1 # BOLT - a post-link optimizer developed to speed up large applications
5 `llvm-bolt <executable> [-o outputfile] <executable>.bolt [-data=perf.fdata] [options]`
17 Display available options (--help-hidden for more)
21 Display all available options
25 Display list of available options (--help-list-hidden for more)
27 - `--help-list-hidden`
29 Display list of all available options
33 Display the version of this program
39 Write bolt info section in the output binary
47 Save recorded profile to a file
49 ### BOLT generic options:
51 - `--align-text=<uint>`
53 Alignment of .text section
57 Allow processing of stripped binaries
59 - `--alt-inst-feature-size=<uint>`
61 Size of feature field in .altinstructions
63 - `--alt-inst-has-padlen`
65 Specify that .altinstructions has padlen field
67 - `--asm-dump[=<dump folder>]`
69 Dump function into assembly
75 - `--bolt-id=<string>`
77 Add any string to tag this execution in the output binary via bolt info section
79 - `--break-funcs=<func1,func2,func3,...>`
81 List of functions to core dump on (debugging)
85 Perform verification of LLVM instruction encoding/decoding. Every instruction
86 in the input is decoded and re-encoded. If the resulting bytes do not match
87 the input, a warning message is printed.
89 - `--comp-dir-override=<string>`
91 Overrides DW_AT_comp_dir, and provides an alternative base location, which is
92 used with DW_AT_dwo_name to construct a path to *.dwo files.
94 - `--create-debug-names-section`
96 Creates .debug_names section, if the input binary doesn't have it already, for
99 - `--cu-processing-batch-size=<uint>`
101 Specifies the size of batches for processing CUs. Higher number has better
102 performance, but more memory usage. Default value is 1.
112 - `--debug-skeleton-cu`
114 Prints out offsets for abbrev and debug_info of Skeleton CUs that get patched.
116 - `--debug-thread-count=<uint>`
118 Specifies the number of threads to be used when processing DWO debug information.
120 - `--dot-tooltip-code`
122 Add basic block instructions as tool tips on nodes
124 - `--dump-alt-instructions`
126 Dump Linux alternative instructions info
128 - `--dump-cg=<string>`
130 Dump callgraph to the given file
134 Dump parsed bolt data for debugging
138 Dump function CFGs to graphviz format after each stage;enable '-print-loops'
139 for color-coded blocks
141 - `--dump-linux-exceptions`
143 Dump Linux kernel exception table
147 Dump raw ORC unwind information (sorted)
149 - `--dump-para-sites`
151 Dump Linux kernel paravitual patch sites
153 - `--dump-pci-fixups`
155 Dump Linux kernel PCI fixup table
159 Dump Linux kernel SMP locks
161 - `--dump-static-calls`
163 Dump Linux kernel static calls
165 - `--dump-static-keys`
167 Dump Linux kernel static keys jump table
169 - `--dwarf-output-path=<string>`
171 Path to where .dwo files or dwp file will be written out to.
175 Path and name to DWP file.
179 Print execution info based on profile
183 Print dyno stats after each stage
185 - `--dyno-stats-scale=<uint>`
187 Scale to be applied while reporting dyno stats
191 Write BOLT Address Translation tables
193 - `--force-data-relocations`
195 Force relocations to data sections to always be processed
199 Force patching of original entry points
201 - `--funcs=<func1,func2,func3,...>`
203 Limit optimizations to functions from the list
205 - `--funcs-file=<string>`
207 File with list of functions to optimize
209 - `--funcs-file-no-regex=<string>`
211 File with list of functions to optimize (non-regex)
213 - `--funcs-no-regex=<func1,func2,func3,...>`
215 Limit optimizations to functions from the list (non-regex)
219 Hot data symbols support (relocation mode)
221 - `--hot-functions-at-end`
223 If reorder-functions is used, order functions putting hottest last
227 Generate hot text symbols. Apply this option to a precompiled binary that
228 manually calls into hugify, such that at runtime hugify call will put hot code
229 into 2M pages. This requires relocation.
