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This page is outdated!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 UID48fb1d48b8a03e63
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 8289e0cf24346aa7:48fb1d48b8a03e63
CompilerGCC 4.9
CPUQualcomm Technologies, Inc MSM8216
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 a880f320e20178b5 1.10 0.25 -O3 -fno-associative-math -fcheck-data-deps -fno-cx-limited-range -fno-dse -fipa-sra -fno-gcse-sm -fno-ira-share-save-slots -flto -fno-peephole -fpeephole2 -fno-reorder-functions -fshrink-wrap -fsplit-ivs-in-unroller -fno-web --param max-variable-expansions-in-unroller=0 --param inline-unit-growth=46 --param max-gcse-insertion-ratio=7 --param gcse-unrestricted-cost=5 --param max-hoist-depth=58 --param max-peel-branches=18 --param max-unswitch-insns=93 --param hot-bb-count-ws-permille=677 --param unlikely-bb-count-fraction=5162 --param tracer-max-code-growth=17 --param omega-max-wild-cards=23 --param max-sched-region-insns=31 --param integer-share-limit=104 --param l1-cache-size=41 --param switch-conversion-max-branch-ratio=1 --param ipa-cp-eval-threshold=974 --param lto-min-partition=1282 --param allow-packed-store-data-races=0 --param tree-reassoc-width=0 --param asan-globals=1 --param asan-instrument-writes=0 --param asan-memintrin=1 -O3 1 0 2 1 milepost-codelet-mibench-telecomm-adpcm-c-src-adpcm-codelet-1-1 default au-0001 1190.4, 1190.4, 1190.4, 1190.4 1 SAMSUNG SM-G530BT Android 4.4.4

[ 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