\n" if ! --> -
\n" if ! -
-This document serves as a high level summary of the optimization features that LLVM provides. Optimizations are implemented as Passes that traverse some portion of a program to either collect information or transform the program. @@ -69,8 +69,7 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if ! bitcode are neither analysis nor transform passes.
The table below provides a quick summary of each pass and links to the more complete pass description later in the document.
-| ANALYSIS PASSES | |
|---|---|
| Option | Name |
| -view-postdom | View postdominance tree of function |
| -view-postdom-only | View postdominance tree of function (with no function bodies) |
This section describes the LLVM Analysis Passes.
-This is a simple N^2 alias analysis accuracy evaluator. Basically, for each function in the program, it simply queries to see how the alias analysis implementation answers alias queries between each pair of @@ -224,10 +223,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
This is the default implementation of the Alias Analysis interface that simply implements a few identities (two different globals cannot alias, @@ -236,18 +235,18 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
Yet to be written.
A pass which can be used to count how many alias queries are being made and how the alias analysis implementation being used responds. @@ -255,10 +254,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
This simple pass checks alias analysis users to ensure that if they create a new value, they do not query AA without informing it of the value. @@ -272,10 +271,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
This pass is a simple dominator construction algorithm for finding forward dominator frontiers. @@ -283,10 +282,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
This pass is a simple dominator construction algorithm for finding forward dominators. @@ -294,10 +293,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
This pass, only available in opt, prints the call graph into a
.dot graph. This graph can then be processed with the "dot" tool
@@ -306,10 +305,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
This pass, only available in opt, prints the control flow graph
into a .dot graph. This graph can then be processed with the
@@ -318,10 +317,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
This pass, only available in opt, prints the control flow graph
into a .dot graph, omitting the function bodies. This graph can
@@ -331,10 +330,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
This pass, only available in opt, prints the dominator tree
into a .dot graph. This graph can then be processed with the
@@ -343,10 +342,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
This pass, only available in opt, prints the dominator tree
into a .dot graph, omitting the function bodies. This graph can
@@ -356,10 +355,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
This pass, only available in opt, prints the post dominator tree
into a .dot graph. This graph can then be processed with the
@@ -368,10 +367,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
This pass, only available in opt, prints the post dominator tree
into a .dot graph, omitting the function bodies. This graph can
@@ -381,10 +380,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
This simple pass provides alias and mod/ref information for global values that do not have their address taken, and keeps track of whether functions @@ -394,20 +393,20 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
This pass collects the count of all instructions and reports them
This analysis calculates and represents the interval partition of a function, or a preexisting interval partition. @@ -420,44 +419,44 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
Bookkeeping for "interesting" users of expressions computed from induction variables.
Interface for lazy computation of value constraint information.
Loop dependence analysis framework, which is used to detect dependences in memory accesses in loops.
LibCall Alias Analysis.
This pass statically checks for common and easily-identified constructs which produce undefined or likely unintended behavior in LLVM IR.
@@ -485,10 +484,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "\n" if !
This analysis is used to identify natural loops and determine the loop depth of various nodes of the CFG. Note that the loops identified may actually be @@ -498,10 +497,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
An analysis that determines, for a given memory operation, what preceding memory operations it depends on. It builds on alias analysis information, and @@ -511,10 +510,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
This pass decodes the debug info metadata in a module and prints in a (sufficiently-prepared-) human-readable form. @@ -524,10 +523,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
Always returns "I don't know" for alias queries. NoAA is unlike other alias analysis implementations, in that it does not chain to a previous analysis. As @@ -536,10 +535,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
The default "no profile" implementation of the abstract
ProfileInfo interface.
@@ -547,10 +546,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
This pass is a simple post-dominator construction algorithm for finding post-dominator frontiers. @@ -558,10 +557,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
This pass is a simple post-dominator construction algorithm for finding post-dominators. @@ -569,18 +568,18 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
Yet to be written.
