pre-commit: PR157614

[zyw-bot]

Link: llvm/llvm-project#157614
Requested by: @andjo403

[zyw-bot]

Diff mode

runner: ariselab-64c-docker
baseline: llvm/llvm-project@7fbfd1c
patch: llvm/llvm-project#157614
sha256: d896e7143c74f67c97c0dac5bac18437889e87e9fad762ba307617208043d8ea
commit: 27c6665

890 files changed, 356758 insertions(+), 361200 deletions(-)

Improvements:
  correlated-value-propagation.NumMinMax 12268 -> 12482 +1.74%
  correlated-value-propagation.NumUDivURemsNarrowed 12862 -> 12945 +0.65%
  correlated-value-propagation.NumAShrsConverted 3634 -> 3652 +0.50%
  correlated-value-propagation.NumMulNUW 61050 -> 61329 +0.46%
  correlated-value-propagation.NumSICmps 63257 -> 63509 +0.40%
  correlated-value-propagation.NumSaturating 2637 -> 2647 +0.38%
  correlated-value-propagation.NumAnd 46599 -> 46763 +0.35%
  correlated-value-propagation.NumSubNUW 37495 -> 37587 +0.25%
  correlated-value-propagation.NumMulNW 138043 -> 138360 +0.23%
  correlated-value-propagation.NumSExt 48030 -> 48129 +0.21%
Regressions:
  licm.NumBOAssociationsHoisted 3597 -> 3595 -0.06%
  simplifycfg.NumLookupTablesHoles 2611 -> 2610 -0.04%
  simplifycfg.NumInvokeSetsFormed 54668 -> 54660 -0.01%
  constmerge.NumIdenticalMerged 15584 -> 15582 -0.01%
  simplifycfg.NumInvokesMerged 148495 -> 148479 -0.01%
  aggressive-instcombine.NumExprsReduced 20572 -> 20570 -0.01%
  memdep.NumCacheNonLocal 21245 -> 21243 -0.01%
  aggressive-instcombine.NumInstrsReduced 68521 -> 68516 -0.01%
  licm.NumSunk 276284 -> 276268 -0.01%
  licm.NumMovedCalls 35440 -> 35438 -0.01%

