15-5894-gaf9a3fe6a52974252516b3eea4c5ab5caae47b4b Author: Richard Biener Date: Tue Dec 3 08:56:35 2024 +0100 tree-optimization/117874 - optimize SLP discovery budget use The following tries to avoid eating into

[gcc r15-5893] Use the number of relevant stmts to limit SLP build

https://gcc.gnu.org/g:d4f3c3efc669afe6e270344ff73c7471fea67ce7 commit r15-5893-gd4f3c3efc669afe6e270344ff73c7471fea67ce7 Author: Richard Biener Date: Tue Dec 3 08:52:48 2024 +0100 Use the number of relevant stmts to limit SLP build The following removes scalar stmt counting from l

[gcc r15-5918] tree-optimization/116083 - SLP discovery slowness

https://gcc.gnu.org/g:27b444a41a6afab8020183c4c5d3361e21635031 commit r15-5918-g27b444a41a6afab8020183c4c5d3361e21635031 Author: Richard Biener Date: Tue Dec 3 14:37:21 2024 +0100 tree-optimization/116083 - SLP discovery slowness One large constant factor of SLP discovery is figur

[gcc r15-5943] middle-end/117801 - failed register coalescing due to GIMPLE schedule

https://gcc.gnu.org/g:dc0dea98c96e02c6b24060170bc88da8d4931bc2 commit r15-5943-gdc0dea98c96e02c6b24060170bc88da8d4931bc2 Author: Richard Biener Date: Wed Nov 27 13:36:19 2024 +0100 middle-end/117801 - failed register coalescing due to GIMPLE schedule For a TSVC testcase we see fai

[gcc r15-5956] rtl-optimization/117922 - add timevar for fold-mem-offsets

https://gcc.gnu.org/g:8772f37e45e9401c9a361548e00c9691424e75e0 commit r15-5956-g8772f37e45e9401c9a361548e00c9691424e75e0 Author: Richard Biener Date: Fri Dec 6 08:08:55 2024 +0100 rtl-optimization/117922 - add timevar for fold-mem-offsets The new fold-mem-offsets RTL pass takes si

[gcc r15-5968] Remove some duplicates reported by genmatch

https://gcc.gnu.org/g:1c07f7a1b45b3d6f46cb25f3bf068d94fe4131e8 commit r15-5968-g1c07f7a1b45b3d6f46cb25f3bf068d94fe4131e8 Author: Richard Biener Date: Thu Dec 5 13:24:27 2024 +0100 Remove some duplicates reported by genmatch genmatch currently has a difficulty to decide whether a d

[gcc r15-5969] More duplicates reported by genmatch

https://gcc.gnu.org/g:e65b09d461d2a4f37fb55d6dcd2e94b1f444daea commit r15-5969-ge65b09d461d2a4f37fb55d6dcd2e94b1f444daea Author: Richard Biener Date: Thu Dec 5 13:47:36 2024 +0100 More duplicates reported by genmatch Here are a bit less obvious cases of duplicate, mostly of the

[gcc r15-5959] testsuite/117714 - gcc.dg/vect/slp-reduc-4.c FAILs on 32-bit SPARC

https://gcc.gnu.org/g:3e14c126de157fe9b87a1ada6e689003735088e1 commit r15-5959-g3e14c126de157fe9b87a1ada6e689003735088e1 Author: Richard Biener Date: Fri Dec 6 09:37:35 2024 +0100 testsuite/117714 - gcc.dg/vect/slp-reduc-4.c FAILs on 32-bit SPARC The testcase tries to ensure we ca

[gcc r15-6039] replace atoi with stroul in c_parser_gimple_parse_bb_spec [PR114541]

https://gcc.gnu.org/g:2b8ecbfe2ce6558637d42fdcb7efb3f878feb56b commit r15-6039-g2b8ecbfe2ce6558637d42fdcb7efb3f878feb56b Author: Heiko Eißfeldt Date: Mon Dec 9 10:39:50 2024 +0100 replace atoi with stroul in c_parser_gimple_parse_bb_spec [PR114541] The full treatment of these inva

[gcc r15-6031] middle-end/117932 - further speedup DF worklist solver

https://gcc.gnu.org/g:6b390f8253b7f6575f18e356610aeb5d83e1140f commit r15-6031-g6b390f8253b7f6575f18e356610aeb5d83e1140f Author: Richard Biener Date: Sat Dec 7 14:43:00 2024 +0100 middle-end/117932 - further speedup DF worklist solver The triple-indirect memory reference we perfor

