package rune
Install
dune-project
Dependency
Authors
Maintainers
Sources
sha256=93abc49d075a1754442ccf495645bc4fdc83e4c66391ec8aca8fa15d2b4f44d2
sha512=5eb958c51f30ae46abded4c96f48d1825f79c7ce03f975f9a6237cdfed0d62c0b4a0774296694def391573d849d1f869919c49008acffca95946b818ad325f6f
doc/llvm_passbuilder/Llvm_passbuilder/index.html
Module Llvm_passbuilderSource
val run_passes :
Llvm.llmodule ->
string ->
Llvm_target.TargetMachine.t ->
llpassbuilder_options ->
(unit, string) Stdlib.resultrun_passes m passes tm opts runs a set of passes over a module. The format of the string passes is the same as opt's -passes argument for the new pass manager. Individual passes may be specified, separated by commas. Full pipelines may also be invoked. See LLVMRunPasses.
Creates a new set of options for a PassBuilder. See llvm::LLVMPassBuilderOptions::LLVMPassBuilderOptions.
Toggles adding the VerifierPass for the PassBuilder. See llvm::LLVMPassBuilderOptions::VerifyEach.
Toggles debug logging. See llvm::LLVMPassBuilderOptions::DebugLogging.
Tuning option to set loop interleaving on/off, set based on opt level. See llvm::PipelineTuningOptions::LoopInterleaving.
Tuning option to enable/disable loop vectorization, set based on opt level. See llvm::PipelineTuningOptions::LoopVectorization.
Tuning option to enable/disable slp loop vectorization, set based on opt level. See llvm::PipelineTuningOptions::SLPVectorization.
Tuning option to enable/disable loop unrolling. Its default value is true. See llvm::PipelineTuningOptions::LoopUnrolling.
val passbuilder_options_set_forget_all_scev_in_loop_unroll :
llpassbuilder_options ->
bool ->
unitTuning option to forget all SCEV loops in LoopUnroll. See llvm::PipelineTuningOptions::ForgetAllSCEVInLoopUnroll.
Tuning option to cap the number of calls to retrive clobbering accesses in MemorySSA, in LICM. See llvm::PipelineTuningOptions::LicmMssaOptCap.
val passbuilder_options_set_licm_mssa_no_acc_for_promotion_cap :
llpassbuilder_options ->
int ->
unitTuning option to disable promotion to scalars in LICM with MemorySSA, if the number of accesses is too large. See llvm::PipelineTuningOptions::LicmMssaNoAccForPromotionCap.
Tuning option to enable/disable call graph profile. See llvm::PipelineTuningOptions::CallGraphProfile.
Tuning option to enable/disable function merging. See llvm::PipelineTuningOptions::MergeFunctions.
Tuning option to override the default inliner threshold. See llvm::PipelineTuningOptions::InlinerThreshold.
Disposes of the options.