7 12 bench/abc/optimized/cecSatG2.ll
23 21 bench/abc/optimized/dauDsd.ll
1 2 bench/abc/optimized/dsdTree.ll
2 2 bench/abseil-cpp/optimized/charconv_bigint.ll
60 78 bench/actix-rs/optimized/1rmq6g519rw6hl6a.ll
68 64 bench/actix-rs/optimized/3q3jkuswkinqn2en.ll
4 6 bench/actix-rs/optimized/comsm606o4zjj7a.ll
17 19 bench/actix-rs/optimized/kyfmdxbalxu1s84.ll
1 1 bench/arrow/optimized/fixed-dtoa.ll
3 2 bench/assimp/optimized/FBXConverter.ll
25 22 bench/bullet3/optimized/btDantzigLCP.ll
33 34 bench/ceres/optimized/problem_impl.ll
12 15 bench/cmake/optimized/archive_read_support_format_7zip.ll
13 18 bench/cmake/optimized/archive_read_support_format_rar.ll
10 10 bench/coreutils-rs/optimized/2cs7fkuvya39n8q8.ll
10 12 bench/coreutils-rs/optimized/nchybjntbm98czw.ll
34 44 bench/curl/optimized/http.ll
12 10 bench/darktable/optimized/decoders_dcraw.ll
18 19 bench/darktable/optimized/introspection_filmicrgb.ll
8 8 bench/delta-rs/optimized/3qkwqfk85qralejq.ll
17 19 bench/diesel-rs/optimized/1mcix0ravw4ybg9k.ll
17 26 bench/duckdb/optimized/format.ll
71 72 bench/duckdb/optimized/prog.ll
6 5 bench/eastl/optimized/TestHeap.ll
1 5 bench/faiss/optimized/IndexLSH.ll
3 5 bench/ffmpeg/optimized/aliaspixdec.ll
1 5 bench/ffmpeg/optimized/parse.ll
42 40 bench/ffmpeg/optimized/pp_bnk.ll
13 10 bench/ffmpeg/optimized/vf_blend.ll
5 5 bench/flatbuffers/optimized/binary_annotator.ll
6 5 bench/fmt/optimized/color-test.ll
7 7 bench/harfbuzz/optimized/hb-subset-plan.ll
6 7 bench/hdf5/optimized/h5tools_dump.ll
31 55 bench/hermes/optimized/APFloat.ll
4 4 bench/hyperscan/optimized/noodle_engine.ll
7 8 bench/icu/optimized/collationfastlatin.ll
5 8 bench/icu/optimized/number_affixutils.ll
9 11 bench/icu/optimized/uiter.ll
2 2 bench/influxdb-rs/optimized/2i5qqg1f0a0eo8eq.ll
5 5 bench/jiff-rs/optimized/diij60a8h6gwjom97cm6sfirq.ll
3 3 bench/jsonnet/optimized/rapidyaml.ll
10 12 bench/just-rs/optimized/53slus9exfz9w045.ll
4 3 bench/lean4/optimized/interrupt.ll
6 10 bench/libdeflate/optimized/prog_util.ll
6 4 bench/linux/optimized/decompress_unlzma.ll
44 40 bench/linux/optimized/entropy_common.ll
6 4 bench/linux/optimized/intel_vdsc.ll
18 28 bench/linux/optimized/keyctl.ll
12 15 bench/llvm/optimized/AnalysisBasedWarnings.ll
6 5 bench/llvm/optimized/CodeGenInstruction.ll
16 24 bench/llvm/optimized/InlineOrder.ll
4 8 bench/llvm/optimized/LegalizerHelper.ll
12 10 bench/llvm/optimized/RISCVMatInt.ll
8 9 bench/llvm/optimized/SelectionDAG.ll
2 6 bench/luajit/optimized/lj_ffrecord.ll
8 8 bench/lvgl/optimized/lv_text.ll
16 20 bench/meshlab/optimized/gltf_loader.ll
4 3 bench/minetest/optimized/mapgen_v6.ll
24 28 bench/mixbox/optimized/mixbox.ll
9 9 bench/mold/optimized/rust-demangle.ll
8 12 bench/ockam-rs/optimized/1gjay75gp38o8e4p.ll
13 21 bench/ockam-rs/optimized/3trfdpf6q4y0wtix.ll
15 19 bench/open3d/optimized/FileGLTF.ll
11 19 bench/openexr/optimized/pack.ll
8 12 bench/openjdk/optimized/abstractInterpreter.ll
10 14 bench/openjdk/optimized/ciEnv.ll
12 16 bench/openjdk/optimized/instanceKlass.ll
12 13 bench/openssl/optimized/asn1_lib.ll
10 9 bench/openssl/optimized/bss_mem.ll
10 13 bench/openusd/optimized/bignum.ll
21 25 bench/openusd/optimized/stbImage.ll
2 3 bench/php/optimized/ir_sccp.ll
12 12 bench/php/optimized/pcre2_jit_compile.ll
37 43 bench/pingora-rs/optimized/22g42cy0ag75yw3gv725oc340.ll
10 11 bench/pingora-rs/optimized/6qou06qwl5au54knfo027km8l.ll
16 20 bench/pola-rs/optimized/02z5h4hjkw80zz1tgff2r8tqu.ll
16 20 bench/pola-rs/optimized/3j6s1s8bkufqpw2g2hh1cvgea.ll
3 6 bench/postgres/optimized/mcv.ll
8 6 bench/postgres/optimized/numutils.ll
22 23 bench/postgres/optimized/tablecmds.ll
9 12 bench/protobuf/optimized/field_mask_util.ll
15 19 bench/raylib/optimized/rmodels.ll
9 13 bench/raylib/optimized/rtextures.ll
29 27 bench/redis/optimized/networking.ll
59 70 bench/rocksdb/optimized/slice.ll
18 17 bench/ruby/optimized/parse.ll
12 10 bench/ruby/optimized/raddrinfo.ll
23 28 bench/ruby/optimized/strscan.ll
20 18 bench/rust-analyzer-rs/optimized/1opoiu8yzxku2bb7.ll
9 16 bench/rust-analyzer-rs/optimized/2qyzpr7jeax5a9tm.ll
9 11 bench/rust-analyzer-rs/optimized/4pru50b7olh1oo5y.ll
1 3 bench/rust-analyzer-rs/optimized/55rapmz9wj276hro.ll
7 9 bench/rustfmt-rs/optimized/1mznjg1e09hdetpr.ll
10 11 bench/softposit-rs/optimized/1e6z9tsqxvhrpdzq.ll
20 21 bench/spike/optimized/s_mulAddF128.ll
10 14 bench/stb/optimized/stb_image.ll
11 14 bench/tls-rs/optimized/1oa4q9ydtxtlathz.ll
13 16 bench/tls-rs/optimized/526n7g8rwc65lc9b.ll
6 8 bench/tokio-rs/optimized/um69cc05lgsv45r.ll
1 3 bench/typst-rs/optimized/2i78fvbm4wocuesi.ll
9 12 bench/uv-rs/optimized/70oya9zcywt1h05ti94rztobj.ll
11 13 bench/uv-rs/optimized/8zwhk62y11bhvj8mpp3621p3f.ll
49 53 bench/wasmtime-rs/optimized/1f5p54q9915bq6gz.ll
11 15 bench/wasmtime-rs/optimized/enal6epyb0tyurl.ll
5 11 bench/wireshark/optimized/packet-dsr.ll
7 13 bench/wireshark/optimized/packet-dvbci.ll
3 4 bench/wireshark/optimized/packet-eap.ll
9 6 bench/wireshark/optimized/packet-gsm_sms.ll
18 20 bench/wireshark/optimized/packet-spice.ll
4 9 bench/wireshark/optimized/strutil.ll
8 8 bench/wolfssl/optimized/sp_int.ll
12 13 bench/yalantinglibs/optimized/chat_room.ll
13 11 bench/yalantinglibs/optimized/test_messages_proto2.pb.ll
32 35 bench/yara-x-rs/optimized/cuynzwkmf4lxp81cl2zzl1pa9.ll
1 1 bench/z3/optimized/dd_pdd.ll
3 9 bench/zed-rs/optimized/1ygefjntcorejwfgj3uddszly.ll
2 4 bench/zed-rs/optimized/3ovedgaw1yidfseciw93ubpv7.ll
6 9 bench/zed-rs/optimized/8uq0kpgi8jqekwanrrlvbr0h1.ll

