2 If Passes.html is up to date, the following "one-liner" should print
5 egrep -e '^<tr><td><a href="#.*">-.*</a></td><td>.*</td></tr>$' \
6 -e '^ <a name=".*">.*</a>$' < Passes.html >html; \
7 perl >help <<'EOT' && diff -u help html; rm -f help html
8 open HTML, "<Passes.html" or die "open: Passes.html: $!\n";
10 m:^<tr><td><a href="#(.*)">-.*</a></td><td>.*</td></tr>$: or next;
11 $order{$1} = sprintf("%03d", 1 + int %order);
13 open HELP, "../Release/bin/opt -help|" or die "open: opt -help: $!\n";
15 m:^ -([^ ]+) +- (.*)$: or next;
17 $o = "000" unless defined $o;
18 push @x, "$o<tr><td><a href=\"#$1\">-$1</a></td><td>$2</td></tr>\n";
19 push @y, "$o <a name=\"$1\">-$1: $2</a>\n";
21 @x = map { s/^\d\d\d//; $_ } sort @x;
22 @y = map { s/^\d\d\d//; $_ } sort @y;
26 This (real) one-liner can also be helpful when converting comments to HTML:
28 perl -e '$/ = undef; for (split(/\n/, <>)) { s:^ *///? ?::; print " <p>\n" if !$on && $_ =~ /\S/; print " </p>\n" if $on && $_ =~ /^\s*$/; print " $_\n"; $on = ($_ =~ /\S/); } print " </p>\n" if $on'
30 ====================================
31 LLVM's Analysis and Transform Passes
32 ====================================
40 This document serves as a high level summary of the optimization features that
41 LLVM provides. Optimizations are implemented as Passes that traverse some
42 portion of a program to either collect information or transform the program.
43 The table below divides the passes that LLVM provides into three categories.
44 Analysis passes compute information that other passes can use or for debugging
45 or program visualization purposes. Transform passes can use (or invalidate)
46 the analysis passes. Transform passes all mutate the program in some way.
47 Utility passes provides some utility but don't otherwise fit categorization.
48 For example passes to extract functions to bitcode or write a module to bitcode
49 are neither analysis nor transform passes. The table of contents above
50 provides a quick summary of each pass and links to the more complete pass
51 description later in the document.
56 This section describes the LLVM Analysis Passes.
58 ``-aa-eval``: Exhaustive Alias Analysis Precision Evaluator
59 -----------------------------------------------------------
61 This is a simple N^2 alias analysis accuracy evaluator. Basically, for each
62 function in the program, it simply queries to see how the alias analysis
63 implementation answers alias queries between each pair of pointers in the
66 This is inspired and adapted from code by: Naveen Neelakantam, Francesco
67 Spadini, and Wojciech Stryjewski.
69 ``-basicaa``: Basic Alias Analysis (stateless AA impl)
70 ------------------------------------------------------
72 A basic alias analysis pass that implements identities (two different globals
73 cannot alias, etc), but does no stateful analysis.
75 ``-basiccg``: Basic CallGraph Construction
76 ------------------------------------------
80 ``-count-aa``: Count Alias Analysis Query Responses
81 ---------------------------------------------------
83 A pass which can be used to count how many alias queries are being made and how
84 the alias analysis implementation being used responds.
86 ``-da``: Dependence Analysis
87 ----------------------------
89 Dependence analysis framework, which is used to detect dependences in memory
92 ``-debug-aa``: AA use debugger
93 ------------------------------
95 This simple pass checks alias analysis users to ensure that if they create a
96 new value, they do not query AA without informing it of the value. It acts as
97 a shim over any other AA pass you want.
99 Yes keeping track of every value in the program is expensive, but this is a
102 ``-domfrontier``: Dominance Frontier Construction
103 -------------------------------------------------
105 This pass is a simple dominator construction algorithm for finding forward
108 ``-domtree``: Dominator Tree Construction
109 -----------------------------------------
111 This pass is a simple dominator construction algorithm for finding forward
115 ``-dot-callgraph``: Print Call Graph to "dot" file
116 --------------------------------------------------
118 This pass, only available in ``opt``, prints the call graph into a ``.dot``
119 graph. This graph can then be processed with the "dot" tool to convert it to
120 postscript or some other suitable format.
122 ``-dot-cfg``: Print CFG of function to "dot" file
123 -------------------------------------------------
125 This pass, only available in ``opt``, prints the control flow graph into a
126 ``.dot`` graph. This graph can then be processed with the :program:`dot` tool
127 to convert it to postscript or some other suitable format.
129 ``-dot-cfg-only``: Print CFG of function to "dot" file (with no function bodies)
130 --------------------------------------------------------------------------------
132 This pass, only available in ``opt``, prints the control flow graph into a
133 ``.dot`` graph, omitting the function bodies. This graph can then be processed
134 with the :program:`dot` tool to convert it to postscript or some other suitable
137 ``-dot-dom``: Print dominance tree of function to "dot" file
138 ------------------------------------------------------------
140 This pass, only available in ``opt``, prints the dominator tree into a ``.dot``
141 graph. This graph can then be processed with the :program:`dot` tool to
142 convert it to postscript or some other suitable format.
