[gcc r16-3082] Add subsections for vectorization and OpenMP/OpenACC target hooks

https://gcc.gnu.org/g:c9d0953ec88a4a4d2d949042ecb44deaeb65961f commit r16-3082-gc9d0953ec88a4a4d2d949042ecb44deaeb65961f Author: Richard Biener Date: Fri Aug 8 14:09:11 2025 +0200 Add subsections for vectorization and OpenMP/OpenACC target hooks Currently they are grouped under Ad

[gcc r16-3081] Remove setting of STMT_VINFO_VECTYPE on non-dataref stmts

https://gcc.gnu.org/g:8ee642035bf8c3bb62e1bf446761a8aa81b78814 commit r16-3081-g8ee642035bf8c3bb62e1bf446761a8aa81b78814 Author: Richard Biener Date: Thu Jul 10 11:37:14 2025 +0200 Remove setting of STMT_VINFO_VECTYPE on non-dataref stmts The following removes early setting of STM

[gcc r16-3078] tree-optimization/121454 - ICE building libgo

https://gcc.gnu.org/g:43949a5271b6c0b14076b736d0c609235d36f7df commit r16-3078-g43949a5271b6c0b14076b736d0c609235d36f7df Author: Richard Biener Date: Fri Aug 8 08:57:58 2025 +0200 tree-optimization/121454 - ICE building libgo The following avoids building BIT_FIELD_REFs of referen

[gcc r15-10202] Bump BASE-VER

https://gcc.gnu.org/g:47dd30079fd532f7c672870d61d3f2ea8cdb1771 commit r15-10202-g47dd30079fd532f7c672870d61d3f2ea8cdb1771 Author: Richard Biener Date: Fri Aug 8 09:32:56 2025 +0200 Bump BASE-VER * BASE-VER: Set to 15.2.1. Diff: --- gcc/BASE-VER | 2 +- 1 file changed, 1

[gcc r16-3076] Modernize vectorizable_lane_reducing

https://gcc.gnu.org/g:19102eab72ea5bdaa220564033f15a0f354d6254 commit r16-3076-g19102eab72ea5bdaa220564033f15a0f354d6254 Author: Richard Biener Date: Thu Aug 7 14:57:09 2025 +0200 Modernize vectorizable_lane_reducing The following avoids STMT_VINFO_VECTYPE usage in vect_is_emu

[gcc r16-3075] Pass SLP node to promotion/demotion costing

https://gcc.gnu.org/g:f38ebe459af792e097d7923dd49dc19d79d52427 commit r16-3075-gf38ebe459af792e097d7923dd49dc19d79d52427 Author: Richard Biener Date: Thu Aug 7 14:45:52 2025 +0200 Pass SLP node to promotion/demotion costing This one was forgotten. Also constants/externals are cos

[gcc r16-3074] Fix typo in getting SLP_TREE_VECTYPE

https://gcc.gnu.org/g:e4678c4cf843370ce00c46810dcf7f79f8ba7be9 commit r16-3074-ge4678c4cf843370ce00c46810dcf7f79f8ba7be9 Author: Richard Biener Date: Thu Aug 7 14:44:47 2025 +0200 Fix typo in getting SLP_TREE_VECTYPE It was applied on the stmt_info rather than the slp_node.

[gcc] Created tag 'releases/gcc-15.2.0'

The signed tag 'releases/gcc-15.2.0' was created pointing to: 5115c7e447fc... Update ChangeLog and version files for release Tagger: Richard Biener Date: Fri Aug 8 06:52:44 2025 + GCC 15.2.0 release

[gcc r16-3066] tree-optimization/121405 - missed VN with aggregate copy

https://gcc.gnu.org/g:53f491ccd1e59fad77fb2cb30d1a58b9e5e5f63c commit r16-3066-g53f491ccd1e59fad77fb2cb30d1a58b9e5e5f63c Author: Richard Biener Date: Wed Aug 6 12:31:13 2025 +0200 tree-optimization/121405 - missed VN with aggregate copy The following handles value-numbering of a B

[gcc r16-3042] Allow fully masked loops with legacy gather/scatter

https://gcc.gnu.org/g:50b1959cab19bdaf4f457f30e70a950ad9d35c6d commit r16-3042-g50b1959cab19bdaf4f457f30e70a950ad9d35c6d Author: Richard Biener Date: Tue Aug 5 14:55:18 2025 +0200 Allow fully masked loops with legacy gather/scatter The following removes the redundant check on supp

[gcc r16-3047] Simplify vect_supportable_dr_alignment

https://gcc.gnu.org/g:54edbeeaac6eb4865fab37374fbdff3a9a2f2e12 commit r16-3047-g54edbeeaac6eb4865fab37374fbdff3a9a2f2e12 Author: Richard Biener Date: Wed Aug 6 09:38:55 2025 +0200 Simplify vect_supportable_dr_alignment The following simplifies vect_supportable_dr_alignment based o

[gcc r16-3043] Use proper alias type for gather/scatter with SLP

https://gcc.gnu.org/g:3da5038ed58578d3373d5d5e371d6a15b172b201 commit r16-3043-g3da5038ed58578d3373d5d5e371d6a15b172b201 Author: Richard Biener Date: Wed Aug 6 09:41:26 2025 +0200 Use proper alias type for gather/scatter with SLP Both vectorizable_load and vectorizable_store compu

[gcc r16-3044] Record gather/scatter scale and base in the SLP tree

https://gcc.gnu.org/g:79e4f386ad836b7f9e1bdaf6aded4585035432ef commit r16-3044-g79e4f386ad836b7f9e1bdaf6aded4585035432ef Author: Richard Biener Date: Mon Aug 4 11:27:54 2025 +0200 Record gather/scatter scale and base in the SLP tree The main gather/scatter discovery happens at SLP

