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This page is outdated!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 UID
8289e0cf24346aa7 (experiment.tune.compiler.flags.gcc.e)
Data UID
89eafdc0af94642c
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 8289e0cf24346aa7:89eafdc0af94642c
Compiler
GCC 4.9
CPU
MT6737T
Objective
min
Improvement key IK1
Main kernel execution time speedup [min]
Improvement key IK2
Code 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
e14e0a893b1e2588
1.24
1.00
-O3 -fbranch-target-load-optimize2 -fbtr-bb-exclusive -fno-ipa-sra -fira-share-save-slots -ftrapping-math -fno-rounding-math -fschedule-insns -fno-section-anchors -fno-sel-sched-pipelining-outer-loops -fstrict-aliasing -fstrict-overflow -fno-tree-builtin-call-dce -fno-tree-dse -fno-tree-forwprop -ftree-pre -fno-tree-sink -fno-tree-sra -ftree-switch-conversion -funsafe-math-optimizations --param inline-unit-growth=0 --param max-peel-times=21 --param max-unswitch-insns=88 --param hot-bb-frequency-fraction=1632 --param tracer-max-code-growth=157 --param max-cse-insns=182 --param iv-max-considered-uses=250 --param scev-max-expr-complexity=0 --param integer-share-limit=20 --param l1-cache-line-size=24 --param graphite-max-nb-scop-params=17 --param loop-max-datarefs-for-datadeps=1866 --param ipa-cp-loop-hint-bonus=124 --param lto-partitions=55 --param allow-packed-load-data-races=1 --param max-tail-merge-comparisons=2 --param max-tail-merge-iterations=3 --param asan-stack=0 --param asan-instrument-reads=0
-O3
1
0
2
1
cbench-consumer-jpeg-d
decode
image-jpeg-dnn-cat
cat.jpg
1300, 1300, 1300, 1300
1
SPA CONDOR ELECTRONICS PGN610
Android 7.0
[ 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
Go Back
Developed by
Grigori Fursin
Implemented as a
CK workflow
Hosted at