144 ``-dot-dom-only``: Print dominance tree of function to "dot" file (with no function bodies)
145 -------------------------------------------------------------------------------------------
147 This pass, only available in ``opt``, prints the dominator tree into a ``.dot``
148 graph, omitting the function bodies. This graph can then be processed with the
149 :program:`dot` tool to convert it to postscript or some other suitable format.
151 ``-dot-postdom``: Print postdominance tree of function to "dot" file
152 --------------------------------------------------------------------
154 This pass, only available in ``opt``, prints the post dominator tree into a
155 ``.dot`` graph. This graph can then be processed with the :program:`dot` tool
156 to convert it to postscript or some other suitable format.
158 ``-dot-postdom-only``: Print postdominance tree of function to "dot" file (with no function bodies)
159 ---------------------------------------------------------------------------------------------------
161 This pass, only available in ``opt``, prints the post dominator tree into a
162 ``.dot`` graph, omitting the function bodies. This graph can then be processed
163 with the :program:`dot` tool to convert it to postscript or some other suitable
166 ``-globalsmodref-aa``: Simple mod/ref analysis for globals
167 ----------------------------------------------------------
169 This simple pass provides alias and mod/ref information for global values that
170 do not have their address taken, and keeps track of whether functions read or
171 write memory (are "pure"). For this simple (but very common) case, we can
172 provide pretty accurate and useful information.
174 ``-instcount``: Counts the various types of ``Instruction``\ s
175 --------------------------------------------------------------
177 This pass collects the count of all instructions and reports them.
179 ``-intervals``: Interval Partition Construction
180 -----------------------------------------------
182 This analysis calculates and represents the interval partition of a function,
183 or a preexisting interval partition.
185 In this way, the interval partition may be used to reduce a flow graph down to
186 its degenerate single node interval partition (unless it is irreducible).
188 ``-iv-users``: Induction Variable Users
189 ---------------------------------------
191 Bookkeeping for "interesting" users of expressions computed from induction
194 ``-lazy-value-info``: Lazy Value Information Analysis
195 -----------------------------------------------------
197 Interface for lazy computation of value constraint information.
199 ``-libcall-aa``: LibCall Alias Analysis
200 ---------------------------------------
202 LibCall Alias Analysis.
204 ``-lint``: Statically lint-checks LLVM IR
205 -----------------------------------------
207 This pass statically checks for common and easily-identified constructs which
208 produce undefined or likely unintended behavior in LLVM IR.
210 It is not a guarantee of correctness, in two ways. First, it isn't
211 comprehensive. There are checks which could be done statically which are not
212 yet implemented. Some of these are indicated by TODO comments, but those
213 aren't comprehensive either. Second, many conditions cannot be checked
214 statically. This pass does no dynamic instrumentation, so it can't check for
215 all possible problems.
217 Another limitation is that it assumes all code will be executed. A store
218 through a null pointer in a basic block which is never reached is harmless, but
219 this pass will warn about it anyway.
221 Optimization passes may make conditions that this pass checks for more or less
222 obvious. If an optimization pass appears to be introducing a warning, it may
223 be that the optimization pass is merely exposing an existing condition in the
226 This code may be run before :ref:`instcombine <passes-instcombine>`. In many
227 cases, instcombine checks for the same kinds of things and turns instructions
228 with undefined behavior into unreachable (or equivalent). Because of this,
229 this pass makes some effort to look through bitcasts and so on.
231 ``-loops``: Natural Loop Information
232 ------------------------------------
234 This analysis is used to identify natural loops and determine the loop depth of
235 various nodes of the CFG. Note that the loops identified may actually be
236 several natural loops that share the same header node... not just a single
239 ``-memdep``: Memory Dependence Analysis
240 ---------------------------------------
242 An analysis that determines, for a given memory operation, what preceding
243 memory operations it depends on. It builds on alias analysis information, and
244 tries to provide a lazy, caching interface to a common kind of alias
247 ``-module-debuginfo``: Decodes module-level debug info
248 ------------------------------------------------------
250 This pass decodes the debug info metadata in a module and prints in a
251 (sufficiently-prepared-) human-readable form.
253 For example, run this pass from ``opt`` along with the ``-analyze`` option, and
254 it'll print to standard output.
256 ``-no-aa``: No Alias Analysis (always returns 'may' alias)
257 ----------------------------------------------------------
259 This is the default implementation of the Alias Analysis interface. It always
260 returns "I don't know" for alias queries. NoAA is unlike other alias analysis
261 implementations, in that it does not chain to a previous analysis. As such it
262 doesn't follow many of the rules that other alias analyses must.
264 ``-postdomfrontier``: Post-Dominance Frontier Construction
265 ----------------------------------------------------------
267 This pass is a simple post-dominator construction algorithm for finding
268 post-dominator frontiers.
270 ``-postdomtree``: Post-Dominator Tree Construction
271 --------------------------------------------------
273 This pass is a simple post-dominator construction algorithm for finding
276 ``-print-alias-sets``: Alias Set Printer
277 ----------------------------------------
281 ``-print-callgraph``: Print a call graph
282 ----------------------------------------
284 This pass, only available in ``opt``, prints the call graph to standard error
285 in a human-readable form.
