I recently enabled timeout detection for the risc-v fuzzer so I'm not sure how interesting this is. Seems like something weird is going on in scheduling that should be short circuited/bailed out. tip-of-tree > ./bin/riscv64-unknown-linux-gnu-gcc red.c -O1 -fschedule-insns -w -ftime-report Time variable usr sys wall GGC phase setup : 0.02 ( 0%) 0.00 ( 0%) 0.02 ( 0%) 2779k ( 1%) phase parsing : 0.72 ( 0%) 0.54 ( 38%) 1.35 ( 0%) 36M ( 13%) phase opt and generate : 917.85 (100%) 0.89 ( 62%) 925.29 (100%) 241M ( 86%) phase last asm : 0.07 ( 0%) 0.00 ( 0%) 0.06 ( 0%) 343k ( 0%) garbage collection : 0.31 ( 0%) 0.01 ( 1%) 0.37 ( 0%) 0 ( 0%) dump files : 0.01 ( 0%) 0.00 ( 0%) 0.00 ( 0%) 0 ( 0%) callgraph construction : 0.06 ( 0%) 0.00 ( 0%) 0.10 ( 0%) 8997k ( 3%) callgraph optimization : 0.02 ( 0%) 0.01 ( 1%) 0.03 ( 0%) 10k ( 0%) callgraph functions expansion : 915.19 (100%) 0.76 ( 53%) 922.36 (100%) 198M ( 71%) callgraph ipa passes : 2.45 ( 0%) 0.10 ( 7%) 2.68 ( 0%) 19M ( 7%) ipa function summary : 0.20 ( 0%) 0.00 ( 0%) 0.21 ( 0%) 5358k ( 2%) ipa dead code removal : 0.00 ( 0%) 0.00 ( 0%) 0.01 ( 0%) 0 ( 0%) ipa inlining heuristics : 0.01 ( 0%) 0.00 ( 0%) 0.00 ( 0%) 192 ( 0%) ipa pure const : 0.01 ( 0%) 0.00 ( 0%) 0.01 ( 0%) 1080 ( 0%) ipa modref : 0.01 ( 0%) 0.00 ( 0%) 0.01 ( 0%) 4336 ( 0%) cfg construction : 0.02 ( 0%) 0.00 ( 0%) 0.02 ( 0%) 28k ( 0%) cfg cleanup : 0.08 ( 0%) 0.00 ( 0%) 0.07 ( 0%) 2496 ( 0%) CFG verifier : 1.27 ( 0%) 0.01 ( 1%) 1.26 ( 0%) 0 ( 0%) trivially dead code : 0.11 ( 0%) 0.00 ( 0%) 0.12 ( 0%) 0 ( 0%) df scan insns : 0.20 ( 0%) 0.04 ( 3%) 0.26 ( 0%) 528 ( 0%) df reaching defs : 1.03 ( 0%) 0.03 ( 2%) 1.07 ( 0%) 0 ( 0%) df live regs : 0.75 ( 0%) 0.00 ( 0%) 0.80 ( 0%) 0 ( 0%) df live&initialized regs : 0.22 ( 0%) 0.00 ( 0%) 0.22 ( 0%) 0 ( 0%) df use-def / def-use chains : 0.00 ( 0%) 0.00 ( 0%) 0.03 ( 0%) 0 ( 0%) df reg dead/unused notes : 0.66 ( 0%) 0.00 ( 0%) 0.68 ( 0%) 7041k ( 2%) register information : 0.27 ( 0%) 0.00 ( 0%) 0.28 ( 0%) 0 ( 0%) alias analysis : 0.45 ( 0%) 0.01 ( 1%) 0.45 ( 0%) 14M ( 5%) alias stmt walking : 0.38 ( 0%) 0.03 ( 2%) 0.50 ( 0%) 3192 ( 0%) register scan : 0.03 ( 0%) 0.00 ( 0%) 0.05 ( 0%) 586k ( 0%) rebuild jump labels : 0.05 ( 0%) 0.00 ( 0%) 0.05 ( 0%) 0 ( 0%) preprocessing : 0.08 ( 0%) 0.09 ( 6%) 0.24 ( 0%) 1861k ( 1%) lexical analysis : 0.24 ( 0%) 0.16 ( 11%) 0.30 ( 0%) 0 ( 