[gcc r15-6030] middle-end/117932 - speed up DF solver

https://gcc.gnu.org/g:57dcb27e7a48151ad5f9a6122c6a40fac31843e9 commit r15-6030-g57dcb27e7a48151ad5f9a6122c6a40fac31843e9 Author: Richard Biener Date: Fri Dec 6 16:36:39 2024 +0100 middle-end/117932 - speed up DF solver The following addresses slow bitmap operations for maintaining

[gcc r15-6045] Free RTL SSA after late-combine

https://gcc.gnu.org/g:91bdaf659d23be61d1fa5c83ef02322bf0872c87 commit r15-6045-g91bdaf659d23be61d1fa5c83ef02322bf0872c87 Author: Richard Biener Date: Mon Dec 9 14:10:34 2024 +0100 Free RTL SSA after late-combine Late-combine fails to release RTL SSA info, leaking memory (as -f

[gcc r15-6044] Assign separate timevar to duplicate computed goto pass

https://gcc.gnu.org/g:75e481c711c0f6698ca063f5a009db71d45ddc0c commit r15-6044-g75e481c711c0f6698ca063f5a009db71d45ddc0c Author: Richard Biener Date: Mon Dec 9 14:09:12 2024 +0100 Assign separate timevar to duplicate computed goto pass It currently shares the timevar with bb-reord

[gcc r15-6484] middle-end/118174 - bogus TER of tailcall

https://gcc.gnu.org/g:f8b559626025d2466c2780af6b62560dda468647 commit r15-6484-gf8b559626025d2466c2780af6b62560dda468647 Author: Richard Biener Date: Tue Dec 31 14:47:03 2024 +0100 middle-end/118174 - bogus TER of tailcall The following avoids applying TER to direct internal funct

[gcc r15-6652] rtl-optimization/118298 - constant iteration loops and #pragma unroll

https://gcc.gnu.org/g:34501ef418da13b361614235077c2162caabab73 commit r15-6652-g34501ef418da13b361614235077c2162caabab73 Author: Richard Biener Date: Tue Jan 7 13:18:27 2025 +0100 rtl-optimization/118298 - constant iteration loops and #pragma unroll When the RTL unroller handles c

[gcc r15-6651] Fixup convert-dfp*.c

https://gcc.gnu.org/g:cda313ba6215d20150db6e952ce62138e9127d21 commit r15-6651-gcda313ba6215d20150db6e952ce62138e9127d21 Author: Richard Biener Date: Tue Jan 7 15:07:12 2025 +0100 Fixup convert-dfp*.c The testcases use -save-temps which doesn't play nice with -flto and multili

[gcc r15-6683] tree-optimization/118269 - SLP reduction chain and early breaks

https://gcc.gnu.org/g:2dd4eb15b63ca1f20a637e8b26567d5064a43a4f commit r15-6683-g2dd4eb15b63ca1f20a637e8b26567d5064a43a4f Author: Richard Biener Date: Tue Jan 7 11:15:43 2025 +0100 tree-optimization/118269 - SLP reduction chain and early breaks When we create the SLP reduction chai

[gcc r15-6684] middle-end/118325 - nonlocal goto lowering

https://gcc.gnu.org/g:640b550425180bdab7bc67edc6a48465f4023888 commit r15-6684-g640b550425180bdab7bc67edc6a48465f4023888 Author: Richard Biener Date: Tue Jan 7 15:11:47 2025 +0100 middle-end/118325 - nonlocal goto lowering When nonlocal goto lowering creates an artificial label it

[gcc r15-7331] tree-optimization/118717 - store commoning vs. abnormals

https://gcc.gnu.org/g:fbcbbfe2bf83eb8b1347144eeca37b06be5a8bb5 commit r15-7331-gfbcbbfe2bf83eb8b1347144eeca37b06be5a8bb5 Author: Richard Biener Date: Mon Feb 3 09:55:50 2025 +0100 tree-optimization/118717 - store commoning vs. abnormals When we sink common stores in cselim or the