[gcc r16-2788] Remove hybrid SLP detection

https://gcc.gnu.org/g:df86ac52fccb2deccb53fb79f71db1fd700476bc commit r16-2788-gdf86ac52fccb2deccb53fb79f71db1fd700476bc Author: Richard Biener Date: Tue Aug 5 13:20:07 2025 +0200 Remove hybrid SLP detection The following removes hybrid SLP detection - it existed as sanity che

[gcc r16-2783] Prune some unnecessary PURE_SLP_STMT checks

https://gcc.gnu.org/g:11be9390a7f06e63179354df3fae1fd5d676abf8 commit r16-2783-g11be9390a7f06e63179354df3fae1fd5d676abf8 Author: Richard Biener Date: Mon Aug 4 14:30:28 2025 +0200 Prune some unnecessary PURE_SLP_STMT checks We now never have hybrid SLP (if we have, we fail immedia

[gcc r16-2785] tree-optimization/121395 - SLP of SIMD calls w/o LHS

https://gcc.gnu.org/g:32b1be7eb434addefe203249746dc6bbdc185403 commit r16-2785-g32b1be7eb434addefe203249746dc6bbdc185403 Author: Richard Biener Date: Mon Aug 4 14:45:53 2025 +0200 tree-optimization/121395 - SLP of SIMD calls w/o LHS The following records the alternate SLP instance

[gcc r16-2784] Rename loop_vect SLP_TYPE and clarify docs

https://gcc.gnu.org/g:9732b57443917fb913959740c471295205bcf2f5 commit r16-2784-g9732b57443917fb913959740c471295205bcf2f5 Author: Richard Biener Date: Mon Aug 4 14:36:03 2025 +0200 Rename loop_vect SLP_TYPE and clarify docs The following renames loop_vect to not_vect, removes the u

[gcc r16-2782] tree-optimization/121382 - avoid UB in IVOPTs inserted step computation

https://gcc.gnu.org/g:5d55cd95e2bbb1114a59ca743671188634e757cc commit r16-2782-g5d55cd95e2bbb1114a59ca743671188634e757cc Author: Richard Biener Date: Tue Aug 5 10:38:03 2025 +0200 tree-optimization/121382 - avoid UB in IVOPTs inserted step computation IVOPTs, when replacing an IV,

[gcc r16-2781] tree-optimization/121370 - avoid UB in building a CHREC

https://gcc.gnu.org/g:afafae097232e700bb7a74a453a048b83ebefccd commit r16-2781-gafafae097232e700bb7a74a453a048b83ebefccd Author: Richard Biener Date: Tue Aug 5 08:59:18 2025 +0200 tree-optimization/121370 - avoid UB in building a CHREC When there is obvious UB involved in the proc

[gcc r16-2748] Fail early when SLP instance build fails

https://gcc.gnu.org/g:446aac86a4bbdf38868c64410cb0e7823299f98a commit r16-2748-g446aac86a4bbdf38868c64410cb0e7823299f98a Author: Richard Biener Date: Mon Aug 4 13:29:16 2025 +0200 Fail early when SLP instance build fails The following makes us fail earlier when parts of the SLP bu

[gcc r16-2729] tree-optimization/121362 - missed FRE through aggregate copy

https://gcc.gnu.org/g:0d276cd378e7a41b9004577a30b9a8ca16ec6b4c commit r16-2729-g0d276cd378e7a41b9004577a30b9a8ca16ec6b4c Author: Richard Biener Date: Sun Aug 3 12:02:31 2025 +0200 tree-optimization/121362 - missed FRE through aggregate copy The following streamlines and generalize

[gcc r16-2720] tree-optimization/90242 - UBSAN error in vn_reference_compute_hash

https://gcc.gnu.org/g:7c99de1c439dbb149cc6a33a214f22f2df1c88f3 commit r16-2720-g7c99de1c439dbb149cc6a33a214f22f2df1c88f3 Author: Richard Biener Date: Fri Aug 1 15:07:15 2025 +0200 tree-optimization/90242 - UBSAN error in vn_reference_compute_hash The following plugs possible overf

[gcc r16-2710] Properly record SLP node when costing a vectorized store

https://gcc.gnu.org/g:39d429631ff7060238dc046db4340a304560227f commit r16-2710-g39d429631ff7060238dc046db4340a304560227f Author: Richard Biener Date: Fri Aug 1 13:27:00 2025 +0200 Properly record SLP node when costing a vectorized store Even when we emit scalar stores we should pa

[gcc r15-10189] arm: fully validate mem_noofs_operand [PR120351]

https://gcc.gnu.org/g:e14a895275d08da5a6bf87f5816d0a1ff6a13931 commit r15-10189-ge14a895275d08da5a6bf87f5816d0a1ff6a13931 Author: Richard Earnshaw Date: Mon May 19 16:19:39 2025 +0100 arm: fully validate mem_noofs_operand [PR120351] It's not enough to just check that a memory oper

[gcc r16-2709] Avoid representing SLP mask by scalar op

https://gcc.gnu.org/g:95aecacc8d9c40100f01433ba214103e756dc984 commit r16-2709-g95aecacc8d9c40100f01433ba214103e756dc984 Author: Richard Biener Date: Fri Aug 1 13:06:51 2025 +0200 Avoid representing SLP mask by scalar op The following removes the scalar mask output from vect_check

[gcc r16-2707] Merge get_group_load_store_type into get_load_store_type

https://gcc.gnu.org/g:86011bba93fc07f38c61b0fbebaea02d3d0d264f commit r16-2707-g86011bba93fc07f38c61b0fbebaea02d3d0d264f Author: Richard Biener Date: Fri Aug 1 11:02:43 2025 +0200 Merge get_group_load_store_type into get_load_store_type The following merges back get_group_load_sto

[gcc r16-2706] Some TLC to vectorizable_store

https://gcc.gnu.org/g:06cd2994cc81e94a38116ea7243181a99105d534 commit r16-2706-g06cd2994cc81e94a38116ea7243181a99105d534 Author: Richard Biener Date: Fri Aug 1 10:45:24 2025 +0200 Some TLC to vectorizable_store The following removes redundant checks and scalar operand uses.