287 ``-print-callgraph-sccs``: Print SCCs of the Call Graph
288 -------------------------------------------------------
290 This pass, only available in ``opt``, prints the SCCs of the call graph to
291 standard error in a human-readable form.
293 ``-print-cfg-sccs``: Print SCCs of each function CFG
294 ----------------------------------------------------
296 This pass, only available in ``opt``, printsthe SCCs of each function CFG to
297 standard error in a human-readable fom.
299 ``-print-dom-info``: Dominator Info Printer
300 -------------------------------------------
302 Dominator Info Printer.
304 ``-print-externalfnconstants``: Print external fn callsites passed constants
305 ----------------------------------------------------------------------------
307 This pass, only available in ``opt``, prints out call sites to external
308 functions that are called with constant arguments. This can be useful when
309 looking for standard library functions we should constant fold or handle in
312 ``-print-function``: Print function to stderr
313 ---------------------------------------------
315 The ``PrintFunctionPass`` class is designed to be pipelined with other
316 ``FunctionPasses``, and prints out the functions of the module as they are
319 ``-print-module``: Print module to stderr
320 -----------------------------------------
322 This pass simply prints out the entire module when it is executed.
324 .. _passes-print-used-types:
326 ``-print-used-types``: Find Used Types
327 --------------------------------------
329 This pass is used to seek out all of the types in use by the program. Note
330 that this analysis explicitly does not include types only used by the symbol
333 ``-regions``: Detect single entry single exit regions
334 -----------------------------------------------------
336 The ``RegionInfo`` pass detects single entry single exit regions in a function,
337 where a region is defined as any subgraph that is connected to the remaining
338 graph at only two spots. Furthermore, an hierarchical region tree is built.
340 ``-scalar-evolution``: Scalar Evolution Analysis
341 ------------------------------------------------
343 The ``ScalarEvolution`` analysis can be used to analyze and catagorize scalar
344 expressions in loops. It specializes in recognizing general induction
345 variables, representing them with the abstract and opaque ``SCEV`` class.
346 Given this analysis, trip counts of loops and other important properties can be
349 This analysis is primarily useful for induction variable substitution and
352 ``-scev-aa``: ScalarEvolution-based Alias Analysis
353 --------------------------------------------------
355 Simple alias analysis implemented in terms of ``ScalarEvolution`` queries.
357 This differs from traditional loop dependence analysis in that it tests for
358 dependencies within a single iteration of a loop, rather than dependencies
359 between different iterations.
361 ``ScalarEvolution`` has a more complete understanding of pointer arithmetic
362 than ``BasicAliasAnalysis``' collection of ad-hoc analyses.
364 ``-targetdata``: Target Data Layout
365 -----------------------------------
367 Provides other passes access to information on how the size and alignment
368 required by the target ABI for various data types.
373 This section describes the LLVM Transform Passes.
375 ``-adce``: Aggressive Dead Code Elimination
376 -------------------------------------------
378 ADCE aggressively tries to eliminate code. This pass is similar to :ref:`DCE
379 <passes-dce>` but it assumes that values are dead until proven otherwise. This
380 is similar to :ref:`SCCP <passes-sccp>`, except applied to the liveness of
383 ``-always-inline``: Inliner for ``always_inline`` functions
384 -----------------------------------------------------------
386 A custom inliner that handles only functions that are marked as "always
389 ``-argpromotion``: Promote 'by reference' arguments to scalars
390 --------------------------------------------------------------
392 This pass promotes "by reference" arguments to be "by value" arguments. In
393 practice, this means looking for internal functions that have pointer
394 arguments. If it can prove, through the use of alias analysis, that an
395 argument is *only* loaded, then it can pass the value into the function instead
396 of the address of the value. This can cause recursive simplification of code
397 and lead to the elimination of allocas (especially in C++ template code like
400 This pass also handles aggregate arguments that are passed into a function,
401 scalarizing them if the elements of the aggregate are only loaded. Note that
402 it refuses to scalarize aggregates which would require passing in more than
403 three operands to the function, because passing thousands of operands for a
404 large array or structure is unprofitable!
406 Note that this transformation could also be done for arguments that are only
407 stored to (returning the value instead), but does not currently. This case
408 would be best handled when and if LLVM starts supporting multiple return values
411 ``-bb-vectorize``: Basic-Block Vectorization
412 --------------------------------------------
414 This pass combines instructions inside basic blocks to form vector
415 instructions. It iterates over each basic block, attempting to pair compatible
416 instructions, repeating this process until no additional pairs are selected for
417 vectorization. When the outputs of some pair of compatible instructions are
418 used as inputs by some other pair of compatible instructions, those pairs are
419 part of a potential vectorization chain. Instruction pairs are only fused into
420 vector instructions when they are part of a chain longer than some threshold
421 length. Moreover, the pass attempts to find the best possible chain for each
422 pair of compatible instructions. These heuristics are intended to prevent
423 vectorization in cases where it would not yield a performance increase of the
426 ``-block-placement``: Profile Guided Basic Block Placement
427 ----------------------------------------------------------
429 This pass is a very simple profile guided basic block placement algorithm. The
430 idea is to put frequently executed blocks together at the start of the function
431 and hopefully increase the number of fall-through conditional branches. If
432 there is no profile information for a particular function, this pass basically
433 orders blocks in depth-first order.
