[llvm] [SLP] Initial vectorization of non-power-of-2 ops. (PR #77790)

Sat Apr 13 02:15:10 PDT 2024

dtcxzyw wrote:

@fhahn This PR breaks our CI: https://github.com/dtcxzyw/llvm-opt-benchmark/issues/514

```
opt: /home/dtcxzyw/WorkSpace/Projects/compilers/llvm-project/llvm/include/llvm/ADT/ArrayRef.h:169: const T& llvm::ArrayRef<T>::front() const [with T = llvm::Value*]: Assertion `!empty()' failed.
PLEASE submit a bug report to https://github.com/llvm/llvm-project/issues/ and include the crash backtrace.
Stack dump:
0.      Program arguments: bin/opt -passes=slp-vectorizer reduced.ll
 #0 0x0000721336bed7b0 llvm::sys::PrintStackTrace(llvm::raw_ostream&, int) (/home/dtcxzyw/WorkSpace/Projects/compilers/LLVM/llvm-build/bin/../lib/libLLVMSupport.so.19.0git+0x1ed7b0)
 #1 0x0000721336beabbf llvm::sys::RunSignalHandlers() (/home/dtcxzyw/WorkSpace/Projects/compilers/LLVM/llvm-build/bin/../lib/libLLVMSupport.so.19.0git+0x1eabbf)
 #2 0x0000721336bead15 SignalHandler(int) Signals.cpp:0:0
 #3 0x0000721336642520 (/lib/x86_64-linux-gnu/libc.so.6+0x42520)
 #4 0x00007213366969fc __pthread_kill_implementation ./nptl/pthread_kill.c:44:76
 #5 0x00007213366969fc __pthread_kill_internal ./nptl/pthread_kill.c:78:10
 #6 0x00007213366969fc pthread_kill ./nptl/pthread_kill.c:89:10
 #7 0x0000721336642476 gsignal ./signal/../sysdeps/posix/raise.c:27:6
 #8 0x00007213366287f3 abort ./stdlib/abort.c:81:7
 #9 0x000072133662871b _nl_load_domain ./intl/loadmsgcat.c:1177:9
#10 0x0000721336639e96 (/lib/x86_64-linux-gnu/libc.so.6+0x39e96)
#11 0x000072133178328c llvm::SLPVectorizerPass::vectorizeStores(llvm::ArrayRef<llvm::StoreInst*>, llvm::slpvectorizer::BoUpSLP&)::'lambda'(std::set<std::pair<unsigned int, int>, llvm::SLPVectorizerPass::vectorizeStores(llvm::ArrayRef<llvm::StoreInst*>, llvm::slpvectorizer::BoUpSLP&)::StoreDistCompare, std::allocator<std::pair<unsigned int, int>>> const&)::operator()(std::set<std::pair<unsigned int, int>, llvm::SLPVectorizerPass::vectorizeStores(llvm::ArrayRef<llvm::StoreInst*>, llvm::slpvectorizer::BoUpSLP&)::StoreDistCompare, std::allocator<std::pair<unsigned int, int>>> const&) const SLPVectorizer.cpp:0:0
#12 0x0000721331783b90 llvm::SLPVectorizerPass::vectorizeStores(llvm::ArrayRef<llvm::StoreInst*>, llvm::slpvectorizer::BoUpSLP&) (/home/dtcxzyw/WorkSpace/Projects/compilers/LLVM/llvm-build/bin/../lib/../lib/libLLVMVectorize.so.19.0git+0x183b90)
#13 0x0000721331784286 llvm::SLPVectorizerPass::vectorizeStoreChains(llvm::slpvectorizer::BoUpSLP&) (/home/dtcxzyw/WorkSpace/Projects/compilers/LLVM/llvm-build/bin/../lib/../lib/libLLVMVectorize.so.19.0git+0x184286)
#14 0x000072133178540f llvm::SLPVectorizerPass::runImpl(llvm::Function&, llvm::ScalarEvolution*, llvm::TargetTransformInfo*, llvm::TargetLibraryInfo*, llvm::AAResults*, llvm::LoopInfo*, llvm::DominatorTree*, llvm::AssumptionCache*, llvm::DemandedBits*, llvm::OptimizationRemarkEmitter*) (.part.0) SLPVectorizer.cpp:0:0
#15 0x0000721331785e1b llvm::SLPVectorizerPass::run(llvm::Function&, llvm::AnalysisManager<llvm::Function>&) (/home/dtcxzyw/WorkSpace/Projects/compilers/LLVM/llvm-build/bin/../lib/../lib/libLLVMVectorize.so.19.0git+0x185e1b)