231 - `--hot-text-move-sections=<sec1,sec2,sec3,...>`
233 List of sections containing functions used for hugifying hot text. BOLT makes
234 sure these functions are not placed on the same page as the hot text.
235 (default='.stub,.mover').
237 - `--insert-retpolines`
239 Run retpoline insertion pass
243 Keep or generate .debug_aranges section if .gdb_index is written
247 Preserve intermediate .o file
251 Skip processing of cold functions
253 - `--log-file=<string>`
255 Redirect journaling to a file instead of stdout/stderr
257 - `--long-jump-labels`
259 Always use long jumps/nops for Linux kernel static keys
261 - `--match-profile-with-function-hash`
263 Match profile with function hash
265 - `--max-data-relocations=<uint>`
267 Maximum number of data relocations to process
269 - `--max-funcs=<uint>`
271 Maximum number of functions to process
275 Use regular size pages for code alignment
279 Disable multithreading
281 - `--pad-funcs=<func1:pad1,func2:pad2,func3:pad3,...>`
283 List of functions to pad with amount of bytes
287 Print mappings in the legend, between characters/blocks and text sections
291 - `--profile-format=<value>`
293 Format to dump profile output in aggregation mode, default is fdata
294 - `fdata`: offset-based plaintext format
295 - `yaml`: dense YAML representation
297 - `--r11-availability=<value>`
299 Determine the availability of r11 before indirect branches
300 - `never`: r11 not available
301 - `always`: r11 available before calls and jumps
302 - `abi`: r11 available before calls but not before jumps
306 Use relocations in the binary (default=autodetect)
310 Remove .symtab section
312 - `--reorder-skip-symbols=<symbol1,symbol2,symbol3,...>`
314 List of symbol names that cannot be reordered
316 - `--reorder-symbols=<symbol1,symbol2,symbol3,...>`
318 List of symbol names that can be reordered
320 - `--retpoline-lfence`
322 Determine if lfence instruction should exist in the retpoline
324 - `--skip-funcs=<func1,func2,func3,...>`
326 List of functions to skip
328 - `--skip-funcs-file=<string>`
330 File with list of functions to skip
334 Trust the input to be from a well-formed source
336 - `--tasks-per-thread=<uint>`
338 Number of tasks to be created per thread
342 Assume that execution stops at trap instruction
344 - `--thread-count=<uint>`
348 - `--top-called-limit=<uint>`
350 Maximum number of functions to print in top called functions section
354 In relocation mode trap upon entry to any function that uses AVX-512
359 Insert traps in old function bodies (relocation mode)
361 - `--update-debug-sections`
363 Update DWARF debug sections of the executable
367 Use GNU_STACK program header for new segment (workaround for issues with
372 Re-use space in old .text if possible (relocation mode)
376 Set verbosity level for diagnostic output
380 Output a single dwarf package file (dwp) instead of multiple non-relocatable
381 dwarf object files (dwo).
383 ### BOLT optimization options:
389 - `--align-blocks-min-size=<uint>`
391 Minimal size of the basic block that should be aligned
393 - `--align-blocks-threshold=<uint>`
395 Align only blocks with frequency larger than containing function execution
396 frequency specified in percent. E.g. 1000 means aligning blocks that are 10
397 times more frequently executed than the containing function.
