Collective Knowledge Aggregator proof-of-concept
Crowd results Raw CK browser Graphs Reports Datasets Models Home

This page is outdated! New version is available here.


Distinct solutions after online classification (auto/crowd-tune GCC compiler flags (minimize execution time))

Scenario UID8289e0cf24346aa7 (experiment.tune.compiler.flags.gcc.e)
Data UID384a4516b1e11bb2
Discuss (optimizations to improve compilers,
semantic/data set/hardware features
to improve predictions
, etc):
GitHub wiki, Google group
Download:[ All solutions in JSON ], [ Solutions' classification in JSON ]
Reproduce all (with reactions):ck replay 8289e0cf24346aa7:384a4516b1e11bb2
CompilerGCC 4.9
CPUQualcomm MSM8226
Objectivemin
Improvement key IK1Main kernel execution time speedup [min]
Improvement key IK2Code size improvement

Improvements (<4% variation) Distinct workload for highest improvement
# Solution UID IK1 IK2 New distinct optimization choices Ref Best species Worst species Touched Iters Program CMD Dataset Dataset file CPU freq (MHz) Cores Platform OS Replay
S1 ff9c34961adba7de 1.78 0.80 -O3 -fdelayed-branch -fearly-inlining -fno-gcse -fgcse-lm -fkeep-static-consts -flive-range-shrinkage -fmodulo-sched-allow-regmoves -fno-move-loop-invariants -fno-partial-inlining -fsched-pressure -fno-sched-spec-insn-heuristic -fno-strict-aliasing -ftree-fre -ftree-loop-if-convert -fno-tree-partial-pre -fno-unroll-all-loops -funroll-loops -fno-vpt -fno-whole-program -falign-jumps=0 -ffp-contract=fast -fira-algorithm=priority --param max-pending-list-length=45 --param large-function-insns=2691 --param ipcp-unit-growth=17 --param gcse-after-reload-critical-fraction=3 --param sms-loop-average-count-threshold=0 --param align-threshold=14 --param iv-consider-all-candidates-bound=48 --param omega-max-eqs=33 --param omega-max-wild-cards=20 --param vect-max-version-for-alias-checks=16 --param min-spec-prob=30 --param selsched-max-lookahead=41 --param sched-mem-true-dep-cost=0 --param ssp-buffer-size=4 --param lra-max-considered-reload-pseudos=364 --param switch-conversion-max-branch-ratio=3 --param loop-invariant-max-bbs-in-loop=5409 --param slp-max-insns-in-bb=1087 --param max-vartrack-size=23039062 --param ipa-cp-loop-hint-bonus=101 --param max-slsr-cand-scan=348863 --param asan-instrument-writes=0 --param uninit-control-dep-attempts=1234 -O3 1 0 2 1 milepost-codelet-mibench-security-pgp-e-src-mpilib-codelet-3-1 default image-ppm-0001 data.ppm 1 SONY D2305 Android 5.1.1
S2 c0329dfb194878b6 1.10 0.82 -O3 -fno-cse-follow-jumps -fcx-limited-range -fdelete-null-pointer-checks -fno-gcse-sm -fno-optimize-sibling-calls -fno-sched-rank-heuristic -fsel-sched-pipelining -fno-single-precision-constant -fno-thread-jumps -ftree-dse -ftree-loop-if-convert-stores -fno-tree-pre -fno-tree-switch-conversion -ftree-vrp -funit-at-a-time -funroll-all-loops -fno-whole-program --param large-function-insns=2259 --param early-inlining-insns=9 --param gcse-after-reload-critical-fraction=12 --param max-peeled-insns=142 --param sms-min-sc=1 --param max-predicted-iterations=189 --param max-grow-copy-bb-insns=3 --param iv-max-considered-uses=177 --param scev-max-expr-size=68 --param vect-max-peeling-for-alignment=25 --param sched-mem-true-dep-cost=0 --param integer-share-limit=474 --param sccvn-max-scc-size=7045 --param switch-conversion-max-branch-ratio=12 --param ipa-cp-eval-threshold=316 --param max-stores-to-sink=2 --param case-values-threshold=0 --param allow-packed-store-data-races=0 --param asan-stack=0 -O3 1 0 2 1 cbench-security-blowfish encode enc-0001 data.enc 1 SONY D2305 Android 5.1.1



[ Participated users, platforms, OS, CPU, GPU, GPGPU, NN, NPU ] [ How to participate ] [ Motivation (PPT) (PDF) ] [ Papers 1 , 2 , 3] [ Android app ] [ Collective training set ] [ Unified AI ]
View entry in raw format

Developed by Grigori Fursin           
Implemented as a CK workflow
                         
   
                      Hosted at