435 ``-break-crit-edges``: Break critical edges in CFG
436 --------------------------------------------------
438 Break all of the critical edges in the CFG by inserting a dummy basic block.
439 It may be "required" by passes that cannot deal with critical edges. This
440 transformation obviously invalidates the CFG, but can update forward dominator
441 (set, immediate dominators, tree, and frontier) information.
443 ``-codegenprepare``: Optimize for code generation
444 -------------------------------------------------
446 This pass munges the code in the input function to better prepare it for
447 SelectionDAG-based code generation. This works around limitations in its
448 basic-block-at-a-time approach. It should eventually be removed.
450 ``-constmerge``: Merge Duplicate Global Constants
451 -------------------------------------------------
453 Merges duplicate global constants together into a single constant that is
454 shared. This is useful because some passes (i.e., TraceValues) insert a lot of
455 string constants into the program, regardless of whether or not an existing
458 ``-constprop``: Simple constant propagation
459 -------------------------------------------
461 This pass implements constant propagation and merging. It looks for
462 instructions involving only constant operands and replaces them with a constant
463 value instead of an instruction. For example:
475 NOTE: this pass has a habit of making definitions be dead. It is a good idea
476 to run a :ref:`Dead Instruction Elimination <passes-die>` pass sometime after
481 ``-dce``: Dead Code Elimination
482 -------------------------------
484 Dead code elimination is similar to :ref:`dead instruction elimination
485 <passes-die>`, but it rechecks instructions that were used by removed
486 instructions to see if they are newly dead.
488 ``-deadargelim``: Dead Argument Elimination
489 -------------------------------------------
491 This pass deletes dead arguments from internal functions. Dead argument
492 elimination removes arguments which are directly dead, as well as arguments
493 only passed into function calls as dead arguments of other functions. This
494 pass also deletes dead arguments in a similar way.
496 This pass is often useful as a cleanup pass to run after aggressive
497 interprocedural passes, which add possibly-dead arguments.
499 ``-deadtypeelim``: Dead Type Elimination
500 ----------------------------------------
502 This pass is used to cleanup the output of GCC. It eliminate names for types
503 that are unused in the entire translation unit, using the :ref:`find used types
504 <passes-print-used-types>` pass.
508 ``-die``: Dead Instruction Elimination
509 --------------------------------------
511 Dead instruction elimination performs a single pass over the function, removing
512 instructions that are obviously dead.
514 ``-dse``: Dead Store Elimination
515 --------------------------------
517 A trivial dead store elimination that only considers basic-block local
520 .. _passes-functionattrs:
522 ``-functionattrs``: Deduce function attributes
523 ----------------------------------------------
525 A simple interprocedural pass which walks the call-graph, looking for functions
526 which do not access or only read non-local memory, and marking them
527 ``readnone``/``readonly``. In addition, it marks function arguments (of
528 pointer type) "``nocapture``" if a call to the function does not create any
529 copies of the pointer value that outlive the call. This more or less means
530 that the pointer is only dereferenced, and not returned from the function or
531 stored in a global. This pass is implemented as a bottom-up traversal of the
534 ``-globaldce``: Dead Global Elimination
535 ---------------------------------------
537 This transform is designed to eliminate unreachable internal globals from the
538 program. It uses an aggressive algorithm, searching out globals that are known
539 to be alive. After it finds all of the globals which are needed, it deletes
540 whatever is left over. This allows it to delete recursive chunks of the
541 program which are unreachable.
543 ``-globalopt``: Global Variable Optimizer
544 -----------------------------------------
546 This pass transforms simple global variables that never have their address
547 taken. If obviously true, it marks read/write globals as constant, deletes
548 variables only stored to, etc.
550 ``-gvn``: Global Value Numbering
551 --------------------------------
553 This pass performs global value numbering to eliminate fully and partially
554 redundant instructions. It also performs redundant load elimination.
558 ``-indvars``: Canonicalize Induction Variables
559 ----------------------------------------------
561 This transformation analyzes and transforms the induction variables (and
562 computations derived from them) into simpler forms suitable for subsequent
563 analysis and transformation.
565 This transformation makes the following changes to each loop with an
566 identifiable induction variable:
568 * All loops are transformed to have a *single* canonical induction variable
569 which starts at zero and steps by one.
570 * The canonical induction variable is guaranteed to be the first PHI node in
571 the loop header block.
572 * Any pointer arithmetic recurrences are raised to use array subscripts.
574 If the trip count of a loop is computable, this pass also makes the following
577 * The exit condition for the loop is canonicalized to compare the induction
578 value against the exit value. This turns loops like:
582 for (i = 7; i*i < 1000; ++i)
588 for (i = 0; i != 25; ++i)
590 * Any use outside of the loop of an expression derived from the indvar is
591 changed to compute the derived value outside of the loop, eliminating the
592 dependence on the exit value of the induction variable. If the only purpose
593 of the loop is to compute the exit value of some derived expression, this
594 transformation will make the loop dead.
596 This transformation should be followed by strength reduction after all of the
597 desired loop transformations have been performed. Additionally, on targets
598 where it is profitable, the loop could be transformed to count down to zero
599 (the "do loop" optimization).