This pass, only available in opt, prints the call graph to
standard error in a human-readable form.
@@ -588,10 +587,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
This pass, only available in opt, prints the SCCs of the call
graph to standard error in a human-readable form.
@@ -599,10 +598,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
This pass, only available in opt, prints the SCCs of each
function CFG to standard error in a human-readable form.
@@ -610,10 +609,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
Pass that prints instructions, and associated debug info:
\n" if !
Dominator Info Printer.
This pass, only available in opt, prints out call sites to
external functions that are called with constant arguments. This can be
@@ -645,10 +644,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
The PrintFunctionPass class is designed to be pipelined with
other FunctionPasses, and prints out the functions of the module
@@ -657,20 +656,20 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
This pass simply prints out the entire module when it is executed.
This pass is used to seek out all of the types in use by the program. Note that this analysis explicitly does not include types only used by the symbol @@ -678,20 +677,20 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
Profiling information that estimates the profiling information in a very crude and unimaginative way.
A concrete implementation of profiling information that loads the information from a profile dump file. @@ -699,16 +698,16 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
Pass that checks profiling information for plausibility.
The RegionInfo pass detects single entry single exit regions in a
function, where a region is defined as any subgraph that is connected to the
@@ -718,10 +717,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
The ScalarEvolution analysis can be used to analyze and
catagorize scalar expressions in loops. It specializes in recognizing general
@@ -737,10 +736,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
Simple alias analysis implemented in terms of ScalarEvolution queries. This differs from traditional loop dependence analysis in that it tests @@ -753,25 +752,26 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
Provides other passes access to information on how the size and alignment required by the the target ABI for various data types.
This section describes the LLVM Transform Passes.
-ADCE aggressively tries to eliminate code. This pass is similar to DCE but it assumes that values are dead until proven otherwise. This is similar to SCCP, except applied to @@ -779,19 +779,19 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
A custom inliner that handles only functions that are marked as "always inline".
This pass promotes "by reference" arguments to be "by value" arguments. In practice, this means looking for internal functions that have pointer @@ -819,10 +819,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
This pass is a very simple profile guided basic block placement algorithm. The idea is to put frequently executed blocks together at the start of the function and hopefully increase the number of fall-through conditional @@ -831,10 +831,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
Break all of the critical edges in the CFG by inserting a dummy basic block. It may be "required" by passes that cannot deal with critical edges. This @@ -844,20 +844,20 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
Merges duplicate global constants together into a single constant that is shared. This is useful because some passes (ie TraceValues) insert a lot of @@ -867,10 +867,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
This file implements constant propagation and merging. It looks for instructions involving only constant operands and replaces them with a constant value instead of an instruction. For example:
@@ -883,10 +883,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "\n" if !
Dead code elimination is similar to dead instruction elimination, but it rechecks instructions that were used by removed @@ -895,10 +895,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
This pass deletes dead arguments from internal functions. Dead argument elimination removes arguments which are directly dead, as well as arguments @@ -913,10 +913,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
This pass is used to cleanup the output of GCC. It eliminate names for types
that are unused in the entire translation unit, using the )) { s:^ *///? ?::; print " \n" if !
Dead instruction elimination performs a single pass over the function, removing instructions that are obviously dead. @@ -936,10 +936,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
A trivial dead store elimination that only considers basic-block local redundant stores. @@ -947,10 +947,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
A simple interprocedural pass which walks the call-graph, looking for functions which do not access or only read non-local memory, and marking them readnone/readonly. In addition, it marks function arguments (of pointer type) @@ -962,10 +962,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
This transform is designed to eliminate unreachable internal globals from the program. It uses an aggressive algorithm, searching out globals that are @@ -976,10 +976,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
This pass transforms simple global variables that never have their address taken. If obviously true, it marks read/write globals as constant, deletes @@ -988,10 +988,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
This pass performs global value numbering to eliminate fully and partially redundant instructions. It also performs redundant load elimination. @@ -999,10 +999,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
This transformation analyzes and transforms the induction variables (and computations derived from them) into simpler forms suitable for subsequent @@ -1050,20 +1050,20 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
Bottom-up inlining of functions into callees.