0%) parser (global) : 0.18 ( 0%) 0.12 ( 8%) 0.35 ( 0%) 10103k ( 4%) parser function body : 0.19 ( 0%) 0.17 ( 12%) 0.40 ( 0%) 24M ( 9%) inline parameters : 0.10 ( 0%) 0.00 ( 0%) 0.07 ( 0%) 128k ( 0%) tree gimplify : 0.04 ( 0%) 0.02 ( 1%) 0.05 ( 0%) 12M ( 5%) tree eh : 0.00 ( 0%) 0.01 ( 1%) 0.00 ( 0%) 10056 ( 0%) tree CFG construction : 0.01 ( 0%) 0.00 ( 0%) 0.02 ( 0%) 8660k ( 3%) tree CFG cleanup : 0.03 ( 0%) 0.00 ( 0%) 0.02 ( 0%) 20k ( 0%) tree copy propagation : 0.02 ( 0%) 0.00 ( 0%) 0.03 ( 0%) 200 ( 0%) tree PTA : 0.36 ( 0%) 0.06 ( 4%) 0.44 ( 0%) 2128k ( 1%) tree SSA other : 0.01 ( 0%) 0.00 ( 0%) 0.00 ( 0%) 0 ( 0%) tree SSA rewrite : 0.03 ( 0%) 0.00 ( 0%) 0.02 ( 0%) 2079k ( 1%) tree SSA incremental : 0.01 ( 0%) 0.00 ( 0%) 0.01 ( 0%) 196k ( 0%) tree operand scan : 0.00 ( 0%) 0.01 ( 1%) 0.02 ( 0%) 3695k ( 1%) dominator optimization : 0.42 ( 0%) 0.00 ( 0%) 0.46 ( 0%) 515k ( 0%) backwards jump threading : 0.00 ( 0%) 0.00 ( 0%) 0.01 ( 0%) 2256 ( 0%) tree CCP : 0.13 ( 0%) 0.01 ( 1%) 0.12 ( 0%) 751k ( 0%) tree FRE : 0.13 ( 0%) 0.00 ( 0%) 0.07 ( 0%) 659k ( 0%) tree code sinking : 0.00 ( 0%) 0.00 ( 0%) 0.02 ( 0%) 63k ( 0%) tree backward propagate : 0.01 ( 0%) 0.00 ( 0%) 0.00 ( 0%) 0 ( 0%) tree forward propagate : 0.03 ( 0%) 0.00 ( 0%) 0.05 ( 0%) 62k ( 0%) tree conservative DCE : 0.06 ( 0%) 0.01 ( 1%) 0.06 ( 0%) 0 ( 0%) tree aggressive DCE : 0.03 ( 0%) 0.01 ( 1%) 0.02 ( 0%) 0 ( 0%) tree DSE : 0.03 ( 0%) 0.00 ( 0%) 0.02 ( 0%) 16k ( 0%) tree loop invariant motion : 0.02 ( 0%) 0.00 ( 0%) 0.02 ( 0%) 2712 ( 0%) tree canonical iv : 0.02 ( 0%) 0.00 ( 0%) 0.02 ( 0%) 234k ( 0%) complete unrolling : 0.05 ( 0%) 0.00 ( 0%) 0.06 ( 0%) 977k ( 0%) tree iv optimization : 0.06 ( 0%) 0.00 ( 0%) 0.07 ( 0%) 2150k ( 1%) tree copy headers : 0.04 ( 0%) 0.00 ( 0%) 0.03 ( 0%) 684k ( 0%) tree SSA verifier : 1.39 ( 0%) 0.01 ( 1%) 1.45 ( 0%) 0 ( 0%) tree STMT verifier : 2.04 ( 0%) 0.03 ( 2%) 2.14 ( 0%) 0 ( 0%) tree strlen optimization : 0.02 ( 0%) 0.00 ( 0%) 0.03 ( 0%) 582k ( 0%) tree modref : 0.04 ( 0%) 0.00 ( 0%) 0.05 ( 0%) 11k ( 0%) callgraph verifier : 0.07 ( 0%) 0.00 ( 0%) 0.10 ( 0%) 0 ( 0%) dominance computation : 0.01 ( 0%) 0.00 ( 0%) 0.03 ( 0%) 0 ( 0%) out of ssa : 0.02 ( 0%) 0.00 ( 0%) 0.02 ( 0%) 376 ( 0%) expand vars : 0.01 ( 0%) 0.00 ( 0%) 0.02 ( 0%) 2102k ( 1%) expand : 0.29 ( 0%) 0.05 ( 3%) 0.37 ( 0%) 63M ( 22%) post expand cleanups : 0.02 ( 