601 ``-inline``: Function Integration/Inlining
602 ------------------------------------------
604 Bottom-up inlining of functions into callees.
606 .. _passes-instcombine:
608 ``-instcombine``: Combine redundant instructions
609 ------------------------------------------------
611 Combine instructions to form fewer, simple instructions. This pass does not
612 modify the CFG. This pass is where algebraic simplification happens.
614 This pass combines things like:
627 This is a simple worklist driven algorithm.
629 This pass guarantees that the following canonicalizations are performed on the
632 #. If a binary operator has a constant operand, it is moved to the right-hand
634 #. Bitwise operators with constant operands are always grouped so that shifts
635 are performed first, then ``or``\ s, then ``and``\ s, then ``xor``\ s.
636 #. Compare instructions are converted from ``<``, ``>``, ``≤``, or ``≥`` to
637 ``=`` or ``≠`` if possible.
638 #. All ``cmp`` instructions on boolean values are replaced with logical
640 #. ``add X, X`` is represented as ``mul X, 2`` ⇒ ``shl X, 1``
641 #. Multiplies with a constant power-of-two argument are transformed into
645 This pass can also simplify calls to specific well-known function calls (e.g.
646 runtime library functions). For example, a call ``exit(3)`` that occurs within
647 the ``main()`` function can be transformed into simply ``return 3``. Whether or
648 not library calls are simplified is controlled by the
649 :ref:`-functionattrs <passes-functionattrs>` pass and LLVM's knowledge of
650 library calls on different targets.
652 ``-internalize``: Internalize Global Symbols
653 --------------------------------------------
655 This pass loops over all of the functions in the input module, looking for a
656 main function. If a main function is found, all other functions and all global
657 variables with initializers are marked as internal.
659 ``-ipconstprop``: Interprocedural constant propagation
660 ------------------------------------------------------
662 This pass implements an *extremely* simple interprocedural constant propagation
663 pass. It could certainly be improved in many different ways, like using a
664 worklist. This pass makes arguments dead, but does not remove them. The
665 existing dead argument elimination pass should be run after this to clean up
668 ``-ipsccp``: Interprocedural Sparse Conditional Constant Propagation
669 --------------------------------------------------------------------
671 An interprocedural variant of :ref:`Sparse Conditional Constant Propagation
674 ``-jump-threading``: Jump Threading
675 -----------------------------------
677 Jump threading tries to find distinct threads of control flow running through a
678 basic block. This pass looks at blocks that have multiple predecessors and
679 multiple successors. If one or more of the predecessors of the block can be
680 proven to always cause a jump to one of the successors, we forward the edge
681 from the predecessor to the successor by duplicating the contents of this
684 An example of when this can occur is code like this:
693 In this case, the unconditional branch at the end of the first if can be
694 revectored to the false side of the second if.
696 ``-lcssa``: Loop-Closed SSA Form Pass
697 -------------------------------------
699 This pass transforms loops by placing phi nodes at the end of the loops for all
700 values that are live across the loop boundary. For example, it turns the left
710 X3 = phi(X1, X2) X3 = phi(X1, X2)
711 ... = X3 + 4 X4 = phi(X3)
714 This is still valid LLVM; the extra phi nodes are purely redundant, and will be
715 trivially eliminated by ``InstCombine``. The major benefit of this
716 transformation is that it makes many other loop optimizations, such as
717 ``LoopUnswitch``\ ing, simpler.
721 ``-licm``: Loop Invariant Code Motion
722 -------------------------------------
724 This pass performs loop invariant code motion, attempting to remove as much
725 code from the body of a loop as possible. It does this by either hoisting code
726 into the preheader block, or by sinking code to the exit blocks if it is safe.
727 This pass also promotes must-aliased memory locations in the loop to live in
728 registers, thus hoisting and sinking "invariant" loads and stores.
730 This pass uses alias analysis for two purposes:
732 #. Moving loop invariant loads and calls out of loops. If we can determine
733 that a load or call inside of a loop never aliases anything stored to, we
734 can hoist it or sink it like any other instruction.
736 #. Scalar Promotion of Memory. If there is a store instruction inside of the
737 loop, we try to move the store to happen AFTER the loop instead of inside of
738 the loop. This can only happen if a few conditions are true:
740 #. The pointer stored through is loop invariant.
741 #. There are no stores or loads in the loop which *may* alias the pointer.
742 There are no calls in the loop which mod/ref the pointer.
744 If these conditions are true, we can promote the loads and stores in the
745 loop of the pointer to use a temporary alloca'd variable. We then use the
746 :ref:`mem2reg <passes-mem2reg>` functionality to construct the appropriate
747 SSA form for the variable.
749 ``-loop-deletion``: Delete dead loops
750 -------------------------------------
752 This file implements the Dead Loop Deletion Pass. This pass is responsible for
753 eliminating loops with non-infinite computable trip counts that have no side
754 effects or volatile instructions, and do not contribute to the computation of
755 the function's return value.
757 .. _passes-loop-extract:
759 ``-loop-extract``: Extract loops into new functions
760 ---------------------------------------------------
762 A pass wrapper around the ``ExtractLoop()`` scalar transformation to extract
763 each top-level loop into its own new function. If the loop is the *only* loop
764 in a given function, it is not touched. This is a pass most useful for
765 debugging via bugpoint.