[gcc r16-2705] Add VMAT_UNINITIALIZED

https://gcc.gnu.org/g:8308de8cc9d21318f69765918c2cc87761892e75 commit r16-2705-g8308de8cc9d21318f69765918c2cc87761892e75 Author: Richard Biener Date: Fri Aug 1 09:47:20 2025 +0200 Add VMAT_UNINITIALIZED We're using VMAT_INVARIANT as default, but we should simply have an uninit

[gcc r16-2704] tree-optimization/121338 - UBSAN error in adjust_setup_cost

https://gcc.gnu.org/g:7d66563e8152360a298dd1dcfe09491853db7f03 commit r16-2704-g7d66563e8152360a298dd1dcfe09491853db7f03 Author: Richard Biener Date: Fri Aug 1 08:55:11 2025 +0200 tree-optimization/121338 - UBSAN error in adjust_setup_cost The following avoids possibly overflowing

[gcc r15-10188] c: Fix crash in c-typeck.cc convert_arguments with indirect calls

https://gcc.gnu.org/g:5768bc24ca731943755ab937abc7617950e80037 commit r15-10188-g5768bc24ca731943755ab937abc7617950e80037 Author: Florian Weimer Date: Fri May 2 11:39:29 2025 +0200 c: Fix crash in c-typeck.cc convert_arguments with indirect calls gcc/c/ PR c/12005

[gcc r15-10187] c: Suppress -Wdeprecated-non-prototype warnings for builtins

https://gcc.gnu.org/g:3913d958e203dbf3c4a0527e7bc03c684e3f05a4 commit r15-10187-g3913d958e203dbf3c4a0527e7bc03c684e3f05a4 Author: Florian Weimer Date: Thu May 1 19:06:45 2025 +0200 c: Suppress -Wdeprecated-non-prototype warnings for builtins Builtins defined with BT_FN_INT_VAR etc

[gcc r16-2702] Use a class hierarchy for vect specific data

https://gcc.gnu.org/g:7a08afbc784ad41263e3cc3d039be631254183bc commit r16-2702-g7a08afbc784ad41263e3cc3d039be631254183bc Author: Richard Biener Date: Wed Jul 30 11:19:03 2025 +0200 Use a class hierarchy for vect specific data The following turns the union into a class hierarchy.

[gcc r16-2703] Put SLP_TREE_SIMD_CLONE_INFO into type specifc data

https://gcc.gnu.org/g:1586a8d707ad7be80184bf29165b4297a2f607d8 commit r16-2703-g1586a8d707ad7be80184bf29165b4297a2f607d8 Author: Richard Biener Date: Wed Jul 30 11:46:17 2025 +0200 Put SLP_TREE_SIMD_CLONE_INFO into type specifc data The following adds vect_simd_clone_data as a con

[gcc r15-10186] tree-optimization/121323 - UBSAN error in ao_ref_init_from_ptr_and_range

https://gcc.gnu.org/g:cb06e4a961b85a06e666f03124f9bd0f99e79f8c commit r15-10186-gcb06e4a961b85a06e666f03124f9bd0f99e79f8c Author: Richard Biener Date: Thu Jul 31 13:04:49 2025 +0200 tree-optimization/121323 - UBSAN error in ao_ref_init_from_ptr_and_range We should check the offset

[gcc r15-10185] tree-optimization/121320 - UBSAN error in ao_ref_init_from_vn_reference

https://gcc.gnu.org/g:3ef3de7a7329d5c76bed31e8b49a2a661c45f72f commit r15-10185-g3ef3de7a7329d5c76bed31e8b49a2a661c45f72f Author: Richard Biener Date: Thu Jul 31 13:06:36 2025 +0200 tree-optimization/121320 - UBSAN error in ao_ref_init_from_vn_reference The multiplication by BITS_

[gcc r15-10172] c++: fix ICE with [[deprecated]] [PR120756]

https://gcc.gnu.org/g:35be043e4d2d47a99ff81b0651bc57566b4d11f9 commit r15-10172-g35be043e4d2d47a99ff81b0651bc57566b4d11f9 Author: Marek Polacek Date: Wed Jun 25 10:29:49 2025 -0400 c++: fix ICE with [[deprecated]] [PR120756] Here we end up with "error reporting routines re-entered

[gcc r16-2681] Fixup wrong change to get_group_load_store_type

https://gcc.gnu.org/g:a4b5618782658ebe420c69ad939275d16ce76121 commit r16-2681-ga4b5618782658ebe420c69ad939275d16ce76121 Author: Richard Biener Date: Thu Jul 31 14:56:27 2025 +0200 Fixup wrong change to get_group_load_store_type The following fixes up the r16-2593-g6ac78317aa6adf

[gcc r16-2679] Extend gimple_fold_inplace API

https://gcc.gnu.org/g:00726defcd8ece9791e052f2aa21262be02a4380 commit r16-2679-g00726defcd8ece9791e052f2aa21262be02a4380 Author: Richard Biener Date: Thu Feb 20 11:45:06 2025 +0100 Extend gimple_fold_inplace API The following allows to specify the valueization hook to be used.