This pass instruments the specified program with counters for edge profiling. Edge profiling can give a reasonable approximation of the hot paths through a @@ -1078,10 +1078,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
This pass instruments the specified program with counters for edge profiling. Edge profiling can give a reasonable approximation of the hot paths through a program, and is used for a wide variety of program transformations. @@ -1089,10 +1089,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
Combine instructions to form fewer, simple instructions. This pass does not modify the CFG This pass is where algebraic @@ -1143,10 +1143,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
This pass loops over all of the functions in the input module, looking for a main function. If a main function is found, all other functions and all @@ -1155,10 +1155,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
This pass implements an extremely simple interprocedural constant propagation pass. It could certainly be improved in many different ways, @@ -1169,10 +1169,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
An interprocedural variant of Sparse Conditional Constant Propagation. @@ -1180,10 +1180,10 @@ perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print "
\n" if !
Jump threading tries to find distinct threads of control flow running through a basic block. This pass looks at blocks that have multiple predecessors and @@ -1209,10 +1209,10 @@ if (X < 3) {
This pass transforms loops by placing phi nodes at the end of the loops for all values that are live across the loop boundary. For example, it turns @@ -1238,10 +1238,10 @@ if (X < 3) {
This pass performs loop invariant code motion, attempting to remove as much code from the body of a loop as possible. It does this by either hoisting @@ -1275,10 +1275,10 @@ if (X < 3) {
This file implements the Dead Loop Deletion Pass. This pass is responsible for eliminating loops with non-infinite computable trip counts that have no @@ -1288,10 +1288,10 @@ if (X < 3) {
A pass wrapper around the ExtractLoop() scalar transformation to
extract each top-level loop into its own new function. If the loop is the
@@ -1301,10 +1301,10 @@ if (X < 3) {
Similar to Extract loops into new functions, this pass extracts one natural loop from the program into a function if it @@ -1313,10 +1313,10 @@ if (X < 3) {
This pass performs a strength reduction on array references inside loops that have as one or more of their components the loop induction variable. This is @@ -1327,18 +1327,18 @@ if (X < 3) {
A simple loop rotation transformation.
This pass performs several transformations to transform natural loops into a simpler form, which makes subsequent analyses and transformations simpler and @@ -1376,10 +1376,10 @@ if (X < 3) {
This pass implements a simple loop unroller. It works best when loops have been canonicalized by the -indvars pass, @@ -1388,10 +1388,10 @@ if (X < 3) {
This pass transforms loops that contain branches on loop-invariant conditions to have multiple loops. For example, it turns the left into the right code: @@ -1418,10 +1418,10 @@ if (X < 3) {
This pass lowers atomic intrinsics to non-atomic form for use in a known non-preemptible environment. @@ -1436,10 +1436,10 @@ if (X < 3) {
This transformation is designed for use by code generators which do not yet support stack unwinding. This pass supports two models of exception handling @@ -1477,10 +1477,10 @@ if (X < 3) {
Lowers setjmp and longjmp to use the LLVM invoke and unwind instructions as necessary. @@ -1506,10 +1506,10 @@ if (X < 3) {
Rewrites switch instructions with a sequence of branches, which allows targets to get away with not implementing the switch instruction until @@ -1518,10 +1518,10 @@ if (X < 3) {
This file promotes memory references to be register references. It promotes alloca instructions which only have loads and @@ -1534,10 +1534,10 @@ if (X < 3) {
This pass performs various transformations related to eliminating memcpy calls, or transforming sets of stores into memset's. @@ -1545,10 +1545,10 @@ if (X < 3) {
This pass looks for equivalent functions that are mergable and folds them. A hash is computed from the function, based on its type and number of @@ -1566,10 +1566,10 @@ if (X < 3) {
Ensure that functions have at most one ret instruction in them. Additionally, it keeps track of which node is the new exit node of the CFG. @@ -1577,20 +1577,20 @@ if (X < 3) {
This pass performs partial inlining, typically by inlining an if statement that surrounds the body of the function.