767 ``-loop-extract-single``: Extract at most one loop into a new function
768 ----------------------------------------------------------------------
770 Similar to :ref:`Extract loops into new functions <passes-loop-extract>`, this
771 pass extracts one natural loop from the program into a function if it can.
772 This is used by :program:`bugpoint`.
774 ``-loop-reduce``: Loop Strength Reduction
775 -----------------------------------------
777 This pass performs a strength reduction on array references inside loops that
778 have as one or more of their components the loop induction variable. This is
779 accomplished by creating a new value to hold the initial value of the array
780 access for the first iteration, and then creating a new GEP instruction in the
781 loop to increment the value by the appropriate amount.
783 ``-loop-rotate``: Rotate Loops
784 ------------------------------
786 A simple loop rotation transformation.
788 ``-loop-simplify``: Canonicalize natural loops
789 ----------------------------------------------
791 This pass performs several transformations to transform natural loops into a
792 simpler form, which makes subsequent analyses and transformations simpler and
795 Loop pre-header insertion guarantees that there is a single, non-critical entry
796 edge from outside of the loop to the loop header. This simplifies a number of
797 analyses and transformations, such as :ref:`LICM <passes-licm>`.
799 Loop exit-block insertion guarantees that all exit blocks from the loop (blocks
800 which are outside of the loop that have predecessors inside of the loop) only
801 have predecessors from inside of the loop (and are thus dominated by the loop
802 header). This simplifies transformations such as store-sinking that are built
805 This pass also guarantees that loops will have exactly one backedge.
807 Note that the :ref:`simplifycfg <passes-simplifycfg>` pass will clean up blocks
808 which are split out but end up being unnecessary, so usage of this pass should
809 not pessimize generated code.
811 This pass obviously modifies the CFG, but updates loop information and
812 dominator information.
814 ``-loop-unroll``: Unroll loops
815 ------------------------------
817 This pass implements a simple loop unroller. It works best when loops have
818 been canonicalized by the :ref:`indvars <passes-indvars>` pass, allowing it to
819 determine the trip counts of loops easily.
821 ``-loop-unswitch``: Unswitch loops
822 ----------------------------------
824 This pass transforms loops that contain branches on loop-invariant conditions
825 to have multiple loops. For example, it turns the left into the right code:
836 This can increase the size of the code exponentially (doubling it every time a
837 loop is unswitched) so we only unswitch if the resultant code will be smaller
840 This pass expects :ref:`LICM <passes-licm>` to be run before it to hoist
841 invariant conditions out of the loop, to make the unswitching opportunity
844 ``-loweratomic``: Lower atomic intrinsics to non-atomic form
845 ------------------------------------------------------------
847 This pass lowers atomic intrinsics to non-atomic form for use in a known
848 non-preemptible environment.
850 The pass does not verify that the environment is non-preemptible (in general
851 this would require knowledge of the entire call graph of the program including
852 any libraries which may not be available in bitcode form); it simply lowers
853 every atomic intrinsic.
855 ``-lowerinvoke``: Lower invokes to calls, for unwindless code generators
856 ------------------------------------------------------------------------
858 This transformation is designed for use by code generators which do not yet
859 support stack unwinding. This pass converts ``invoke`` instructions to
860 ``call`` instructions, so that any exception-handling ``landingpad`` blocks
861 become dead code (which can be removed by running the ``-simplifycfg`` pass
864 ``-lowerswitch``: Lower ``SwitchInst``\ s to branches
865 -----------------------------------------------------
867 Rewrites switch instructions with a sequence of branches, which allows targets
868 to get away with not implementing the switch instruction until it is
873 ``-mem2reg``: Promote Memory to Register
874 ----------------------------------------
876 This file promotes memory references to be register references. It promotes
877 alloca instructions which only have loads and stores as uses. An ``alloca`` is
878 transformed by using dominator frontiers to place phi nodes, then traversing
879 the function in depth-first order to rewrite loads and stores as appropriate.
880 This is just the standard SSA construction algorithm to construct "pruned" SSA
883 ``-memcpyopt``: MemCpy Optimization
884 -----------------------------------
886 This pass performs various transformations related to eliminating ``memcpy``
887 calls, or transforming sets of stores into ``memset``\ s.
889 ``-mergefunc``: Merge Functions
890 -------------------------------
892 This pass looks for equivalent functions that are mergable and folds them.
894 Total-ordering is introduced among the functions set: we define comparison
895 that answers for every two functions which of them is greater. It allows to
896 arrange functions into the binary tree.
898 For every new function we check for equivalent in tree.
900 If equivalent exists we fold such functions. If both functions are overridable,
901 we move the functionality into a new internal function and leave two
902 overridable thunks to it.
904 If there is no equivalent, then we add this function to tree.
906 Lookup routine has O(log(n)) complexity, while whole merging process has
907 complexity of O(n*log(n)).
910 :doc:`this <MergeFunctions>`
911 article for more details.