[gcc r16-2671] tree-optimization/121320 - UBSAN error in ao_ref_init_from_vn_reference

https://gcc.gnu.org/g:ff6f7d8e005ae94ffd55f1dba727e28531c3daf3 commit r16-2671-gff6f7d8e005ae94ffd55f1dba727e28531c3daf3 Author: Richard Biener Date: Thu Jul 31 13:06:36 2025 +0200 tree-optimization/121320 - UBSAN error in ao_ref_init_from_vn_reference The multiplication by BITS_P

[gcc r16-2670] tree-optimization/121323 - UBSAN error in ao_ref_init_from_ptr_and_range

https://gcc.gnu.org/g:0be8ffbc854410b65d45a72e843dc18b13fc7b0f commit r16-2670-g0be8ffbc854410b65d45a72e843dc18b13fc7b0f Author: Richard Biener Date: Thu Jul 31 13:04:49 2025 +0200 tree-optimization/121323 - UBSAN error in ao_ref_init_from_ptr_and_range We should check the offset

[gcc r15-10171] tree-optimization/121256 - re-instantiate check on SLP node

https://gcc.gnu.org/g:295dd134318068d253f09786b523718fd6951967 commit r15-10171-g295dd134318068d253f09786b523718fd6951967 Author: Richard Biener Date: Thu Jul 31 12:20:34 2025 +0200 tree-optimization/121256 - re-instantiate check on SLP node On the gcc-15 branch we can have no SLP

[gcc r16-2665] Add checks for node in aarch64 vector cost modeling

https://gcc.gnu.org/g:51c03570b3e65029fd4ac5d5ace5bddadec7ba9b commit r16-2665-g51c03570b3e65029fd4ac5d5ace5bddadec7ba9b Author: Richard Biener Date: Thu Jul 31 13:23:09 2025 +0200 Add checks for node in aarch64 vector cost modeling After removing STMT_VINFO_MEMORY_ACCESS_TYPE we

[gcc r16-2662] Do not bother with fake verifying of shared DRs

https://gcc.gnu.org/g:60b1938de008cc26a13210f1e5e49e23e64def3c commit r16-2662-g60b1938de008cc26a13210f1e5e49e23e64def3c Author: Richard Biener Date: Thu Jul 31 09:26:16 2025 +0200 Do not bother with fake verifying of shared DRs The following avoids comparing the shared DRs agains

[gcc r16-2663] Remove STMT_VINFO_MEMORY_ACCESS_TYPE

https://gcc.gnu.org/g:a5aa73e547d7383bcefa79704f072800ec94340b commit r16-2663-ga5aa73e547d7383bcefa79704f072800ec94340b Author: Richard Biener Date: Thu Jul 24 15:20:02 2025 +0200 Remove STMT_VINFO_MEMORY_ACCESS_TYPE This should be present only on SLP nodes now. The RISC-V chang

[gcc r15-10168] Fix ICE in speculative devirtualization

https://gcc.gnu.org/g:1b42af2d10bd6ec11ffa25431445cd609384e690 commit r15-10168-g1b42af2d10bd6ec11ffa25431445cd609384e690 Author: Jan Hubicka Date: Fri Jul 11 12:37:24 2025 +0200 Fix ICE in speculative devirtualization This patch fixes ICE bilding lto1 with autoprofiledbootstrap a

[gcc r15-10167] tree-optimization/121256 - properly support SLP in vectorizable recurrence

https://gcc.gnu.org/g:e442d6f4964daec3c9dfeb52101d9f80ff363f27 commit r15-10167-ge442d6f4964daec3c9dfeb52101d9f80ff363f27 Author: Richard Biener Date: Sun Jul 27 18:42:25 2025 +0200 tree-optimization/121256 - properly support SLP in vectorizable recurrence We failed to build the c

[gcc r15-10166] tree-optimization/121130 - vectorizable_call cannot handle .MASK_CALL

https://gcc.gnu.org/g:5e9d3957c6fea548c9c4a8504261e669add9512a commit r15-10166-g5e9d3957c6fea548c9c4a8504261e669add9512a Author: Richard Biener Date: Wed Jul 30 12:34:20 2025 +0200 tree-optimization/121130 - vectorizable_call cannot handle .MASK_CALL The following makes it correc

[gcc r16-2658] Disallow scan-store vectorization in epilogues

https://gcc.gnu.org/g:8cae7cc099f533b2882898dc05528fcbb70ea9f2 commit r16-2658-g8cae7cc099f533b2882898dc05528fcbb70ea9f2 Author: Richard Biener Date: Wed Jul 30 15:05:19 2025 +0200 Disallow scan-store vectorization in epilogues The following disallows vectorizing epilogues contain

[gcc r16-2657] Avoid passing vectype != NULL when costing scalar IL

https://gcc.gnu.org/g:fb59c5719c17a04ecfd58b5e566eccd6d2ac583a commit r16-2657-gfb59c5719c17a04ecfd58b5e566eccd6d2ac583a Author: Richard Biener Date: Tue Jul 29 09:20:42 2025 +0200 Avoid passing vectype != NULL when costing scalar IL The following makes sure to not leak a set vect

[gcc r16-2645] [x86] factor out worker from ix86_builtin_vectorization_cost

https://gcc.gnu.org/g:e9ad533dd2ea233c41d41928e8d71a824e09eaf9 commit r16-2645-ge9ad533dd2ea233c41d41928e8d71a824e09eaf9 Author: Richard Biener Date: Wed Jul 30 13:01:18 2025 +0200 [x86] factor out worker from ix86_builtin_vectorization_cost The following factors out a worker that

