pre-commit: PR153842

[zyw-bot]

Link: llvm/llvm-project#153842
Requested by: @dtcxzyw

[zyw-bot]

Diff mode

runner: ariselab-64c-docker
baseline: llvm/llvm-project@f0967fc
patch: llvm/llvm-project#153842
sha256: 41780d234462fe032de407acf4fcc2db8b62460c814642a7e7f2a5e5d23ff705
commit: 04bc5fc

75 files changed, 36010 insertions(+), 36339 deletions(-)

Improvements:
  bdce.NumSimplified 5953 -> 5954 +0.02%
  instcombine.NumDeadInst 42243355 -> 42244425 +0.00%
  gvn.NumGVNInstr 155558 -> 155560 +0.00%
  gvn.NumGVNPRE 155558 -> 155560 +0.00%
  instcombine.NumConstProp 158945 -> 158946 +0.00%
  simplifycfg.NumSpeculations 399969 -> 399971 +0.00%
  simplifycfg.NumSinkCommonInstrs 803151 -> 803155 +0.00%
  correlated-value-propagation.NumAddNSW 276889 -> 276890 +0.00%
  simplifycfg.NumSinkCommonCode 370722 -> 370723 +0.00%
  instcombine.NumCombined 124882231 -> 124882561 +0.00%
Regressions:
  simplifycfg.NumFoldValueComparisonIntoPredecessors 512764 -> 512756 -0.00%
  correlated-value-propagation.NumDeadCases 70120 -> 70119 -0.00%
  local.NumPHICSEs 197869 -> 197868 -0.00%
  gvn.IsValueFullyAvailableInBlockNumSpeculationsMax 602860 -> 602859 -0.00%
  correlated-value-propagation.NumPhis 1291680 -> 1291678 -0.00%
  globalsmodref-aa.NumReadMemFunctions 1203882 -> 1203881 -0.00%
  jump-threading.NumFolds 2641504 -> 2641502 -0.00%
  gvn.NumGVNSimpl 4555297 -> 4555294 -0.00%
  abstract-call-sites.NumDirectAbstractCallSites 9569263 -> 9569258 -0.00%
  instsimplify.NumSimplified 2526305 -> 2526304 -0.00%

2 2 bench/arrow/optimized/function_internal.ll
90 90 bench/gromacs/optimized/checkpoint.ll
3 5 bench/linux/optimized/memtype.ll
6 6 bench/llvm/optimized/CGCoroutine.ll
51 51 bench/llvm/optimized/TypePrinter.ll
13 15 bench/oiio/optimized/strutil.ll
3 3 bench/oiio/optimized/typedesc.ll
78 90 bench/opencv/optimized/connectedcomponents.ll
87 86 bench/openjdk/optimized/awt_parseImage.ll
113 117 bench/postgres/optimized/array_userfuncs.ll
53 59 bench/postgres/optimized/be-fsstubs.ll
110 118 bench/postgres/optimized/brin_minmax_multi.ll
60 62 bench/postgres/optimized/brin_tuple.ll
110 116 bench/postgres/optimized/cryptohashfuncs.ll
30 31 bench/postgres/optimized/datum.ll
39 41 bench/postgres/optimized/dbsize.ll
102 108 bench/postgres/optimized/dependencies.ll
122 128 bench/postgres/optimized/detoast.ll
79 83 bench/postgres/optimized/dict.ll
81 85 bench/postgres/optimized/encode.ll
281 297 bench/postgres/optimized/hashfunc.ll
115 119 bench/postgres/optimized/indextuple.ll
229 245 bench/postgres/optimized/json.ll
70 78 bench/postgres/optimized/jsonb.ll
98 106 bench/postgres/optimized/jsonb_gin.ll
52 54 bench/postgres/optimized/jsonb_util.ll
86 36 bench/postgres/optimized/logicalfuncs.ll
37 41 bench/postgres/optimized/mbutils.ll
89 95 bench/postgres/optimized/multirangetypes.ll
100 106 bench/postgres/optimized/mvdistinct.ll
71 75 bench/postgres/optimized/network.ll
13 13 bench/postgres/optimized/nodeSubplan.ll
17 17 bench/postgres/optimized/quote.ll
21 23 bench/postgres/optimized/rangetypes.ll
152 156 bench/postgres/optimized/regress.ll
113 119 bench/postgres/optimized/selfuncs.ll
95 105 bench/postgres/optimized/spgutils.ll
99 103 bench/postgres/optimized/to_tsany.ll
52 54 bench/postgres/optimized/toast_compression.ll
37 39 bench/postgres/optimized/toast_helper.ll
151 155 bench/postgres/optimized/ts_selfuncs.ll
137 145 bench/postgres/optimized/tsginidx.ll
12 12 bench/z3/optimized/nlsat_simple_checker.ll

[github-actions]

The provided diff contains several changes across multiple LLVM IR files. Here are the major changes summarized:

1. Optimization of Conditional Checks and Switch Statements: In several files, conditional checks involving comparisons and bit manipulations are optimized. For example, in bench/llvm/optimized/TypePrinter.ll, a switch statement is replaced with a series of comparisons and conditional branches, which can be more efficient in certain scenarios.

2. Simplification of Phi Nodes: Phi nodes in various functions are simplified or restructured. For instance, in bench/gromacs/optimized/checkpoint.ll, phi nodes are updated to reflect changes in control flow, ensuring that the correct values are selected based on the incoming paths.

3. Control Flow Adjustments: Control flow is adjusted in multiple functions to improve performance. In bench/postgres/optimized/hashfunc.ll, the control flow is modified to reduce the number of branches and improve the predictability of the code.

4. Memory Access and Alignment: Changes in memory access patterns and alignment are observed. For example, in bench/postgres/optimized/detoast.ll, memory access patterns are optimized to reduce the number of loads and stores, and alignment is adjusted to ensure better performance.

5. Function Call and Return Simplification: Function calls and returns are simplified in some cases. In bench/postgres/optimized/json.ll, function calls are streamlined to reduce overhead, and return paths are optimized to minimize the number of instructions executed.

These changes collectively aim to improve the performance and efficiency of the generated code by optimizing conditional checks, simplifying control flow, and improving memory access patterns.

model: qwen-plus-latest
CompletionUsage(completion_tokens=331, prompt_tokens=111192, total_tokens=111523, completion_tokens_details=None, prompt_tokens_details=None)