399 - `--align-functions=<uint>`
401 Align functions at a given value (relocation mode)
403 - `--align-functions-max-bytes=<uint>`
405 Maximum number of bytes to use to align functions
409 Assume the ABI is never violated
411 - `--block-alignment=<uint>`
413 Boundary to use for alignment of basic blocks
415 - `--bolt-seed=<uint>`
417 Seed for randomization
419 - `--cg-from-perf-data`
421 Use perf data directly when constructing the call graph for stale functions
423 - `--cg-ignore-recursive-calls`
425 Ignore recursive calls when constructing the call graph
427 - `--cg-use-split-hot-size`
429 Use hot/cold data on basic blocks to determine hot sizes for call graph
432 - `--cold-threshold=<uint>`
434 Tenths of percents of main entry frequency to use as a threshold when
435 evaluating whether a basic block is cold (0 means it is only considered cold
436 if the block has zero samples). Default: 0
438 - `--elim-link-veneers`
440 Run veneer elimination pass
442 - `--eliminate-unreachable`
444 Eliminate unreachable code
446 - `--equalize-bb-counts`
448 Use same count for BBs that should have equivalent count (used in non-LBR and
451 - `--execution-count-threshold=<uint>`
453 Perform profiling accuracy-sensitive optimizations only if function execution
454 count >= the threshold (default: 0)
456 - `--fix-block-counts`
458 Adjust block counts based on outgoing branch counts
460 - `--fix-func-counts`
462 Adjust function counts based on basic blocks execution count
464 - `--force-inline=<func1,func2,func3,...>`
466 List of functions to always consider for inlining
468 - `--frame-opt=<value>`
470 Optimize stack frame accesses
471 - `none`: do not perform frame optimization
472 - `hot`: perform FOP on hot functions
473 - `all`: perform FOP on all functions
475 - `--frame-opt-rm-stores`
477 Apply additional analysis to remove stores (experimental)
479 - `--function-order=<string>`
481 File containing an ordered list of functions to use for function reordering
483 - `--generate-function-order=<string>`
485 File to dump the ordered list of functions to use for function reordering
487 - `--generate-link-sections=<string>`
489 Generate a list of function sections in a format suitable for inclusion in a
494 Share stubs across functions
498 Automatically put hot code on 2MB page(s) (hugify) at runtime. No manual call
499 to hugify is needed in the binary (which is what --hot-text relies on).