0%) 0.00 ( 0%) 0.02 ( 0%) 5184 ( 0%) lower subreg : 0.01 ( 0%) 0.00 ( 0%) 0.01 ( 0%) 0 ( 0%) forward prop : 0.25 ( 0%) 0.04 ( 3%) 0.29 ( 0%) 1448k ( 1%) CSE : 0.64 ( 0%) 0.01 ( 1%) 0.65 ( 0%) 1083k ( 0%) dead code elimination : 0.07 ( 0%) 0.00 ( 0%) 0.07 ( 0%) 0 ( 0%) dead store elim1 : 0.70 ( 0%) 0.01 ( 1%) 0.74 ( 0%) 2507k ( 1%) dead store elim2 : 0.61 ( 0%) 0.00 ( 0%) 0.61 ( 0%) 2385k ( 1%) loop init : 0.00 ( 0%) 0.00 ( 0%) 0.02 ( 0%) 1257k ( 0%) loop invariant motion : 0.15 ( 0%) 0.00 ( 0%) 0.13 ( 0%) 23k ( 0%) loop fini : 0.00 ( 0%) 0.00 ( 0%) 0.01 ( 0%) 0 ( 0%) auto inc dec : 0.06 ( 0%) 0.00 ( 0%) 0.06 ( 0%) 144 ( 0%) branch prediction : 0.01 ( 0%) 0.00 ( 0%) 0.03 ( 0%) 58k ( 0%) combiner : 1.12 ( 0%) 0.03 ( 2%) 1.17 ( 0%) 22M ( 8%) scheduling : 895.36 ( 97%) 0.19 ( 13%) 901.31 ( 97%) 15M ( 6%) integrated RA : 2.93 ( 0%) 0.21 ( 15%) 3.28 ( 0%) 52M ( 19%) LRA non-specific : 0.41 ( 0%) 0.01 ( 1%) 0.44 ( 0%) 218k ( 0%) LRA virtuals elimination : 0.11 ( 0%) 0.00 ( 0%) 0.10 ( 0%) 42k ( 0%) LRA reload inheritance : 0.03 ( 0%) 0.00 ( 0%) 0.03 ( 0%) 4816 ( 0%) LRA create live ranges : 0.12 ( 0%) 0.00 ( 0%) 0.12 ( 0%) 411k ( 0%) LRA hard reg assignment : 0.10 ( 0%) 0.00 ( 0%) 0.10 ( 0%) 0 ( 0%) reload : 0.02 ( 0%) 0.00 ( 0%) 0.02 ( 0%) 144 ( 0%) reload CSE regs : 0.18 ( 0%) 0.00 ( 0%) 0.17 ( 0%) 2470k ( 1%) thread pro- & epilogue : 0.27 ( 0%) 0.00 ( 0%) 0.26 ( 0%) 18k ( 0%) hard reg cprop : 0.21 ( 0%) 0.00 ( 0%) 0.21 ( 0%) 11k ( 0%) reorder blocks : 0.03 ( 0%) 0.00 ( 0%) 0.04 ( 0%) 89k ( 0%) shorten branches : 0.04 ( 0%) 0.01 ( 1%) 0.04 ( 0%) 0 ( 0%) final : 0.15 ( 0%) 0.00 ( 0%) 0.15 ( 0%) 385k ( 0%) straight-line strength reduction : 0.02 ( 0%) 0.00 ( 0%) 0.02 ( 0%) 234k ( 0%) initialize rtl : 0.01 ( 0%) 0.00 ( 0%) 0.00 ( 0%) 17k ( 0%) access analysis : 0.00 ( 0%) 0.00 ( 0%) 0.02 ( 0%) 512 ( 0%) rest of compilation : 0.21 ( 0%) 0.00 ( 0%) 0.24 ( 0%) 1531k ( 1%) remove unused locals : 0.02 ( 0%) 0.00 ( 0%) 0.03 ( 0%) 0 ( 0%) address taken : 0.03 ( 0%) 0.00 ( 0%) 0.03 ( 0%) 0 ( 0%) verify loop closed : 0.01 ( 0%) 0.00 ( 0%) 0.01 ( 0%) 0 ( 0%) verify RTL sharing : 2.30 ( 0%) 0.02 ( 1%) 2.40 ( 0%) 0 ( 0%) TOTAL : 918.66 1.43 926.72 280M Extra diagnostic checks enabled; compiler may run slowly. Configure with --enable-checking=release to disable checks. Bisected with a 2 minute timeout. First bad commit: r13-5154-g733a1b777f1 