[gcc r16-2636] tree-optimization/121130 - vectorizable_call cannot handle .MASK_CALL

https://gcc.gnu.org/g:a0d2de441ea0898e508c36567fbb77f161afa3c8 commit r16-2636-ga0d2de441ea0898e508c36567fbb77f161afa3c8 Author: Richard Biener Date: Wed Jul 30 12:34:20 2025 +0200 tree-optimization/121130 - vectorizable_call cannot handle .MASK_CALL The following makes it correct

[gcc r16-2630] tree-optimization/121304 - set memory_access_type before reading it

https://gcc.gnu.org/g:10a5762984965fb849f91f1227b03b8bba149bb5 commit r16-2630-g10a5762984965fb849f91f1227b03b8bba149bb5 Author: Richard Biener Date: Wed Jul 30 09:44:07 2025 +0200 tree-optimization/121304 - set memory_access_type before reading it The following re-orders gather/s

[gcc r16-2603] Another testcase for PR120687

https://gcc.gnu.org/g:9c48cbb259851873397abc632422f3f44f253db0 commit r16-2603-g9c48cbb259851873397abc632422f3f44f253db0 Author: Richard Biener Date: Tue Jul 29 14:01:46 2025 +0200 Another testcase for PR120687 This shows reassoc is harmful even with len == 3. PR

[gcc r16-2601] tree-optimization/120687 - avoid disturbing reduction chains in reassoc

https://gcc.gnu.org/g:e8a51144c02e1cf210db5763e435802ac6fa6ad9 commit r16-2601-ge8a51144c02e1cf210db5763e435802ac6fa6ad9 Author: Richard Biener Date: Tue Jul 29 10:05:32 2025 +0200 tree-optimization/120687 - avoid disturbing reduction chains in reassoc Reassoc carefully ranks oper

[gcc r16-2594] Un-factor vectorizable_load parts

https://gcc.gnu.org/g:a1ad36f5a899e3d4c2c97f4f68d283bef3d386b1 commit r16-2594-ga1ad36f5a899e3d4c2c97f4f68d283bef3d386b1 Author: Richard Biener Date: Mon Jul 28 15:04:01 2025 +0200 Un-factor vectorizable_load parts When the costing refactoring happened we ended up with some strang

[gcc r16-2593] Eliminate gather-scatter-info offset_dt member

https://gcc.gnu.org/g:6ac78317aa6adf285568974afae5f22e4a309c97 commit r16-2593-g6ac78317aa6adf285568974afae5f22e4a309c97 Author: Richard Biener Date: Mon Jul 28 14:22:44 2025 +0200 Eliminate gather-scatter-info offset_dt member The following removes this only set member. Sligthly

[gcc r16-2582] Const correctness for gather-scatter info

https://gcc.gnu.org/g:fda0a873df12a46ff4cdc89229b0de6179b465c1 commit r16-2582-gfda0a873df12a46ff4cdc89229b0de6179b465c1 Author: Richard Biener Date: Mon Jul 28 13:46:18 2025 +0200 Const correctness for gather-scatter info The following adds const qualification to gather_scatter_i

[gcc r16-2578] tree-optimization/121256 - properly support SLP in vectorizable recurrence

https://gcc.gnu.org/g:26dc9aa285b53551c55d3d660bb6da21d59d7023 commit r16-2578-g26dc9aa285b53551c55d3d660bb6da21d59d7023 Author: Richard Biener Date: Sun Jul 27 18:42:25 2025 +0200 tree-optimization/121256 - properly support SLP in vectorizable recurrence We failed to build the co

[gcc r16-2576] Move STMT_VINFO_TYPE to SLP_TREE_TYPE

https://gcc.gnu.org/g:973456e69123be203386baa3f7806d7bd38a9cd4 commit r16-2576-g973456e69123be203386baa3f7806d7bd38a9cd4 Author: Richard Biener Date: Sun Jul 27 18:42:18 2025 +0200 Move STMT_VINFO_TYPE to SLP_TREE_TYPE I am at a point where I want to store additional information f

[gcc r16-2518] Remove now redundant vect_get_vec_defs overload

https://gcc.gnu.org/g:cebf2a85d2de9aea3fa8ba0d8dc45945e56f0eb9 commit r16-2518-gcebf2a85d2de9aea3fa8ba0d8dc45945e56f0eb9 Author: Richard Biener Date: Fri Jul 25 15:03:36 2025 +0200 Remove now redundant vect_get_vec_defs overload The following removes the vect_get_vec_defs overload

[gcc r16-2517] Tidy vect_is_simple_use API for SLP only

https://gcc.gnu.org/g:91eb409242218d5bb66fb32cba7dad3c8ab77814 commit r16-2517-g91eb409242218d5bb66fb32cba7dad3c8ab77814 Author: Richard Biener Date: Fri Jul 25 14:04:50 2025 +0200 Tidy vect_is_simple_use API for SLP only The following removes one vect_is_simple_use overload that

[gcc r16-2516] Fix and simplify vect_model_simple_cost

https://gcc.gnu.org/g:e38c88d5deb4fd4ef103b63691d7be028a54c3dc commit r16-2516-ge38c88d5deb4fd4ef103b63691d7be028a54c3dc Author: Richard Biener Date: Fri Jul 25 13:38:48 2025 +0200 Fix and simplify vect_model_simple_cost There's no longer any call without SLP node so we don't need

[gcc r16-2513] Reduce the number of STMT_VINFO_VECTYPE uses

https://gcc.gnu.org/g:fe3069ae2386ca48614dc627accec20b3fcfe722 commit r16-2513-gfe3069ae2386ca48614dc627accec20b3fcfe722 Author: Richard Biener Date: Fri Jul 25 12:34:44 2025 +0200 Reduce the number of STMT_VINFO_VECTYPE uses The following adjusts uses of STMT_VINFO_VECTYPE to loo

[gcc r16-2512] Remove unused stmt_vectype

https://gcc.gnu.org/g:c2bb709a9cd580309016ed95b6e2abc9cb7a26e8 commit r16-2512-gc2bb709a9cd580309016ed95b6e2abc9cb7a26e8 Author: Richard Biener Date: Fri Jul 25 11:45:17 2025 +0200 Remove unused stmt_vectype The following removes an unused accessor to data that should go away.

