I'm measuring how optimization flag number affects the coverage of GCC when it is compiling some test cases. I know GCC has predefined optimization levels, such as -O1 and -O2. I also know how to obtain the list of concrete flags enabled by a predefined level by this command:
$ gcc -O2 -Q --help=optimizer
The following options control optimizations:
-O<number>
-Ofast
-Og
-Os
-faggressive-loop-optimizations [enabled]
-falign-functions [enabled]
-falign-functions=
-falign-jumps [enabled]
-falign-jumps=
-falign-labels [enabled]
-falign-labels=
-falign-loops [enabled]
-falign-loops=
-fallocation-dce [enabled]
-fallow-store-data-races [disabled]
-fassociative-math [disabled]
-fassume-phsa [available in BRIG]
-fasynchronous-unwind-tables [enabled]
-fauto-inc-dec [enabled]
-fbranch-count-reg [enabled]
-fbranch-probabilities [disabled]
-fcaller-saves [enabled]
-fcode-hoisting [enabled]
-fcombine-stack-adjustments [enabled]
-fcompare-elim [enabled]
-fconserve-stack [disabled]
-fcprop-registers [enabled]
-fcrossjumping [enabled]
-fcse-follow-jumps [enabled]
-fcx-fortran-rules [disabled]
-fcx-limited-range [disabled]
-fdce [enabled]
-fdefer-pop [enabled]
-fdelayed-branch [disabled]
-fdelete-dead-exceptions [disabled]
-fdelete-null-pointer-checks [enabled]
-fdevirtualize [enabled]
-fdevirtualize-speculatively [enabled]
-fdse [enabled]
-fearly-inlining [enabled]
-fexceptions [disabled]
-fexcess-precision=[fast|standard] [default]
-fexpensive-optimizations [enabled]
-ffast-math
-ffinite-loops [disabled]
-ffinite-math-only [disabled]
-ffloat-store [disabled]
-fforward-propagate [enabled]
-ffp-contract=[off|on|fast] fast
-ffp-int-builtin-inexact [enabled]
-ffunction-cse [enabled]
-fgcse [enabled]
-fgcse-after-reload [disabled]
-fgcse-las [disabled]
-fgcse-lm [enabled]
-fgcse-sm [disabled]
-fgraphite [disabled]
-fgraphite-identity [disabled]
-fguess-branch-probability [enabled]
-fhandle-exceptions -fexceptions
-fhoist-adjacent-loads [enabled]
-fif-conversion [enabled]
-fif-conversion2 [enabled]
-findirect-inlining [enabled]
-finline [enabled]
-finline-atomics [enabled]
-finline-functions [enabled]
-finline-functions-called-once [enabled]
-finline-small-functions [enabled]
-fipa-bit-cp [enabled]
-fipa-cp [enabled]
-fipa-cp-clone [disabled]
-fipa-icf [enabled]
-fipa-icf-functions [enabled]
-fipa-icf-variables [enabled]
-fipa-profile [enabled]
-fipa-pta [disabled]
-fipa-pure-const [enabled]
-fipa-ra [enabled]
-fipa-reference [enabled]
-fipa-reference-addressable [enabled]
-fipa-sra [enabled]
-fipa-stack-alignment [enabled]
-fipa-vrp [enabled]
-fira-algorithm=[CB|priority] CB
-fira-hoist-pressure [enabled]
-fira-loop-pressure [disabled]
-fira-region=[one|all|mixed] [default]
-fira-share-save-slots [enabled]
-fira-share-spill-slots [enabled]
-fisolate-erroneous-paths-attribute [disabled]
-fisolate-erroneous-paths-dereference [enabled]
-fivopts [enabled]
-fjump-tables [enabled]
-fkeep-gc-roots-live [disabled]
-flifetime-dse [enabled]
-flifetime-dse=<0,2> 2
-flimit-function-alignment [disabled]
-flive-patching -flive-
patching=inline-clone
-flive-patching=[inline-only-static|inline-clone] [default]
-flive-range-shrinkage [disabled]
-floop-interchange [disabled]
-floop-nest-optimize [disabled]
-floop-parallelize-all [disabled]
-floop-unroll-and-jam [disabled]
-flra-remat [enabled]
-fmath-errno [enabled]
-fmodulo-sched [disabled]
-fmodulo-sched-allow-regmoves [disabled]
-fmove-loop-invariants [enabled]
-fnon-call-exceptions [disabled]
-fnothrow-opt [available in C++,
ObjC++]
-fomit-frame-pointer [enabled]
-fopt-info [disabled]
-foptimize-sibling-calls [enabled]
-foptimize-strlen [enabled]
-fpack-struct [disabled]
-fpack-struct=<number>
-fpartial-inlining [enabled]
-fpatchable-function-entry=
-fpeel-loops [disabled]
-fpeephole [enabled]
-fpeephole2 [enabled]
-fplt [enabled]
-fpredictive-commoning [disabled]
-fprefetch-loop-arrays [enabled]
-fprintf-return-value [enabled]
-fprofile-partial-training [disabled]
-fprofile-reorder-functions [disabled]
-freciprocal-math [disabled]
-free [enabled]
-freg-struct-return [disabled]
-frename-registers [enabled]