913 ``-mergereturn``: Unify function exit nodes
914 -------------------------------------------
916 Ensure that functions have at most one ``ret`` instruction in them.
917 Additionally, it keeps track of which node is the new exit node of the CFG.
919 ``-partial-inliner``: Partial Inliner
920 -------------------------------------
922 This pass performs partial inlining, typically by inlining an ``if`` statement
923 that surrounds the body of the function.
925 ``-prune-eh``: Remove unused exception handling info
926 ----------------------------------------------------
928 This file implements a simple interprocedural pass which walks the call-graph,
929 turning invoke instructions into call instructions if and only if the callee
930 cannot throw an exception. It implements this as a bottom-up traversal of the
933 ``-reassociate``: Reassociate expressions
934 -----------------------------------------
936 This pass reassociates commutative expressions in an order that is designed to
937 promote better constant propagation, GCSE, :ref:`LICM <passes-licm>`, PRE, etc.
939 For example: 4 + (x + 5) ⇒ x + (4 + 5)
941 In the implementation of this algorithm, constants are assigned rank = 0,
942 function arguments are rank = 1, and other values are assigned ranks
943 corresponding to the reverse post order traversal of current function (starting
944 at 2), which effectively gives values in deep loops higher rank than values not
947 ``-reg2mem``: Demote all values to stack slots
948 ----------------------------------------------
950 This file demotes all registers to memory references. It is intended to be the
951 inverse of :ref:`mem2reg <passes-mem2reg>`. By converting to ``load``
952 instructions, the only values live across basic blocks are ``alloca``
953 instructions and ``load`` instructions before ``phi`` nodes. It is intended
954 that this should make CFG hacking much easier. To make later hacking easier,
955 the entry block is split into two, such that all introduced ``alloca``
956 instructions (and nothing else) are in the entry block.
958 ``-scalarrepl``: Scalar Replacement of Aggregates (DT)
959 ------------------------------------------------------
961 The well-known scalar replacement of aggregates transformation. This transform
962 breaks up ``alloca`` instructions of aggregate type (structure or array) into
963 individual ``alloca`` instructions for each member if possible. Then, if
964 possible, it transforms the individual ``alloca`` instructions into nice clean
967 This combines a simple scalar replacement of aggregates algorithm with the
968 :ref:`mem2reg <passes-mem2reg>` algorithm because they often interact,
969 especially for C++ programs. As such, iterating between ``scalarrepl``, then
970 :ref:`mem2reg <passes-mem2reg>` until we run out of things to promote works
975 ``-sccp``: Sparse Conditional Constant Propagation
976 --------------------------------------------------
978 Sparse conditional constant propagation and merging, which can be summarized
981 * Assumes values are constant unless proven otherwise
982 * Assumes BasicBlocks are dead unless proven otherwise
983 * Proves values to be constant, and replaces them with constants
984 * Proves conditional branches to be unconditional
986 Note that this pass has a habit of making definitions be dead. It is a good
987 idea to run a :ref:`DCE <passes-dce>` pass sometime after running this pass.
989 .. _passes-simplifycfg:
991 ``-simplifycfg``: Simplify the CFG
992 ----------------------------------
994 Performs dead code elimination and basic block merging. Specifically:
996 * Removes basic blocks with no predecessors.
997 * Merges a basic block into its predecessor if there is only one and the
998 predecessor only has one successor.
999 * Eliminates PHI nodes for basic blocks with a single predecessor.
1000 * Eliminates a basic block that only contains an unconditional branch.
1002 ``-sink``: Code sinking
1003 -----------------------
1005 This pass moves instructions into successor blocks, when possible, so that they
1006 aren't executed on paths where their results aren't needed.
1008 ``-strip``: Strip all symbols from a module
1009 -------------------------------------------
1011 Performs code stripping. This transformation can delete:
1013 * names for virtual registers
1014 * symbols for internal globals and functions
1017 Note that this transformation makes code much less readable, so it should only
1018 be used in situations where the strip utility would be used, such as reducing
1019 code size or making it harder to reverse engineer code.
1021 ``-strip-dead-debug-info``: Strip debug info for unused symbols
1022 ---------------------------------------------------------------
1024 .. FIXME: this description is the same as for -strip
1026 performs code stripping. this transformation can delete:
1028 * names for virtual registers
1029 * symbols for internal globals and functions
1032 note that this transformation makes code much less readable, so it should only
1033 be used in situations where the strip utility would be used, such as reducing
1034 code size or making it harder to reverse engineer code.
1036 ``-strip-dead-prototypes``: Strip Unused Function Prototypes
1037 ------------------------------------------------------------
1039 This pass loops over all of the functions in the input module, looking for dead
1040 declarations and removes them. Dead declarations are declarations of functions
1041 for which no implementation is available (i.e., declarations for unused library
1044 ``-strip-debug-declare``: Strip all ``llvm.dbg.declare`` intrinsics
1045 -------------------------------------------------------------------
1047 .. FIXME: this description is the same as for -strip
1049 This pass implements code stripping. Specifically, it can delete:
1051 #. names for virtual registers
1052 #. symbols for internal globals and functions
1053 #. debug information
1055 Note that this transformation makes code much less readable, so it should only
1056 be used in situations where the 'strip' utility would be used, such as reducing
1057 code size or making it harder to reverse engineer code.