[gcc r16-2509] Remove STMT_VINFO_VEC_STMTS

https://gcc.gnu.org/g:22e5711c66a7319cd4bf28346b3b3dcda130ed75 commit r16-2509-g22e5711c66a7319cd4bf28346b3b3dcda130ed75 Author: Richard Biener Date: Fri Jul 25 09:50:18 2025 +0200 Remove STMT_VINFO_VEC_STMTS The following removes the last uses of STMT_VINFO_VEC_STMTS and the

[gcc r16-2508] Remove load interleaving code

https://gcc.gnu.org/g:735b470346bcbd2947531dc1065ee1f4f4ede050 commit r16-2508-g735b470346bcbd2947531dc1065ee1f4f4ede050 Author: Richard Biener Date: Fri Jul 25 09:34:32 2025 +0200 Remove load interleaving code The following removes the non-SLP load interleaving code which was

[gcc r16-2507] Remove store interleaving support

https://gcc.gnu.org/g:fcadd6d32398271db9a4935d51a0066648962674 commit r16-2507-gfcadd6d32398271db9a4935d51a0066648962674 Author: Richard Biener Date: Fri Jul 25 09:40:27 2025 +0200 Remove store interleaving support The following removes the non-SLP store interleaving support which

[gcc r16-2506] Remove vect_get_vec_defs_for_operand

https://gcc.gnu.org/g:727276ab918449cc82f22685ddc8fb5d559b8f33 commit r16-2506-g727276ab918449cc82f22685ddc8fb5d559b8f33 Author: Richard Biener Date: Fri Jul 25 09:19:47 2025 +0200 Remove vect_get_vec_defs_for_operand This removes vect_get_vec_defs_for_operand and its remaining us

[gcc r16-2505] Remove VMAT_CONTIGUOUS_PERMUTE

https://gcc.gnu.org/g:85d081b13fcc03c9403d175f1fac8b42aa018566 commit r16-2505-g85d081b13fcc03c9403d175f1fac8b42aa018566 Author: Richard Biener Date: Fri Jul 25 09:09:35 2025 +0200 Remove VMAT_CONTIGUOUS_PERMUTE This VMAT was used for interleaving which was non-SLP only. The

[gcc r16-2504] Remove dead code from vectorizable_store

https://gcc.gnu.org/g:fab1cd65070d9374cda161a881c4f2f9be38ee8a commit r16-2504-gfab1cd65070d9374cda161a881c4f2f9be38ee8a Author: Richard Biener Date: Fri Jul 25 09:04:49 2025 +0200 Remove dead code from vectorizable_store There's dead code in the else block of a if (!costing_p) bl

[gcc r16-2488] Remove vec_stmt from vectorizable_* API

https://gcc.gnu.org/g:5865c0b81fe8c3046da3ec7c6cd9a38aa2fd709d commit r16-2488-g5865c0b81fe8c3046da3ec7c6cd9a38aa2fd709d Author: Richard Biener Date: Thu Jul 24 14:14:24 2025 +0200 Remove vec_stmt from vectorizable_* API The following removes the non-SLP gimple **vec_stmt argument

[gcc r16-2486] Remove non-SLP path from vectorizable_early_exit

https://gcc.gnu.org/g:292a9bee46db91a1b6368c1d1fd1b43a529489ed commit r16-2486-g292a9bee46db91a1b6368c1d1fd1b43a529489ed Author: Richard Biener Date: Thu Jul 24 13:47:14 2025 +0200 Remove non-SLP path from vectorizable_early_exit This removes the non-SLP path from vectorizable_ear

[gcc r16-2485] Remove non-SLP path from vectorizable_simd_clone_call

https://gcc.gnu.org/g:ccd3ed47764a11e0ef4ed1284c8ee41258dd5337 commit r16-2485-gccd3ed47764a11e0ef4ed1284c8ee41258dd5337 Author: Richard Biener Date: Thu Jul 24 12:46:01 2025 +0200 Remove non-SLP path from vectorizable_simd_clone_call This removes the non-SLP path from vectorizabl

[gcc r16-2476] Remove non-SLP path from vectorizable_comparison

https://gcc.gnu.org/g:0152a18509369a949eaf118f9e9abd5f239167ac commit r16-2476-g0152a18509369a949eaf118f9e9abd5f239167ac Author: Richard Biener Date: Thu Jul 24 13:30:32 2025 +0200 Remove non-SLP path from vectorizable_comparison This removes the non-SLP path from vectorizable_com

[gcc r16-2475] Remove non-SLP path from vectorizable_condition

https://gcc.gnu.org/g:febfd99cacc0608d52688d626b64b47b470d commit r16-2475-gfebfd99cacc0608d52688d626b64b47b470d Author: Richard Biener Date: Thu Jul 24 13:24:28 2025 +0200 Remove non-SLP path from vectorizable_condition The following removes the non-SLP paths from vectori

[gcc r16-2474] Remove non-SLP path from vectorizable_scan_store

https://gcc.gnu.org/g:e2019172dcf748d4880e1cc85f2d92b905d3ad6e commit r16-2474-ge2019172dcf748d4880e1cc85f2d92b905d3ad6e Author: Richard Biener Date: Thu Jul 24 13:19:20 2025 +0200 Remove non-SLP path from vectorizable_scan_store This removes the non-SLP paths from vectorizable_sc

[gcc r16-2473] Remove non-SLP path from vectorizable_shift

https://gcc.gnu.org/g:70f6c2aa3117b41c516f81b376d73b358156a867 commit r16-2473-g70f6c2aa3117b41c516f81b376d73b358156a867 Author: Richard Biener Date: Thu Jul 24 13:11:05 2025 +0200 Remove non-SLP path from vectorizable_shift This removes the non-SLP paths from vectorizable_shift.