503 Fold functions with identical code
507 Alias for --indirect-call-promotion
509 - `--icp-calls-remaining-percent-threshold=<uint>`
511 The percentage threshold against remaining unpromoted indirect call count for
512 the promotion for calls
516 Alias for --indirect-call-promotion-calls-topn
518 - `--icp-calls-total-percent-threshold=<uint>`
520 The percentage threshold against total count for the promotion for calls
522 - `--icp-eliminate-loads`
524 Enable load elimination using memory profiling data when performing ICP
526 - `--icp-funcs=<func1,func2,func3,...>`
528 List of functions to enable ICP for
532 Only promote call targets eligible for inlining
534 - `--icp-jt-remaining-percent-threshold=<uint>`
536 The percentage threshold against remaining unpromoted indirect call count for
537 the promotion for jump tables
541 Alias for --icp-jump-tables-targets
545 Alias for --indirect-call-promotion-jump-tables-topn
547 - `--icp-jt-total-percent-threshold=<uint>`
549 The percentage threshold against total count for the promotion for jump tables
551 - `--icp-jump-tables-targets`
553 For jump tables, optimize indirect jmp targets instead of indices
555 - `--icp-mp-threshold`
557 Alias for --indirect-call-promotion-mispredict-threshold
559 - `--icp-old-code-sequence`
561 Use old code sequence for promoted calls
563 - `--icp-top-callsites=<uint>`
565 Optimize hottest calls until at least this percentage of all indirect calls
566 frequency is covered. 0 = all callsites
570 Alias for --indirect-call-promotion-topn
574 Alias for --indirect-call-promotion-use-mispredicts
576 - `--indirect-call-promotion=<value>`
578 Indirect call promotion
579 - `none`: do not perform indirect call promotion
580 - `calls`: perform ICP on indirect calls
581 - `jump-tables`: perform ICP on jump tables
582 - `all`: perform ICP on calls and jump tables
584 - `--indirect-call-promotion-calls-topn=<uint>`
586 Limit number of targets to consider when doing indirect call promotion on
589 - `--indirect-call-promotion-jump-tables-topn=<uint>`
591 Limit number of targets to consider when doing indirect call promotion on jump
594 - `--indirect-call-promotion-topn=<uint>`
596 Limit number of targets to consider when doing indirect call promotion. 0 = no
599 - `--indirect-call-promotion-use-mispredicts`
601 Use misprediction frequency for determining whether or not ICP should be
602 applied at a callsite. The -indirect-call-promotion-mispredict-threshold
603 value will be used by this heuristic
605 - `--infer-fall-throughs`
607 Infer execution count for fall-through blocks
609 - `--infer-stale-profile`
611 Infer counts from stale profile data.
619 Adjust function profile after inlining
621 - `--inline-limit=<uint>`
623 Maximum number of call sites to inline
625 - `--inline-max-iters=<uint>`
627 Maximum number of inline iterations
631 Inline memcpy using 'rep movsb' instruction (X86-only)
633 - `--inline-small-functions`
635 Inline functions if increase in size is less than defined by -inline-small-
638 - `--inline-small-functions-bytes=<uint>`
640 Max number of bytes for the function to be considered small for inlining
645 Instrument code to generate accurate profile data
647 - `--iterative-guess`
649 In non-LBR mode, guess edge counts using iterative technique
651 - `--jt-footprint-optimize-for-icache`
653 With jt-footprint-reduction, only process PIC jumptables and turn off other
654 transformations that increase code size
656 - `--jt-footprint-reduction`
658 Make jump tables size smaller at the cost of using more instructions at jump
661 - `--jump-tables=<value>`
663 Jump tables support (default=basic)
664 - `none`: do not optimize functions with jump tables
665 - `basic`: optimize functions with jump tables
666 - `move`: move jump tables to a separate section
667 - `split`: split jump tables section into hot and cold based on function
669 - `aggressive`: aggressively split jump tables section based on usage of the
674 Keep no-op instructions. By default they are removed.
676 - `--lite-threshold-count=<uint>`
678 Similar to '-lite-threshold-pct' but specify threshold using absolute function
679 call count. I.e. limit processing to functions executed at least the specified
682 - `--lite-threshold-pct=<uint>`
684 Threshold (in percent) for selecting functions to process in lite mode. Higher
685 threshold means fewer functions to process. E.g threshold of 90 means only top
686 10 percent of functions with profile will be processed.
688 - `--match-with-call-graph`
690 Match functions with call graph
692 - `--memcpy1-spec=<func1,func2:cs1:cs2,func3:cs1,...>`
694 List of functions with call sites for which to specialize memcpy() for size 1
696 - `--min-branch-clusters`
698 Use a modified clustering algorithm geared towards minimizing branches
700 - `--name-similarity-function-matching-threshold=<uint>`
702 Match functions using namespace and edit distance.