This file implements a simple interprocedural pass which walks the call-graph, turning invoke instructions into call instructions if and @@ -1600,10 +1600,10 @@ if (X < 3) {
This pass reassociates commutative expressions in an order that is designed to promote better constant propagation, GCSE, LICM, PRE, etc. @@ -1623,10 +1623,10 @@ if (X < 3) {
This file demotes all registers to memory references. It is intented to be the inverse of -mem2reg. By converting to @@ -1640,10 +1640,10 @@ if (X < 3) {
The well-known scalar replacement of aggregates transformation. This transform breaks up alloca instructions of aggregate type (structure @@ -1662,10 +1662,10 @@ if (X < 3) {
Sparse conditional constant propagation and merging, which can be summarized as: @@ -1685,10 +1685,10 @@ if (X < 3) {
Applies a variety of small optimizations for calls to specific well-known function calls (e.g. runtime library functions). For example, a call @@ -1698,10 +1698,10 @@ if (X < 3) {
Performs dead code elimination and basic block merging. Specifically:
@@ -1717,20 +1717,20 @@ if (X < 3) {This pass moves instructions into successor blocks, when possible, so that they aren't executed on paths where their results aren't needed.
This pass finds functions that return a struct (using a pointer to the struct as the first argument of the function, marked with the 'sret' attribute) and @@ -1750,10 +1750,10 @@ if (X < 3) {
performs code stripping. this transformation can delete:
@@ -1772,10 +1772,10 @@ if (X < 3) {performs code stripping. this transformation can delete:
@@ -1794,10 +1794,10 @@ if (X < 3) {This pass loops over all of the functions in the input module, looking for dead declarations and removes them. Dead declarations are declarations of @@ -1807,10 +1807,10 @@ if (X < 3) {
This pass implements code stripping. Specifically, it can delete:
This pass implements code stripping. Specifically, it can delete:
This file transforms calls of the current function (self recursion) followed by a return instruction with a branch to the entry of the function, creating @@ -1875,10 +1875,10 @@ if (X < 3) {
This pass performs a limited form of tail duplication, intended to simplify CFGs by removing some unconditional branches. This pass is necessary to @@ -1888,17 +1888,18 @@ if (X < 3) {
This section describes the LLVM Utility Passes.
-Same as dead argument elimination, but deletes arguments to functions which are external. This is only for use by
This pass is used by bugpoint to extract all blocks from the module into their own functions.
This is a little utility pass that gives instructions names, this is mostly useful when diffing the effect of an optimization because deleting an unnamed instruction can change all other instruction numbering, making the @@ -1928,10 +1929,10 @@ if (X < 3) {
Ensures that the module is in the form required by the Module Verifier pass. @@ -1944,10 +1945,10 @@ if (X < 3) {
Verifies an LLVM IR code. This is useful to run after an optimization which is undergoing testing. Note that llvm-as verifies its input before @@ -1995,20 +1996,20 @@ if (X < 3) {
Displays the control flow graph using the GraphViz tool.
Displays the control flow graph using the GraphViz tool, but omitting function bodies. @@ -2016,20 +2017,20 @@ if (X < 3) {
Displays the dominator tree using the GraphViz tool.
Displays the dominator tree using the GraphViz tool, but omitting function bodies. @@ -2037,26 +2038,28 @@ if (X < 3) {
Displays the post dominator tree using the GraphViz tool.
Displays the post dominator tree using the GraphViz tool, but omitting function bodies.