[gcc r16-2472] Remove non-SLP path from vectorizable_assignment

https://gcc.gnu.org/g:8f9373c80572595fd1cb3df61b8301cf1647ccff commit r16-2472-g8f9373c80572595fd1cb3df61b8301cf1647ccff Author: Richard Biener Date: Thu Jul 24 13:07:33 2025 +0200 Remove non-SLP path from vectorizable_assignment This removes the non-SLP paths from vectorizable_as

[gcc r16-2470] Remove non-SLP path from vectorizable_bswap

https://gcc.gnu.org/g:a9394c914739becd6239c8aaa277c9b8e32a4e5e commit r16-2470-ga9394c914739becd6239c8aaa277c9b8e32a4e5e Author: Richard Biener Date: Thu Jul 24 11:34:45 2025 +0200 Remove non-SLP path from vectorizable_bswap * tree-vect-stmts.cc (vectorizable_bswap): Remov

[gcc r16-2469] Remove non-SLP path from vectorizable_recurr

https://gcc.gnu.org/g:6a0983c3db7c2b048d91171306c34ecc1100154e commit r16-2469-g6a0983c3db7c2b048d91171306c34ecc1100154e Author: Richard Biener Date: Thu Jul 24 11:17:57 2025 +0200 Remove non-SLP path from vectorizable_recurr This removes the non-SLP paths from vectorizable_recurr

[gcc r16-2471] Remove non-SLP path from vectorizable_call

https://gcc.gnu.org/g:c69a34add852a76020feb1b64ba04bfaeb0b5e16 commit r16-2471-gc69a34add852a76020feb1b64ba04bfaeb0b5e16 Author: Richard Biener Date: Thu Jul 24 11:39:11 2025 +0200 Remove non-SLP path from vectorizable_call This removes the non-SLP paths from vectorizable_call, pr

[gcc r15-10061] tree-optimization/121202 - fix vector stmt placement

https://gcc.gnu.org/g:4d0c17354e42d57d3f94a7b66ab59299b076ee68 commit r15-10061-g4d0c17354e42d57d3f94a7b66ab59299b076ee68 Author: Richard Biener Date: Tue Jul 22 13:02:03 2025 +0200 tree-optimization/121202 - fix vector stmt placement When we have a vector shift with a scalar the

[gcc r16-2443] tree-optimization/121220 - improve sinking of stores

https://gcc.gnu.org/g:44b24b62a44e6208d654513cf567dd6900a2fb56 commit r16-2443-g44b24b62a44e6208d654513cf567dd6900a2fb56 Author: Richard Biener Date: Wed Jul 23 09:40:24 2025 +0200 tree-optimization/121220 - improve sinking of stores We currently do only very restricted store sink

[gcc r16-2437] [aarch64] check for non-NULL vectype in aarch64_vector_costs::add_stmt_cost

https://gcc.gnu.org/g:9a8d91219d65f7cb304b1d678957f95061bfe02b commit r16-2437-g9a8d91219d65f7cb304b1d678957f95061bfe02b Author: Richard Biener Date: Tue Jul 22 15:04:16 2025 +0200 [aarch64] check for non-NULL vectype in aarch64_vector_costs::add_stmt_cost With a patch still in de

[gcc r16-2436] middle-end/121216 - ICE with VLA const string initializer

https://gcc.gnu.org/g:6acf9501771b8a26643fe6b887eb2d9b6d008b47 commit r16-2436-g6acf9501771b8a26643fe6b887eb2d9b6d008b47 Author: Richard Biener Date: Tue Jul 22 15:41:20 2025 +0200 middle-end/121216 - ICE with VLA const string initializer constant_byte_string fails to consider the

[gcc r16-2422] tree-optimization/121202 - fix vector stmt placement

https://gcc.gnu.org/g:bdfb5cc5aa6959a6959fc0cf98da08db89c81032 commit r16-2422-gbdfb5cc5aa6959a6959fc0cf98da08db89c81032 Author: Richard Biener Date: Tue Jul 22 13:02:03 2025 +0200 tree-optimization/121202 - fix vector stmt placement When we have a vector shift with a scalar the s

[gcc r16-2397] Fix gcc.dg/vect/slp-28.c

https://gcc.gnu.org/g:7698f886e1f61335eeb7af2e68fb6fb1c07a7256 commit r16-2397-g7698f886e1f61335eeb7af2e68fb6fb1c07a7256 Author: Richard Biener Date: Tue Jul 22 08:44:25 2025 +0200 Fix gcc.dg/vect/slp-28.c gcc.dg/vect/slp-28.c is now vectorized as expected even on targets with

[gcc r16-2381] tree-optimization/121194 - check LC PHIs can be vectorized

https://gcc.gnu.org/g:8ca85e95daca7f7c56d297fa7f2328e5385d446c commit r16-2381-g8ca85e95daca7f7c56d297fa7f2328e5385d446c Author: Richard Biener Date: Mon Jul 21 12:48:45 2025 +0200 tree-optimization/121194 - check LC PHIs can be vectorized With bools we can have the usual mismatch