706 Disable all inlining (overrides other inlining options)
710 Do not scan cold functions for external references (may result in slower binary)
712 - `--peepholes=<value>`
714 Enable peephole optimizations
715 - `none`: disable peepholes
716 - `double-jumps`: remove double jumps when able
717 - `tailcall-traps`: insert tail call traps
718 - `useless-branches`: remove useless conditional branches
719 - `all`: enable all peephole optimizations
723 Optimize PLT calls (requires linking with -znow)
724 - `none`: do not optimize PLT calls
725 - `hot`: optimize executed (hot) PLT calls
726 - `all`: optimize all PLT calls
728 - `--preserve-blocks-alignment`
730 Try to preserve basic block alignment
732 - `--profile-ignore-hash`
734 Ignore hash while reading function profile
736 - `--profile-use-dfs`
738 Use DFS order for YAML profile
742 Reassign registers so as to avoid using REX prefixes in hot code
744 - `--reorder-blocks=<value>`
746 Change layout of basic blocks in a function
747 - `none`: do not reorder basic blocks
748 - `reverse`: layout blocks in reverse order
749 - `normal`: perform optimal layout based on profile
750 - `branch-predictor`: perform optimal layout prioritizing branch predictions
751 - `cache`: perform optimal layout prioritizing I-cache behavior
752 - `cache+`: perform layout optimizing I-cache behavior
753 - `ext-tsp`: perform layout optimizing I-cache behavior
754 - `cluster-shuffle`: perform random layout of clusters
756 - `--reorder-data=<section1,section2,section3,...>`
758 List of sections to reorder
760 - `--reorder-data-algo=<value>`
762 Algorithm used to reorder data sections
763 - `count`: sort hot data by read counts
764 - `funcs`: sort hot data by hot function usage and count
766 - `--reorder-data-inplace`
768 Reorder data sections in place
770 - `--reorder-data-max-bytes=<uint>`
772 Maximum number of bytes to reorder
774 - `--reorder-data-max-symbols=<uint>`
776 Maximum number of symbols to reorder
778 - `--reorder-functions=<value>`
780 Reorder and cluster functions (works only with relocations)
781 - `none`: do not reorder functions
782 - `exec-count`: order by execution count
783 - `hfsort`: use hfsort algorithm
784 - `hfsort+`: use cache-directed sort
785 - `cdsort`: use cache-directed sort
786 - `pettis-hansen`: use Pettis-Hansen algorithm
787 - `random`: reorder functions randomly
788 - `user`: use function order specified by -function-order
790 - `--reorder-functions-use-hot-size`
792 Use a function's hot size when doing clustering
794 - `--report-bad-layout=<uint>`
796 Print top <uint> functions with suboptimal code layout on input
800 Print the list of functions with stale profile
802 - `--runtime-hugify-lib=<string>`
804 Specify file name of the runtime hugify library
806 - `--runtime-instrumentation-lib=<string>`
808 Specify file name of the runtime instrumentation library
810 - `--sctc-mode=<value>`
812 Mode for simplify conditional tail calls
813 - `always`: always perform sctc
814 - `preserve`: only perform sctc when branch direction is preserved
815 - `heuristic`: use branch prediction data to control sctc
817 - `--sequential-disassembly`
819 Performs disassembly sequentially
821 - `--shrink-wrapping-threshold=<uint>`
823 Percentage of prologue execution count to use as threshold when evaluating
824 whether a block is cold enough to be profitable to move eligible spills there
826 - `--simplify-conditional-tail-calls`
828 Simplify conditional tail calls by removing unnecessary jumps
830 - `--simplify-rodata-loads`
832 Simplify loads from read-only sections by replacing the memory operand with
833 the constant found in the corresponding section
835 - `--split-align-threshold=<uint>`
837 When deciding to split a function, apply this alignment while doing the size
838 comparison (see -split-threshold). Default value: 2.
842 Outline as many cold basic blocks as possible
846 Split C++ exception handling code
848 - `--split-functions`
850 Split functions into fragments
852 - `--split-strategy=<value>`
854 Strategy used to partition blocks into fragments
855 - `profile2`: split each function into a hot and cold fragment using profiling
857 - `cdsplit`: split each function into a hot, warm, and cold fragment using
858 profiling information
859 - `random2`: split each function into a hot and cold fragment at a randomly
860 chosen split point (ignoring any available profiling information)
861 - `randomN`: split each function into N fragments at a randomly chosen split
862 points (ignoring any available profiling information)
863 - `all`: split all basic blocks of each function into fragments such that each
864 fragment contains exactly a single basic block
866 - `--split-threshold=<uint>`
868 Split function only if its main size is reduced by more than given amount of
869 bytes. Default value: 0, i.e. split iff the size is reduced. Note that on some
870 architectures the size can increase after splitting.