1059 ``-strip-nondebug``: Strip all symbols, except dbg symbols, from a module
1060 -------------------------------------------------------------------------
1062 .. FIXME: this description is the same as for -strip
1064 This pass implements code stripping. Specifically, it can delete:
1066 #. names for virtual registers
1067 #. symbols for internal globals and functions
1068 #. debug information
1070 Note that this transformation makes code much less readable, so it should only
1071 be used in situations where the 'strip' utility would be used, such as reducing
1072 code size or making it harder to reverse engineer code.
1074 ``-tailcallelim``: Tail Call Elimination
1075 ----------------------------------------
1077 This file transforms calls of the current function (self recursion) followed by
1078 a return instruction with a branch to the entry of the function, creating a
1079 loop. This pass also implements the following extensions to the basic
1082 #. Trivial instructions between the call and return do not prevent the
1083 transformation from taking place, though currently the analysis cannot
1084 support moving any really useful instructions (only dead ones).
1085 #. This pass transforms functions that are prevented from being tail recursive
1086 by an associative expression to use an accumulator variable, thus compiling
1087 the typical naive factorial or fib implementation into efficient code.
1088 #. TRE is performed if the function returns void, if the return returns the
1089 result returned by the call, or if the function returns a run-time constant
1090 on all exits from the function. It is possible, though unlikely, that the
1091 return returns something else (like constant 0), and can still be TRE'd. It
1092 can be TRE'd if *all other* return instructions in the function return the
1094 #. If it can prove that callees do not access theier caller stack frame, they
1095 are marked as eligible for tail call elimination (by the code generator).
1100 This section describes the LLVM Utility Passes.
1102 ``-deadarghaX0r``: Dead Argument Hacking (BUGPOINT USE ONLY; DO NOT USE)
1103 ------------------------------------------------------------------------
1105 Same as dead argument elimination, but deletes arguments to functions which are
1106 external. This is only for use by :doc:`bugpoint <Bugpoint>`.
1108 ``-extract-blocks``: Extract Basic Blocks From Module (for bugpoint use)
1109 ------------------------------------------------------------------------
1111 This pass is used by bugpoint to extract all blocks from the module into their
1114 ``-instnamer``: Assign names to anonymous instructions
1115 ------------------------------------------------------
1117 This is a little utility pass that gives instructions names, this is mostly
1118 useful when diffing the effect of an optimization because deleting an unnamed
1119 instruction can change all other instruction numbering, making the diff very
1124 ``-verify``: Module Verifier
1125 ----------------------------
1127 Verifies an LLVM IR code. This is useful to run after an optimization which is
1128 undergoing testing. Note that llvm-as verifies its input before emitting
1129 bitcode, and also that malformed bitcode is likely to make LLVM crash. All
1130 language front-ends are therefore encouraged to verify their output before
1131 performing optimizing transformations.
1133 #. Both of a binary operator's parameters are of the same type.
1134 #. Verify that the indices of mem access instructions match other operands.
1135 #. Verify that arithmetic and other things are only performed on first-class
1136 types. Verify that shifts and logicals only happen on integrals f.e.
1137 #. All of the constants in a switch statement are of the correct type.
1138 #. The code is in valid SSA form.
1139 #. It is illegal to put a label into any other type (like a structure) or to
1141 #. Only phi nodes can be self referential: ``%x = add i32 %x``, ``%x`` is
1143 #. PHI nodes must have an entry for each predecessor, with no extras.
1144 #. PHI nodes must be the first thing in a basic block, all grouped together.
1145 #. PHI nodes must have at least one entry.
1146 #. All basic blocks should only end with terminator insts, not contain them.
1147 #. The entry node to a function must not have predecessors.
1148 #. All Instructions must be embedded into a basic block.
1149 #. Functions cannot take a void-typed parameter.
1150 #. Verify that a function's argument list agrees with its declared type.
1151 #. It is illegal to specify a name for a void value.
1152 #. It is illegal to have an internal global value with no initializer.
1153 #. It is illegal to have a ``ret`` instruction that returns a value that does
1154 not agree with the function return value type.
1155 #. Function call argument types match the function prototype.
1156 #. All other things that are tested by asserts spread about the code.
1158 Note that this does not provide full security verification (like Java), but
1159 instead just tries to ensure that code is well-formed.
1161 ``-view-cfg``: View CFG of function
1162 -----------------------------------
1164 Displays the control flow graph using the GraphViz tool.
1166 ``-view-cfg-only``: View CFG of function (with no function bodies)
1167 ------------------------------------------------------------------
1169 Displays the control flow graph using the GraphViz tool, but omitting function
1172 ``-view-dom``: View dominance tree of function
1173 ----------------------------------------------
1175 Displays the dominator tree using the GraphViz tool.
1177 ``-view-dom-only``: View dominance tree of function (with no function bodies)
1178 -----------------------------------------------------------------------------
1180 Displays the dominator tree using the GraphViz tool, but omitting function
1183 ``-view-postdom``: View postdominance tree of function
1184 ------------------------------------------------------
1186 Displays the post dominator tree using the GraphViz tool.
1188 ``-view-postdom-only``: View postdominance tree of function (with no function bodies)
1189 -------------------------------------------------------------------------------------
1191 Displays the post dominator tree using the GraphViz tool, but omitting function