#16 0x000072133206f636 llvm::detail::PassModel<llvm::Function, llvm::SLPVectorizerPass, llvm::AnalysisManager<llvm::Function>>::run(llvm::Function&, llvm::AnalysisManager<llvm::Function>&) (/home/dtcxzyw/WorkSpace/Projects/compilers/LLVM/llvm-build/bin/../lib/../lib/libLLVMPasses.so.19.0git+0x6f636)
#17 0x000072132f51eec1 llvm::PassManager<llvm::Function, llvm::AnalysisManager<llvm::Function>>::run(llvm::Function&, llvm::AnalysisManager<llvm::Function>&) (/home/dtcxzyw/WorkSpace/Projects/compilers/LLVM/llvm-build/bin/../lib/../lib/libLLVMCore.so.19.0git+0x31eec1)
#18 0x00007213358a45d6 llvm::detail::PassModel<llvm::Function, llvm::PassManager<llvm::Function, llvm::AnalysisManager<llvm::Function>>, llvm::AnalysisManager<llvm::Function>>::run(llvm::Function&, llvm::AnalysisManager<llvm::Function>&) (/home/dtcxzyw/WorkSpace/Projects/compilers/LLVM/llvm-build/bin/../lib/../lib/libLLVMX86CodeGen.so.19.0git+0xa45d6)
#19 0x000072132f51db9b llvm::ModuleToFunctionPassAdaptor::run(llvm::Module&, llvm::AnalysisManager<llvm::Module>&) (/home/dtcxzyw/WorkSpace/Projects/compilers/LLVM/llvm-build/bin/../lib/../lib/libLLVMCore.so.19.0git+0x31db9b)
#20 0x00007213358a59c6 llvm::detail::PassModel<llvm::Module, llvm::ModuleToFunctionPassAdaptor, llvm::AnalysisManager<llvm::Module>>::run(llvm::Module&, llvm::AnalysisManager<llvm::Module>&) (/home/dtcxzyw/WorkSpace/Projects/compilers/LLVM/llvm-build/bin/../lib/../lib/libLLVMX86CodeGen.so.19.0git+0xa59c6)
#21 0x000072132f51b9d1 llvm::PassManager<llvm::Module, llvm::AnalysisManager<llvm::Module>>::run(llvm::Module&, llvm::AnalysisManager<llvm::Module>&) (/home/dtcxzyw/WorkSpace/Projects/compilers/LLVM/llvm-build/bin/../lib/../lib/libLLVMCore.so.19.0git+0x31b9d1)
#22 0x000072133702e5c5 llvm::runPassPipeline(llvm::StringRef, llvm::Module&, llvm::TargetMachine*, llvm::TargetLibraryInfoImpl*, llvm::ToolOutputFile*, llvm::ToolOutputFile*, llvm::ToolOutputFile*, llvm::StringRef, llvm::ArrayRef<llvm::PassPlugin>, llvm::ArrayRef<std::function<void (llvm::PassBuilder&)>>, llvm::opt_tool::OutputKind, llvm::opt_tool::VerifierKind, bool, bool, bool, bool, bool, bool, bool) (/home/dtcxzyw/WorkSpace/Projects/compilers/LLVM/llvm-build/bin/../lib/libLLVMOptDriver.so.19.0git+0x245c5)
#23 0x0000721337039216 optMain (/home/dtcxzyw/WorkSpace/Projects/compilers/LLVM/llvm-build/bin/../lib/libLLVMOptDriver.so.19.0git+0x2f216)
#24 0x0000721336629d90 __libc_start_call_main ./csu/../sysdeps/nptl/libc_start_call_main.h:58:16
#25 0x0000721336629e40 call_init ./csu/../csu/libc-start.c:128:20
#26 0x0000721336629e40 __libc_start_main ./csu/../csu/libc-start.c:379:5
#27 0x0000566543d65095 _start (bin/opt+0x1095)
Aborted (core dumped)
```

Reduced test case:
```
; opt -passes=slp-vectorizer reduced.ll -S

target datalayout = "e-m:e-p270:32:32-p271:32:32-p272:64:64-i64:64-i128:128-f80:128-n8:16:32:64-S128"
target triple = "x86_64-unknown-linux-gnu"

define void @test(i24 %0, ptr %.pre) {
  br label %2

2:                                                ; preds = %2, %1
  %.sroa.28.0.extract.trunc = trunc i24 %0 to i8
  store i8 %.sroa.28.0.extract.trunc, ptr %.pre, align 1
  %3 = getelementptr i8, ptr %.pre, i64 1
  store i8 0, ptr %3, align 1
  br label %2
}
```

https://github.com/llvm/llvm-project/pull/77790