-freorder-blocks [enabled]
-freorder-blocks-algorithm=[simple|stc] stc
-freorder-blocks-and-partition [enabled]
-freorder-functions [enabled]
-frerun-cse-after-loop [enabled]
-freschedule-modulo-scheduled-loops [disabled]
-frounding-math [disabled]
-frtti [available in C++,
D, ObjC++]
-fsave-optimization-record [disabled]
-fsched-critical-path-heuristic [enabled]
-fsched-dep-count-heuristic [enabled]
-fsched-group-heuristic [enabled]
-fsched-interblock [enabled]
-fsched-last-insn-heuristic [enabled]
-fsched-pressure [disabled]
-fsched-rank-heuristic [enabled]
-fsched-spec [enabled]
-fsched-spec-insn-heuristic [enabled]
-fsched-spec-load [disabled]
-fsched-spec-load-dangerous [disabled]
-fsched-stalled-insns [disabled]
-fsched-stalled-insns-dep [enabled]
-fsched-stalled-insns-dep=<number>
-fsched-stalled-insns=<number>
-fsched2-use-superblocks [disabled]
-fschedule-fusion [enabled]
-fschedule-insns [disabled]
-fschedule-insns2 [enabled]
-fsection-anchors [disabled]
-fsel-sched-pipelining [disabled]
-fsel-sched-pipelining-outer-loops [disabled]
-fsel-sched-reschedule-pipelined [disabled]
-fselective-scheduling [disabled]
-fselective-scheduling2 [disabled]
-fshort-enums [enabled]
-fshort-wchar [disabled]
-fshrink-wrap [enabled]
-fshrink-wrap-separate [enabled]
-fsignaling-nans [disabled]
-fsigned-zeros [enabled]
-fsimd-cost-model=[unlimited|dynamic|cheap] unlimited
-fsingle-precision-constant [disabled]
-fsplit-ivs-in-unroller [enabled]
-fsplit-loops [disabled]
-fsplit-paths [disabled]
-fsplit-wide-types [enabled]
-fsplit-wide-types-early [disabled]
-fssa-backprop [enabled]
-fssa-phiopt [enabled]
-fstack-check=[no|generic|specific]
-fstack-clash-protection [disabled]
-fstack-protector [disabled]
-fstack-protector-all [disabled]
-fstack-protector-explicit [disabled]
-fstack-protector-strong [disabled]
-fstack-reuse=[all|named_vars|none] all
-fstdarg-opt [enabled]
-fstore-merging [enabled]
-fstrict-aliasing [enabled]
-fstrict-enums [available in C++,
ObjC++]
-fstrict-volatile-bitfields [enabled]
-fthread-jumps [enabled]
-fno-threadsafe-statics [available in C++,
ObjC++]
-ftoplevel-reorder [enabled]
-ftracer [disabled]
-ftrapping-math [enabled]
-ftrapv [disabled]
-ftree-bit-ccp [enabled]
-ftree-builtin-call-dce [enabled]
-ftree-ccp [enabled]
-ftree-ch [enabled]
-ftree-coalesce-vars [enabled]
-ftree-copy-prop [enabled]
-ftree-cselim [enabled]
-ftree-dce [enabled]
-ftree-dominator-opts [enabled]
-ftree-dse [enabled]
-ftree-forwprop [enabled]
-ftree-fre [enabled]
-ftree-loop-distribute-patterns [enabled]
-ftree-loop-distribution [disabled]
-ftree-loop-if-convert [enabled]
-ftree-loop-im [enabled]
-ftree-loop-ivcanon [enabled]
-ftree-loop-optimize [enabled]
-ftree-loop-vectorize [disabled]
-ftree-lrs [disabled]
-ftree-parallelize-loops=<number> 1
-ftree-partial-pre [disabled]
-ftree-phiprop [enabled]
-ftree-pre [enabled]
-ftree-pta [enabled]
-ftree-reassoc [enabled]
-ftree-scev-cprop [enabled]
-ftree-sink [enabled]
-ftree-slp-vectorize [disabled]
-ftree-slsr [enabled]
-ftree-sra [enabled]
-ftree-switch-conversion [enabled]
-ftree-tail-merge [enabled]
-ftree-ter [enabled]
-ftree-vectorize
-ftree-vrp [enabled]
-funconstrained-commons [disabled]
-funroll-all-loops [disabled]
-funroll-loops [disabled]
-funsafe-math-optimizations [disabled]
-funswitch-loops [disabled]
-funwind-tables [disabled]
-fvar-tracking [enabled]
-fvar-tracking-assignments [enabled]
-fvar-tracking-assignments-toggle [disabled]
-fvar-tracking-uninit [disabled]
-fvariable-expansion-in-unroller [disabled]
-fvect-cost-model=[unlimited|dynamic|cheap] cheap
-fversion-loops-for-strides [disabled]
-fvpt [disabled]
-fweb [enabled]
-fwrapv [disabled]
-fwrapv-pointer [disabled]
I compiled some test cases with -O2, it turns out that 440 files and 167,035 lines of GCC source code were covered. But when I concatenated every enabled flag into a string S and replace -O2 with S when compiling the same test cases, I found only 390 files and 90,385 lines were covered.
Can someone tell me why? I suspect there is some mechanism in GCC that makes S and -O2 behave very different.