[github-actions]

The provided diff contains multiple changes across various LLVM IR files. Below is a high-level summary of up to 5 major changes, focusing on significant transformations and ignoring minor formatting or renaming changes:

1. Loop Induction Variable Initialization Fix: In several files (e.g., cecSatG2.ll, dauDsd.ll), the incoming block for phi nodes in loop headers was corrected from .lr.ph89 to .lr.ph88, ensuring correct initialization of induction variables. This reflects a fix in loop preheader identification or loop rotation, impacting loop optimization correctness.

2. Control Flow Simplification and Critical Edge Handling: In dauDsd.ll, the .loopexit block was modified to remove a predecessor (Abc_TtSuppOnlyOne.exit.thread) and a new block .loopexit.thread was introduced. This suggests a transformation to reduce critical edges and improve control flow structure, possibly for better optimization or debuggability.

3. Elimination of Redundant smax Call: In dsdTree.ll, a redundant @llvm.smax.i32 call was removed. The result of smax with a constant 1 was replaced by a direct zext of the original value, indicating a missed optimization that was later corrected, simplifying the IR and reducing unnecessary operations.

4. Switch Table and Control Flow Optimization in assimp: In FBXConverter.ll, a new switch table @switch.table._ZN6Assimp3FBX12FBXConverter30NameTransformationCompPropertyENS1_18TransformationCompE was added, and an existing switch table was updated to remove a poison entry. This indicates improved code generation for switch statements, likely enabling better jump table optimizations.

5. Phi Node and Control Flow Restructuring in actix-rs: Multiple files under actix-rs/optimized/ show extensive restructuring of basic blocks and phi nodes, particularly around string slicing and boundary checks. For example, redundant checks and branches were eliminated, and new .thread blocks were introduced to merge control flow paths, improving performance by reducing branching and enabling better SROA (Scalar Replacement of Aggregates).

These changes collectively reflect improvements in loop handling, control flow optimization, removal of redundant operations, and better code generation for switches and structured control flow.

model: qwen-plus-latest
CompletionUsage(completion_tokens=498, prompt_tokens=112696, total_tokens=113194, completion_tokens_details=None, prompt_tokens_details=None)