r13-5154-g733a1b777f1 > ./bin/riscv64-unknown-linux-gnu-gcc red.c -O1 -fschedule-insns -w -ftime-report; /work/patrick/notify.sh Time variable usr sys wall GGC phase setup : 0.00 ( 0%) 0.00 ( 0%) 0.01 ( 0%) 1539k ( 1%) phase parsing : 0.61 ( 0%) 0.51 ( 35%) 1.12 ( 0%) 36M ( 13%) phase opt and generate :1342.05 (100%) 0.93 ( 64%)1348.25 (100%) 246M ( 87%) phase last asm : 0.05 ( 0%) 0.01 ( 1%) 0.06 ( 0%) 89k ( 0%) garbage collection : 0.39 ( 0%) 0.01 ( 1%) 0.41 ( 0%) 0 ( 0%) callgraph construction : 0.12 ( 0%) 0.02 ( 1%) 0.09 ( 0%) 9000k ( 3%) callgraph optimization : 0.02 ( 0%) 0.01 ( 1%) 0.02 ( 0%) 4112 ( 0%) callgraph functions expansion :1339.07 (100%) 0.81 ( 56%)1345.12 (100%) 202M ( 71%) callgraph ipa passes : 2.68 ( 0%) 0.10 ( 7%) 2.81 ( 0%) 21M ( 7%) ipa function summary : 0.08 ( 0%) 0.01 ( 1%) 0.09 ( 0%) 7113k ( 2%) ipa dead code removal : 0.01 ( 0%) 0.00 ( 0%) 0.01 ( 0%) 0 ( 0%) ipa inlining heuristics : 0.01 ( 0%) 0.00 ( 0%) 0.01 ( 0%) 2536 ( 0%) ipa pure const : 0.01 ( 0%) 0.00 ( 0%) 0.00 ( 0%) 1080 ( 0%) ipa free inline summary : 0.00 ( 0%) 0.00 ( 0%) 0.01 ( 0%) 0 ( 0%) ipa modref : 0.01 ( 0%) 0.00 ( 0%) 0.02 ( 0%) 6256 ( 0%) cfg construction : 0.02 ( 0%) 0.00 ( 0%) 0.02 ( 0%) 28k ( 0%) cfg cleanup : 0.15 ( 0%) 0.00 ( 0%) 0.14 ( 0%) 2496 ( 0%) CFG verifier : 1.46 ( 0%) 0.01 ( 1%) 1.48 ( 0%) 0 ( 0%) trivially dead code : 0.12 ( 0%) 0.00 ( 0%) 0.12 ( 0%) 0 ( 0%) df scan insns : 0.18 ( 0%) 0.06 ( 4%) 0.24 ( 0%) 624 ( 0%) df reaching defs : 1.28 ( 0%) 0.02 ( 1%) 1.29 ( 0%) 0 ( 0%) df live regs : 0.87 ( 0%) 0.00 ( 0%) 0.83 ( 0%) 0 ( 0%) df live&initialized regs : 0.25 ( 0%) 0.00 ( 0%) 0.23 ( 0%) 0 ( 0%) df use-def / def-use chains : 0.01 ( 0%) 0.00 ( 0%) 0.02 ( 0%) 0 ( 0%) df reg dead/unused notes : 0.70 ( 0%) 0.00 ( 0%) 0.69 ( 0%) 6981k ( 2%) register information : 0.33 ( 0%) 0.00 ( 0%) 0.33 ( 0%) 0 ( 0%) alias analysis : 0.46 ( 0%) 0.00 ( 0%) 0.44 ( 0%) 14M ( 5%) alias stmt walking : 0.54 ( 0%) 0.02 ( 1%) 0.59 ( 0%) 3192 ( 0%) register scan : 0.02 ( 0%) 0.00 ( 0%) 0.03 ( 0%) 591k ( 0%) rebuild jump labels : 0.05 ( 0%) 0.00 ( 0%) 0.05 ( 0%) 0 ( 0%) preprocessing : 0.12 ( 0%) 0.10 ( 7%) 0.40 ( 0%) 1602k ( 1%) lexical analysis : 0.17 ( 0%) 0.20 ( 14%) 0.26 ( 0%) 0 ( 0%) parser (global) : 0.16 ( 0%) 0.09 ( 6%) 0.22 ( 0%) 10M ( 4%) parser function body : 0.11 ( 0%) 0.11 ( 8%) 0.16 ( 0%) 24M ( 9%) parser inl. func. body : 0.01 ( 0%) 0.00 ( 0%) 0.01 ( 0%) 24k ( 0%) inline parameters : 0.08 ( 0%) 0.00 ( 0%) 0.08 ( 0%) 128k ( 0%) tree gimplify : 0.05 ( 0%) 0.00 ( 0%) 0.08 ( 0%) 12M ( 5%) tree eh : 0.00 ( 0%) 0.00 ( 0%) 0.01 ( 0%) 10056 ( 0%) tree CFG construction : 0.01 ( 0%) 0.00 ( 0%) 0.01 ( 0%) 8669k ( 3%) tree CFG cleanup : 0.01 ( 0%) 0.00 ( 0%) 0.02 ( 0%) 25k ( 0%) tree copy propagation : 0.04 ( 0%) 0.00 ( 0%) 0.04 ( 0%) 200 ( 0%) tree PTA : 0.40 ( 0%) 0.07 ( 5%) 0.45 ( 0%) 2127k ( 1%) tree SSA rewrite : 0.02 ( 0%) 0.00 ( 0%) 0.03 ( 0%) 2079k ( 1%) tree SSA incremental : 0.02 ( 0%) 0.00 ( 0%) 0.01 ( 0%) 200k ( 0%) tree operand scan : 0.03 ( 0%) 0.03 ( 2%) 0.04 ( 0%) 3691k ( 1%) dominator optimization : 0.80 ( 0%) 0.00 ( 0%) 0.79 ( 0%) 2918k ( 1%) backwards jump threading : 0.01 ( 0%) 0.00 ( 0%) 0.01 ( 0%) 2560 ( 0%) tree CCP : 0.12 ( 0%) 0.00 ( 0%) 0.13 ( 0%) 734k ( 0%) tree reassociation : 0.01 ( 0%) 0.00 ( 0%) 0.01 ( 0%) 96 ( 0%) tree FRE : 0.11 ( 0%) 0.00 ( 0%) 0.12 ( 0%) 684k ( 0%) tree code sinking : 0.03 ( 0%) 0.00 ( 0%) 0.01 ( 0%) 63k ( 0%) tree linearize phis : 0.01 ( 0%) 0.00 ( 0%) 0.01 ( 0%) 11k ( 0%) tree forward propagate : 0.05 ( 0%) 0.00 ( 0%) 0.04 ( 0%) 67k ( 0%) tree conservative DCE : 0.02 ( 0%) 0.04 ( 3%) 0.06 ( 0%) 0 ( 0%) tree aggressive DCE : 0.04 ( 0%) 0.00 ( 0%) 0.08 ( 0%) 16k ( 0%) tree DSE : 0.02 ( 0%) 0.00 ( 0%) 0.03 ( 0%) 0 ( 0%) tree loop invariant motion : 0.03 ( 0%) 0.00 ( 0%) 0.03 ( 0%) 2616 ( 0%) tree canonical iv : 0.01 ( 0%) 0.00 ( 0%) 0.01 ( 0%) 236k ( 0%) complete unrolling : 0.06 ( 0%) 0.00 ( 0%) 0.06 ( 0%) 980k ( 0%) tree iv optimization : 0.05 ( 0%) 0.00 ( 0%) 0.05 ( 0%) 2196k ( 1%) tree SSA uncprop : 0.01 ( 0%) 0.00 ( 0%) 0.00 ( 0%) 0 ( 0%) tree SSA verifier : 1.82 ( 0%) 0.00 ( 0%) 1.77 ( 0%) 0 ( 0%) tree STMT verifier : 2.38 ( 0%) 0.00 ( 0%) 2.35 ( 0%) 0 ( 0%) tree strlen optimization : 0.03 ( 0%) 0.00 ( 0%) 0.03 ( 0%) 631k ( 0%) tree modref : 0.04 ( 0%) 0.00 ( 0%) 0.04 ( 0%) 12k ( 0%) callgraph verifier : 0.09 ( 0%) 0.00 ( 0%) 0.11 ( 0%) 0 ( 0%) dominance computation : 0.05 ( 0%) 0.00 ( 0%) 0.03 ( 0%) 0 ( 0%) out of ssa : 0.03 ( 0%) 0.00 ( 0%) 0.04 ( 0%) 480 ( 0%) expand vars : 0.01 ( 0%) 0.00 ( 0%) 0.01 ( 0%) 2133k ( 1%) expand : 0.29 ( 0%) 0.01 ( 1%) 0.31 ( 0%) 61M ( 22%) post expand cleanups : 0.02 ( 0%) 0.00 ( 0%) 0.02 ( 0%) 5656 ( 0%) varconst : 0.00 ( 0%) 0.00 ( 0%) 0.02 ( 0%) 