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This page is outdated!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 UIDc4c80aecdb23b433
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:c4c80aecdb23b433
CompilerGCC 4.9
CPUSamsung GrandPrimePlus LTE LTN OPEN rev04 board based on MT6737T
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 45758edb8c77f868 1.72 1.03 -O3 -fcaller-saves -fconserve-stack -fno-compare-elim -fno-dse -fno-hoist-adjacent-loads -fipa-pure-const -fno-merge-constants -fsched-spec -fomit-frame-pointer -fno-partial-inlining -freciprocal-math -fsched-spec-load -fno-sched-spec-load-dangerous -fno-signaling-nans -fstrict-aliasing -ftree-bit-ccp -ftree-loop-linear -fno-unsafe-loop-optimizations -falign-functions=0 -falign-jumps=0 -ftree-parallelize-loops=0 --param min-inline-recursive-probability=14 --param min-vect-loop-bound=1 --param max-unroll-times=3 --param hot-bb-frequency-fraction=1093 --param align-threshold=194 --param tracer-min-branch-ratio=28 --param iv-always-prune-cand-set-bound=0 --param scev-max-expr-size=71 --param max-cselib-memory-locations=509 --param max-sched-region-insns=22 --param sched-spec-prob-cutoff=96 --param selsched-max-lookahead=9 --param prefetch-latency=65 --param sccvn-max-alias-queries-per-access=1413 --param ipa-cp-array-index-hint-bonus=81 --param allow-packed-load-data-races=1 -O3 1 0 2 1 milepost-codelet-mibench-automotive-basicmath-cubic-codelet-3-1 default 1040, 1040, 1040, 1040 1 SAMSUNG SM-G532M Android 6.0.1



[ 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