[gcc r16-2373] Remove bougs minimum VF compute

https://gcc.gnu.org/g:1eb17c6b20e662b4aa4b33e7e84ceccf6b6c41d8 commit r16-2373-g1eb17c6b20e662b4aa4b33e7e84ceccf6b6c41d8 Author: Richard Biener Date: Mon Jul 21 10:40:13 2025 +0200 Remove bougs minimum VF compute The following removes the minimum VF compute from dataref analysis

[gcc r16-2346] tree-optimization/121126 - properly verify live LC PHIs

https://gcc.gnu.org/g:8bb7234a86f3432658f9fa29d7957d1d0c929832 commit r16-2346-g8bb7234a86f3432658f9fa29d7957d1d0c929832 Author: Richard Biener Date: Fri Jul 18 08:28:59 2025 +0200 tree-optimization/121126 - properly verify live LC PHIs The following makes sure we analyze live LC

[gcc r15-10004] tree-optimization/120924 - up --param uninit-max-chain-len

https://gcc.gnu.org/g:e1828db8e75a56eebf8ede53e1ae4fcbd0f3376f commit r15-10004-ge1828db8e75a56eebf8ede53e1ae4fcbd0f3376f Author: Richard Biener Date: Fri Jul 18 09:02:09 2025 +0200 tree-optimization/120924 - up --param uninit-max-chain-len The PR shows that the uninit analysis li

[gcc r16-2342] Remove non-SLP path from vectorizable_live_operation

https://gcc.gnu.org/g:9892d6c4d757c6bf84f0964ff078a9d9f7e9e7b5 commit r16-2342-g9892d6c4d757c6bf84f0964ff078a9d9f7e9e7b5 Author: Richard Biener Date: Fri Jul 18 09:37:01 2025 +0200 Remove non-SLP path from vectorizable_live_operation This removes paths gated by !slp_node and propa

[gcc r15-10002] tree-optimization/121059 - fixup loop mask query

https://gcc.gnu.org/g:b37d0b2f6203ec3978459bc4cbce9bd9aa110025 commit r15-10002-gb37d0b2f6203ec3978459bc4cbce9bd9aa110025 Author: Richard Biener Date: Mon Jul 14 14:09:28 2025 +0200 tree-optimization/121059 - fixup loop mask query When we opportunistically mask an operand of a AND

[gcc r15-10001] tree-optimization/121049 - avoid loop masking with even/odd reduction

https://gcc.gnu.org/g:185a8bae0f2ce46c65ec6a541c2fe7299df6d7f4 commit r15-10001-g185a8bae0f2ce46c65ec6a541c2fe7299df6d7f4 Author: Richard Biener Date: Wed Jul 16 15:07:58 2025 +0200 tree-optimization/121049 - avoid loop masking with even/odd reduction The following disables loop m

[gcc r15-10000] tree-optimization/121035 - handle stray VN values without expression

https://gcc.gnu.org/g:aa28a331e9f7c526540184d4b748198b8e5ad338 commit r15-1-gaa28a331e9f7c526540184d4b748198b8e5ad338 Author: Richard Biener Date: Wed Jul 16 20:19:44 2025 +0200 tree-optimization/121035 - handle stray VN values without expression When VN iterates we can end up

[gcc r16-2341] tree-optimization/120924 - up --param uninit-max-chain-len

https://gcc.gnu.org/g:cf9a479e3f909d5217e954788eb3c5b569e4bc52 commit r16-2341-gcf9a479e3f909d5217e954788eb3c5b569e4bc52 Author: Richard Biener Date: Fri Jul 18 09:02:09 2025 +0200 tree-optimization/120924 - up --param uninit-max-chain-len The PR shows that the uninit analysis lim

[gcc r16-2333] tree-optimization/121048 - move check for only having vector(1)

https://gcc.gnu.org/g:175b49933ffeea6d4a690e40bba006854977f43f commit r16-2333-g175b49933ffeea6d4a690e40bba006854977f43f Author: Richard Biener Date: Thu Jul 17 08:51:11 2025 +0200 tree-optimization/121048 - move check for only having vector(1) The following moves rejecting loop v

[gcc r16-2312] Reject single lane vector types for SLP build

https://gcc.gnu.org/g:246ecf81612518196f4aa45fe96f85831fe408ff commit r16-2312-g246ecf81612518196f4aa45fe96f85831fe408ff Author: Richard Biener Date: Thu Jul 10 13:04:00 2025 +0200 Reject single lane vector types for SLP build The following makes us never consider vector(1) T type

[gcc r16-2311] tree-optimization/121035 - handle stray VN values without expression

https://gcc.gnu.org/g:9af57c471087a3a1b87621bce1208d6c77ba2a4a commit r16-2311-g9af57c471087a3a1b87621bce1208d6c77ba2a4a Author: Richard Biener Date: Wed Jul 16 20:19:44 2025 +0200 tree-optimization/121035 - handle stray VN values without expression When VN iterates we can end up

[gcc r16-2299] tree-optimization/121049 - avoid loop masking with even/odd reduction

https://gcc.gnu.org/g:bc5570f7ef796fa7f5ab89b34ed9de2be5299f0e commit r16-2299-gbc5570f7ef796fa7f5ab89b34ed9de2be5299f0e Author: Richard Biener Date: Wed Jul 16 15:07:58 2025 +0200 tree-optimization/121049 - avoid loop masking with even/odd reduction The following disables loop ma

  1   2   3   4   5   6   7   8   9   10   >