872 - `--stale-matching-max-func-size=<uint>`
874 The maximum size of a function to consider for inference.
876 - `--stale-matching-min-matched-block=<uint>`
878 Percentage threshold of matched basic blocks at which stale profile inference
881 - `--stale-threshold=<uint>`
883 Maximum percentage of stale functions to tolerate (default: 100)
887 Turn on the stoke analysis
891 Strip 'repz' prefix from 'repz retq' sequence (on by default)
893 - `--tail-duplication=<value>`
895 Duplicate unconditional branches that cross a cache line
896 - `none`: do not apply
897 - `aggressive`: aggressive strategy
898 - `moderate`: moderate strategy
899 - `cache`: cache-aware duplication strategy
901 - `--tsp-threshold=<uint>`
903 Maximum number of hot basic blocks in a function for which to use a precise
904 TSP solution while re-ordering basic blocks
906 - `--use-aggr-reg-reassign`
908 Use register liveness analysis to try to find more opportunities for -reg-
909 reassign optimization
911 - `--use-compact-aligner`
913 Use compact approach for aligning functions
915 - `--use-edge-counts`
917 Use edge count data when doing clustering
921 Verify the CFG after every pass
923 - `--x86-align-branch-boundary-hot-only`
925 Only apply branch boundary alignment in hot code
927 - `--x86-strip-redundant-address-size`
929 Remove redundant Address-Size override prefix
931 ### BOLT options in relocation mode:
933 - `--align-macro-fusion=<value>`
935 Fix instruction alignment for macro-fusion (x86 relocation mode)
936 - `none`: do not insert alignment no-ops for macro-fusion
937 - `hot`: only insert alignment no-ops on hot execution paths (default)
938 - `all`: always align instructions to allow macro-fusion
940 ### BOLT instrumentation options:
942 `llvm-bolt <executable> -instrument [-o outputfile] <instrumented-executable>`
944 - `--conservative-instrumentation`
946 Disable instrumentation optimizations that sacrifice profile accuracy (for
947 debugging, default: false)
949 - `--instrument-calls`
951 Record profile for inter-function control flow activity (default: true)
953 - `--instrument-hot-only`
955 Only insert instrumentation on hot functions (needs profile, default: false)
957 - `--instrumentation-binpath=<string>`
959 Path to instrumented binary in case if /proc/self/map_files is not accessible
960 due to access restriction issues
962 - `--instrumentation-file=<string>`
964 File name where instrumented profile will be saved (default: /tmp/prof.fdata)
966 - `--instrumentation-file-append-pid`
968 Append PID to saved profile file name (default: false)
970 - `--instrumentation-no-counters-clear`
972 Don't clear counters across dumps (use with instrumentation-sleep-time option)
974 - `--instrumentation-sleep-time=<uint>`
976 Interval between profile writes (default: 0 = write only at program end).
977 This is useful for service workloads when you want to dump profile every X
978 minutes or if you are killing the program and the profile is not being dumped
981 - `--instrumentation-wait-forks`
983 Wait until all forks of instrumented process will finish (use with
984 instrumentation-sleep-time option)
986 ### BOLT printing options:
990 Print aliases when printing objects
994 Print functions after each stage
998 Print functions after CFG construction
1000 - `--print-debug-info`
1002 Print debug info when printing functions
1006 Print function after disassembly
1008 - `--print-dyno-opcode-stats=<uint>`
1010 Print per instruction opcode dyno stats and the functionnames:BB offsets of
1011 the nth highest execution counts
1013 - `--print-dyno-stats-only`
1015 While printing functions output dyno-stats and skip instructions
1017 - `--print-exceptions`
1019 Print exception handling data
1023 Print global symbols after disassembly
1025 - `--print-jump-tables`
1031 Print loop related information
1033 - `--print-mem-data`
1035 Print memory data annotations when printing functions
1037 - `--print-normalized`
1039 Print functions after CFG is normalized
1041 - `--print-only=<func1,func2,func3,...>`
1043 List of functions to print
1047 Print ORC unwind information for instructions
1051 Print functions after attaching profile
1053 - `--print-profile-stats`
1055 Print profile quality/bias analysis
1057 - `--print-pseudo-probes=<value>`
1059 Print pseudo probe info
1060 - `decode`: decode probes section from binary