[gcc r15-7332] c++/79786 - bougs invocation of DATA_ABI_ALIGNMENT macro

https://gcc.gnu.org/g:6ec19825b4e72611cdbd4749feed67b61392aa81 commit r15-7332-g6ec19825b4e72611cdbd4749feed67b61392aa81 Author: Richard Biener Date: Mon Feb 3 11:27:20 2025 +0100 c++/79786 - bougs invocation of DATA_ABI_ALIGNMENT macro The first argument is supposed to be a type,

[gcc r15-7299] Do not rely on non-SLP analysis for SLP outer loop vectorization

https://gcc.gnu.org/g:9e3ceed8d50912e271e84389ed80aeea47184e59 commit r15-7299-g9e3ceed8d50912e271e84389ed80aeea47184e59 Author: Richard Biener Date: Thu Jan 30 14:52:14 2025 +0100 Do not rely on non-SLP analysis for SLP outer loop vectorization We end up relying on non-SLP analys

[gcc r15-7301] debug/100530 - Revert QUAL_ADDR_SPACE handling from dwarf2out.cc

https://gcc.gnu.org/g:319f1d042179b381becf4bf1d0f6b9dab6e84884 commit r15-7301-g319f1d042179b381becf4bf1d0f6b9dab6e84884 Author: Richard Biener Date: Fri Jan 31 08:56:39 2025 +0100 debug/100530 - Revert QUAL_ADDR_SPACE handling from dwarf2out.cc The bug clearly shows that r8-4385-

[gcc r15-7343] lto/113207 - fix free_lang_data_in_type

https://gcc.gnu.org/g:a55e14b239181381204c615335929b3316d75370 commit r15-7343-ga55e14b239181381204c615335929b3316d75370 Author: Richard Biener Date: Mon Feb 3 14:27:01 2025 +0100 lto/113207 - fix free_lang_data_in_type When we process function types we strip volatile and const qu

[gcc r15-7344] rtl-optimization/117611 - ICE in simplify_shift_const_1

https://gcc.gnu.org/g:5b46c01c50662a1c730e6658ea4307d4f80da578 commit r15-7344-g5b46c01c50662a1c730e6658ea4307d4f80da578 Author: Richard Biener Date: Mon Feb 3 15:27:30 2025 +0100 rtl-optimization/117611 - ICE in simplify_shift_const_1 The following checks we have a scalar int shi

[gcc r15-7352] c/118742 - gimple FE parsing of unary operators of C promoted args

https://gcc.gnu.org/g:4c8c9c9ee22bdb306439020a5a8fc5b4ee935f5b commit r15-7352-g4c8c9c9ee22bdb306439020a5a8fc5b4ee935f5b Author: Richard Biener Date: Tue Feb 4 10:54:48 2025 +0100 c/118742 - gimple FE parsing of unary operators of C promoted args The GIMPLE FE currently invokes pa

[gcc r15-7349] tree-optimization/117113 - ICE with unroll-and-jam

https://gcc.gnu.org/g:0675eb17480bada678bf2769d39732027abcd6d0 commit r15-7349-g0675eb17480bada678bf2769d39732027abcd6d0 Author: Richard Biener Date: Mon Feb 3 15:12:52 2025 +0100 tree-optimization/117113 - ICE with unroll-and-jam When there's an inner loop without virtual header

[gcc r15-7383] tree-optimization/118749 - bogus alignment peeling causes misaligned access

https://gcc.gnu.org/g:927e10bfce569947354cdd6b35c29b74e25c7816 commit r15-7383-g927e10bfce569947354cdd6b35c29b74e25c7816 Author: Richard Biener Date: Wed Feb 5 10:28:25 2025 +0100 tree-optimization/118749 - bogus alignment peeling causes misaligned access The vectorizer thinks it

[gcc r15-7384] rtl-optimization/117922 - disable fold-mem-offsets for highly connected CFG

https://gcc.gnu.org/g:677122c9df1b55a791a54426269f7a8ce794f947 commit r15-7384-g677122c9df1b55a791a54426269f7a8ce794f947 Author: Richard Biener Date: Wed Feb 5 13:17:47 2025 +0100 rtl-optimization/117922 - disable fold-mem-offsets for highly connected CFG The PR shows fold-mem-off

[gcc r15-7411] tree-optimization/115538 - possible wrong-code with SLP conversion

https://gcc.gnu.org/g:4931a637479aba35e35c50a86f58ecd6262bc487 commit r15-7411-g4931a637479aba35e35c50a86f58ecd6262bc487 Author: Richard Biener Date: Fri Feb 7 08:46:31 2025 +0100 tree-optimization/115538 - possible wrong-code with SLP conversion The following fixes a latent issue