88k ( 0%) lower subreg : 0.01 ( 0%) 0.00 ( 0%) 0.01 ( 0%) 0 ( 0%) forward prop : 0.22 ( 0%) 0.02 ( 1%) 0.24 ( 0%) 1193k ( 0%) CSE : 0.60 ( 0%) 0.01 ( 1%) 0.61 ( 0%) 863k ( 0%) dead code elimination : 0.08 ( 0%) 0.00 ( 0%) 0.09 ( 0%) 0 ( 0%) dead store elim1 : 0.76 ( 0%) 0.00 ( 0%) 0.77 ( 0%) 2501k ( 1%) dead store elim2 : 0.79 ( 0%) 0.00 ( 0%) 0.77 ( 0%) 2492k ( 1%) loop init : 0.02 ( 0%) 0.00 ( 0%) 0.01 ( 0%) 1305k ( 0%) loop invariant motion : 0.13 ( 0%) 0.00 ( 0%) 0.18 ( 0%) 34k ( 0%) loop fini : 0.00 ( 0%) 0.00 ( 0%) 0.01 ( 0%) 0 ( 0%) branch prediction : 0.00 ( 0%) 0.00 ( 0%) 0.01 ( 0%) 60k ( 0%) combiner : 0.94 ( 0%) 0.00 ( 0%) 0.94 ( 0%) 25M ( 9%) if-conversion : 0.00 ( 0%) 0.00 ( 0%) 0.01 ( 0%) 10k ( 0%) scheduling :1315.49 ( 98%) 0.37 ( 26%)1320.88 ( 98%) 15M ( 5%) integrated RA : 4.20 ( 0%) 0.22 ( 15%) 4.58 ( 0%) 53M ( 19%) LRA non-specific : 0.60 ( 0%) 0.00 ( 0%) 0.63 ( 0%) 245k ( 0%) LRA virtuals elimination : 0.20 ( 0%) 0.00 ( 0%) 0.20 ( 0%) 43k ( 0%) LRA reload inheritance : 0.05 ( 0%) 0.00 ( 0%) 0.05 ( 0%) 5040 ( 0%) LRA create live ranges : 0.19 ( 0%) 0.00 ( 0%) 0.19 ( 0%) 412k ( 0%) LRA hard reg assignment : 0.15 ( 0%) 0.00 ( 0%) 0.15 ( 0%) 0 ( 0%) reload : 0.02 ( 0%) 0.00 ( 0%) 0.02 ( 0%) 168 ( 0%) reload CSE regs : 0.18 ( 0%) 0.00 ( 0%) 0.19 ( 0%) 2494k ( 1%) thread pro- & epilogue : 0.28 ( 0%) 0.00 ( 0%) 0.28 ( 0%) 20k ( 0%) hard reg cprop : 0.21 ( 0%) 0.00 ( 0%) 0.20 ( 0%) 22k ( 0%) reorder blocks : 0.05 ( 0%) 0.00 ( 0%) 0.03 ( 0%) 99k ( 0%) shorten branches : 0.06 ( 0%) 0.00 ( 0%) 0.05 ( 0%) 0 ( 0%) final : 0.13 ( 0%) 0.00 ( 0%) 0.15 ( 0%) 624 ( 0%) straight-line strength reduction : 0.03 ( 0%) 0.00 ( 0%) 0.04 ( 0%) 234k ( 0%) access analysis : 0.01 ( 0%) 0.00 ( 0%) 0.02 ( 0%) 416 ( 0%) rest of compilation : 0.23 ( 0%) 0.01 ( 1%) 0.27 ( 0%) 2179k ( 1%) remove unused locals : 0.02 ( 0%) 0.00 ( 0%) 0.04 ( 0%) 0 ( 0%) address taken : 0.01 ( 0%) 0.00 ( 0%) 0.03 ( 0%) 0 ( 0%) verify RTL sharing : 2.54 ( 0%) 0.00 ( 0%) 2.57 ( 0%) 0 ( 0%) TOTAL :1342.71 1.45 1349.44 284M Extra diagnostic checks enabled; compiler may run slowly. Configure with --enable-checking=release to disable checks. I don't have the tooling to reliably reduce a slow compile issue so I've attached the unreduced preprocessed file I was testing with. Any tips for reducing a case like this would be appreciated. Testcase found via fuzzer.
