LLVMLLVM Compiler SystemThree primary LLVM componentsThe LLVM C/C++ CompilerThe LLVM C/C++ Compiler (cont)Looking into events at compile-timeExample Intermediate RepresentationCurrent Uses/Projects made with LLVMFuture of LLVMLLVMLLVMDeveloped by University of Developed by University of Illinois at Urbana-Champaign Illinois at Urbana-Champaign CIS deptCIS deptCisc 471Cisc 471Matthew WarnerMatthew WarnerLLVM Compiler SystemLLVM Compiler SystemLLVM = Low Level Virtual Machine LLVM = Low Level Virtual Machine The LLVM Compiler InfrastructureThe LLVM Compiler InfrastructureProvides reusable components for building Provides reusable components for building compilerscompilersReduce the time/cost to build a new compilerReduce the time/cost to build a new compilerBuild static compilers, JITs, trace-based Build static compilers, JITs, trace-based optimizers, ...optimizers, ...The LLVM Compiler FrameworkThe LLVM Compiler FrameworkEnd-to-end compilers using the LLVM infrastructureEnd-to-end compilers using the LLVM infrastructureC and C++ are robust and aggressive:C and C++ are robust and aggressive:Java, Scheme and others are in developmentJava, Scheme and others are in developmentEmit C code or native code for X86, Sparc, PowerPCEmit C code or native code for X86, Sparc, PowerPCThree primary LLVM Three primary LLVM componentscomponentsThe LLVM The LLVM Virtual Instruction SetVirtual Instruction SetThe common language- and target-independent IRThe common language- and target-independent IRInternal (IR) and external (persistent) Internal (IR) and external (persistent) representationrepresentationA collection of well-integrated librariesA collection of well-integrated librariesAnalyses, optimizations, code generators, JIT Analyses, optimizations, code generators, JIT compiler, garbage collection support, profiling, …compiler, garbage collection support, profiling, …A collection of tools built from the librariesA collection of tools built from the librariesAssemblers, automatic debugger, linker, code Assemblers, automatic debugger, linker, code generator, compiler driver, modular optimizer, …generator, compiler driver, modular optimizer, …The LLVM C/C++ The LLVM C/C++ CompilerCompilerFrom the high level, it is a standard compiler:From the high level, it is a standard compiler:Compatible with standard makefilesCompatible with standard makefilesUses GCC 3.4 C and C++ parserUses GCC 3.4 C and C++ parserDistinguishing features:Distinguishing features:Uses LLVM optimizers, not GCC optimizersUses LLVM optimizers, not GCC optimizers.o files contain LLVM IR/bytecode, not machine code.o files contain LLVM IR/bytecode, not machine codeExecutable can be bytecode (JIT’d) or machine codeExecutable can be bytecode (JIT’d) or machine codellvmg++llvmgccC fileC++ file.o file.o filellvm linker executableCompile Time Link TimeThe LLVM C/C++ Compiler The LLVM C/C++ Compiler (cont)(cont)Standard compiler organization, which uses LLVM as midlevel IR:– Language specific front-end lowers code to LLVM IR– Language/target independent optimizers improve code– Code generator converts LLVM code to target (e.g. IA64) codeLooking into events at Looking into events at compile-timecompile-timellvmgccC file .o file llvmg++C++ file .o fileModified version of G++Emits LLVM IR as text fileLowers C++ AST to LLVMModified version of GCCEmits LLVM IR as text fileLowers C AST to LLVMLLVM IR ParserLLVM Verifier40 LLVM Analysis & Optimization PassesLLVM .bc File WriterC to LLVM FrontendCompile-time OptimizerC++ to LLVM FrontendCompile-time Optimizer“cc1” “cc1plus” “gccas”“gccas”Dead Global Elimination, IP Constant Propagation, Dead Argument Elimination, Inlining, Reassociation, LICM, Loop Opts, Memory Promotion, Dead Store Elimination, ADCE, …Example Intermediate Example Intermediate RepresentationRepresentation; Declare the string constant as a global constant...; Declare the string constant as a global [email protected]@.LC0 = = internalinternal constant [13 x i8] c"hello world\0A\00" constant [13 x i8] c"hello world\0A\00" ; [13 x i8]*; [13 x i8]* ; External declaration of the puts function; External declaration of the puts functiondeclare i32 @puts(i8 *) declare i32 @puts(i8 *) ; i32(i8 *)*; i32(i8 *)* ; Definition of main function; Definition of main functiondefine i32 @main() { define i32 @main() { ; i32()*; i32()* ; Convert [13x i8 ]* to i8 *...; Convert [13x i8 ]* to i8 *... %cast210 = getelementptr %cast210 = getelementptr [13 x i8 ]* @.LC0, i64 0, i64 0 [13 x i8 ]* @.LC0, i64 0, i64 0 ; i8 *; i8 * ; Call puts function to write out the string to stdout...; Call puts function to write out the string to stdout... call i32 @puts(i8 * %cast210) call i32 @puts(i8 * %cast210)ret i32 0 ret i32 0 } }Current Uses/Projects made Current Uses/Projects made with LLVMwith LLVMRegister Allocation solving by Using Register Allocation solving by Using Puzzle Solving techniquesPuzzle Solving techniquesScheme compilerScheme compilerJIT Compiler for LLVM bytecodeJIT Compiler for LLVM bytecodeFuture of LLVMFuture of LLVMCurrently a partial version of the Currently a partial version of the Java compiler exists and need to be Java compiler exists and need to be extendedextendedImprove optimizations for Improve optimizations for performanceperformanceMost of LLVM’s current optimizations Most of LLVM’s current optimizations are performance based, need are performance based, need optimizations for sizeoptimizations for