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This page is outdated!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 UID20193453da8bad47
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 8289e0cf24346aa7:20193453da8bad47
CompilerGCC 4.9
CPUhi6620oem
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 76d21ea259446749 1.10 1.12 -O3 -fno-compare-elim -fcse-skip-blocks -fdata-sections -fno-devirtualize-speculatively -fno-expensive-optimizations -ffast-math -ffloat-store -fipa-pta -fira-hoist-pressure -fno-isolate-erroneous-paths-dereference -fno-ivopts -floop-interchange -floop-parallelize-all -fmerge-all-constants -fno-branch-count-reg -ftoplevel-reorder -fno-partial-inlining -fpredictive-commoning -fno-sched-spec-load-dangerous -fno-tree-builtin-call-dce -ftree-coalesce-vars -ftree-pta -ftree-vectorize -funsafe-math-optimizations -finline-limit=0 --param predictable-branch-outcome=10 --param large-stack-frame=442 --param hot-bb-frequency-fraction=1543 --param lim-expensive=29 --param iv-consider-all-candidates-bound=36 --param omega-hash-table-size=205 --param max-cselib-memory-locations=756 --param sink-frequency-threshold=21 --param max-sched-region-insns=182 --param sched-mem-true-dep-cost=0 --param max-last-value-rtl=1305 --param max-sched-ready-insns=139 --param prefetch-latency=36 --param ira-loop-reserved-regs=4 --param loop-invariant-max-bbs-in-loop=3046 --param min-insn-to-prefetch-ratio=9 --param ipa-cp-array-index-hint-bonus=58 --param asan-stack=1 -O3 1 0 2 1 cbench-network-dijkstra default cdataset-dijkstra-0001 data.txt 1 HUAWEI P7-L10 Android 4.4.2



[ 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