[gcc r15-7425] jit/118780 - make sure to include dlfcn.h when plugin support is disabled

https://gcc.gnu.org/g:e22962538f64bb6e5ac87977ec8a5d86f4ef21cb commit r15-7425-ge22962538f64bb6e5ac87977ec8a5d86f4ef21cb Author: Richard Biener Date: Fri Feb 7 14:42:23 2025 +0100 jit/118780 - make sure to include dlfcn.h when plugin support is disabled The following makes the dlf

[gcc r15-7472] tree-optimization/118817 - missed folding of PRE inserted code

https://gcc.gnu.org/g:0a1d2ea57722c248777e1130de076e28c443ff8b commit r15-7472-g0a1d2ea57722c248777e1130de076e28c443ff8b Author: Richard Biener Date: Tue Feb 11 10:29:18 2025 +0100 tree-optimization/118817 - missed folding of PRE inserted code When PRE inserts code it is not fully

[gcc r15-7264] middle-end/118684 - fix fallout of wrong stack local alignment fix

https://gcc.gnu.org/g:7b02b8f65ef60be77f3f93945e2a6b463edaa0aa commit r15-7264-g7b02b8f65ef60be77f3f93945e2a6b463edaa0aa Author: Richard Biener Date: Wed Jan 29 08:58:39 2025 +0100 middle-end/118684 - fix fallout of wrong stack local alignment fix When we expand BIT_FIELD_REF we

[gcc r15-7287] middle-end/118695 - missed misalign handling in MEM_REF expansion

https://gcc.gnu.org/g:9ba2de71815c0ea6cc940ecb50af7cc1a84579f7 commit r15-7287-g9ba2de71815c0ea6cc940ecb50af7cc1a84579f7 Author: Richard Biener Date: Thu Jan 30 11:22:37 2025 +0100 middle-end/118695 - missed misalign handling in MEM_REF expansion When MEM_REF expansion of a non-ME

[gcc r15-6072] tree-optimization/117912 - bogus address equivalences for __builtin_object_size

https://gcc.gnu.org/g:233972ab3b5338d7a5d1d7af9108c1f366170e44 commit r15-6072-g233972ab3b5338d7a5d1d7af9108c1f366170e44 Author: Richard Biener Date: Thu Dec 5 10:47:13 2024 +0100 tree-optimization/117912 - bogus address equivalences for __builtin_object_size VN again is the culp

[gcc r15-6309] middle-end/118062 - bogus lowering of vector compares

https://gcc.gnu.org/g:cfe1ad3c488693a10fafb39d68c8cabc6e48daa7 commit r15-6309-gcfe1ad3c488693a10fafb39d68c8cabc6e48daa7 Author: Richard Biener Date: Tue Dec 17 11:23:02 2024 +0100 middle-end/118062 - bogus lowering of vector compares The generic expand_vector_piecewise routine su

[gcc r15-6830] More memory leak fixes

https://gcc.gnu.org/g:bedf26c201894d340a4eb61927879fd0d82e3102 commit r15-6830-gbedf26c201894d340a4eb61927879fd0d82e3102 Author: Richard Biener Date: Fri Jan 10 16:25:35 2025 +0100 More memory leak fixes The following were found compiling SPEC CPU 2017 with valgrind.

[gcc r15-7098] tree-optimization/118569 - LC SSA broken after unrolling

https://gcc.gnu.org/g:1d25e357c047f48cee50df318e965a0fbf84e2fe commit r15-7098-g1d25e357c047f48cee50df318e965a0fbf84e2fe Author: Richard Biener Date: Tue Jan 21 09:45:41 2025 +0100 tree-optimization/118569 - LC SSA broken after unrolling The following amends the previous fix to ma

[gcc r15-7149] tree-optimization/118558 - fix alignment compute with VMAT_CONTIGUOUS_REVERSE

https://gcc.gnu.org/g:7f1deb47a70ff804f0b2cce7be7e5fe8ba13 commit r15-7149-g7f1deb47a70ff804f0b2cce7be7e5fe8ba13 Author: Richard Biener Date: Tue Jan 21 14:58:43 2025 +0100 tree-optimization/118558 - fix alignment compute with VMAT_CONTIGUOUS_REVERSE There are calls to dr

<    1   2   3   4   5   6   7   8   9   10   >