1061 - `address_conversion`: update address2ProbesMap with output block address
1062 - `encoded_probes`: display the encoded probes in binary section
1063 - `all`: enable all debugging printout
1065 - `--print-relocations`
1067 Print relocations when printing functions/objects
1069 - `--print-reordered-data`
1071 Print section contents after reordering
1073 - `--print-retpoline-insertion`
1075 Print functions after retpoline insertion pass
1079 Print all SDT markers
1081 - `--print-sections`
1083 Print all registered sections
1087 Print names of functions with unknown control flow
1091 Print time spent constructing binary functions
1095 Print time spent in rewriting passes
1097 - `--print-after-branch-fixup`
1099 Print function after fixing local branches
1101 - `--print-after-jt-footprint-reduction`
1103 Print function after jt-footprint-reduction pass
1105 - `--print-after-lowering`
1107 Print function after instruction lowering
1109 - `--print-cache-metrics`
1111 Calculate and print various metrics for instruction cache
1113 - `--print-clusters`
1117 - `--print-estimate-edge-counts`
1119 Print function after edge counts are set for no-LBR profile
1121 - `--print-finalized`
1123 Print function after CFG is finalized
1125 - `--print-fix-relaxations`
1127 Print functions after fix relaxations pass
1129 - `--print-fix-riscv-calls`
1131 Print functions after fix RISCV calls pass
1135 Print functions after frame optimizer pass
1137 - `--print-function-statistics=<uint>`
1139 Print statistics about basic block ordering
1143 Print functions after ICF optimization
1147 Print functions after indirect call promotion
1151 Print functions after inlining optimization
1153 - `--print-large-functions`
1155 Print functions that could not be overwritten due to excessive size
1159 Print functions after longjmp pass
1161 - `--print-optimize-bodyless`
1163 Print functions after bodyless optimization
1165 - `--print-output-address-range`
1167 Print output address range for each basic block in the function
1168 whenBinaryFunction::print is called
1170 - `--print-peepholes`
1172 Print functions after peephole optimization
1176 Print functions after PLT optimization
1178 - `--print-regreassign`
1180 Print functions after regreassign pass
1182 - `--print-reordered`
1184 Print functions after layout optimization
1186 - `--print-reordered-functions`
1188 Print functions after clustering
1192 Print functions after conditional tail call simplification
1194 - `--print-simplify-rodata-loads`
1196 Print functions after simplification of RO data loads
1198 - `--print-sorted-by=<value>`
1200 Print functions sorted by order of dyno stats
1201 - `executed-forward-branches`: executed forward branches
1202 - `taken-forward-branches`: taken forward branches
1203 - `executed-backward-branches`: executed backward branches
1204 - `taken-backward-branches`: taken backward branches
1205 - `executed-unconditional-branches`: executed unconditional branches
1206 - `all-function-calls`: all function calls
1207 - `indirect-calls`: indirect calls
1208 - `PLT-calls`: PLT calls
1209 - `executed-instructions`: executed instructions
1210 - `executed-load-instructions`: executed load instructions
1211 - `executed-store-instructions`: executed store instructions
1212 - `taken-jump-table-branches`: taken jump table branches
1213 - `taken-unknown-indirect-branches`: taken unknown indirect branches
1214 - `total-branches`: total branches
1215 - `taken-branches`: taken branches
1216 - `non-taken-conditional-branches`: non-taken conditional branches
1217 - `taken-conditional-branches`: taken conditional branches
1218 - `all-conditional-branches`: all conditional branches
1219 - `linker-inserted-veneer-calls`: linker-inserted veneer calls
1220 - `all`: sorted by all names
1222 - `--print-sorted-by-order=<value>`
1224 Use ascending or descending order when printing functions ordered by dyno stats
1228 Print functions after code splitting
1232 Print functions after stoke analysis
1236 Print functions after unreachable code elimination
1238 - `--print-veneer-elimination`
1240 Print functions after veneer elimination pass
1244 Print time spent in each optimization
1246 - `--print-all-options`
1248 Print all option values after command line parsing
1252 Print non-default options after command line parsing