Created attachment 57640 [details] Raw testcase and headers
Created attachment 57641 [details] unreduced preprocessed testcase
The first attachment is empty (perhaps you made a non-recursive archive when you meant to recursively zip a directory).
Created attachment 57642 [details] Raw testcase and headers Thanks for catching that, here's the zip file with actual files in it this time ;)
It appears sched-deps is O(N*M) given N reg_pending_barriers and M distinct pseudos in a region (or even a basic block). For instance, on the following testcase #define x10(x) x x x x x x x x x x #define x100(x) x10(x10(x)) #define x10000(x) x100(x100(x)) void f(int); void g(int *p) { #if 1 x10000(f(*p++);) #else x10000(asm("" :: "r"(*p++));) #endif } gcc -O -fschedule-insns invokes add_dependence 20000 times for each asm/call after the first. There is a loop for (i = 0; i < (unsigned)deps->max_reg; i++) { struct deps_reg *reg_last = &deps->reg_last[i]; reg_last->sets = alloc_INSN_LIST (insn, reg_last->sets); SET_REGNO_REG_SET (&deps->reg_last_in_use, i); } that registers the insn with reg_pending_barrier != 0 in reg_last->sets of each pseudo, and then all those reg_last->sets will be inspected on the next reg_pending_barrier insn.
Isn't the "scheduling window" limited somehow? Can we impose an upper bound on the number of dependences by placing a "virtual barrier" when we hit that limit? I don't know the structure of the scheduler or it's dependence analysis framework. In other places where GCC faces this usual quadraticness in dependence analysis we have such limits and try to cope with it as good as we can, worst give up completely.
Yea, there are various limits on the size of various lists the scheduler maintains. This looks independent of those various clamps.
If we want to get rid of the compilation time regression sooner rather than later, I can suggest limiting my change only to functions that call setjmp: diff --git a/gcc/sched-deps.cc b/gcc/sched-deps.cc index c23218890f..ae23f55274 100644 --- a/gcc/sched-deps.cc +++ b/gcc/sched-deps.cc @@ -3695,7 +3695,7 @@ deps_analyze_insn (class deps_desc *deps, rtx_insn *insn) CANT_MOVE (insn) = 1; - if (!reload_completed) + if (!reload_completed && cfun->calls_setjmp) { /* Scheduling across calls may increase register pressure by extending live ranges of pseudos over the call. Worse, in presence of setjmp That way we retain the "correctness fix" part of r13-5154-g733a1b777f1 and keep the previous status quo on normal functions (quadraticness on asms like demonstrated in comment #5 would also remain).
As far as I understand the testcase is from fuzzing so not "real", so I think this proposed "fix" isn't necessary (and it's not a real fix, adding a setjmp call at the end of the function will restore it).
Indeed, but OTOH according to bug 84402 comment 58 it caused a noticeable hit on gimple-match.cc compilation: 733a1b777f16cd397b43a242d9c31761f66d3da8 13th January 2023 sched-deps: do not schedule pseudos across calls [PR108117] (Alexander Monakov) Stage 2: +14% Stage 3: +9% In any case, if the proposed band-aid is unnecessary, that's fine with me.