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This page is outdated!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 UID35876e5541b08d51
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 8289e0cf24346aa7:35876e5541b08d51
CompilerGCC 6.1.0
CPUIntel(R) Xeon(R) CPU E5-2630 v3 @ 2.40GHz
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 14bf3d94b2392703 3.49 0.79 -O3 -fno-guess-branch-probability -fno-tree-loop-if-convert -funroll-all-loops -O3 2 0 61 94 milepost-codelet-mibench-consumer-jpeg-c-src-jchuff-codelet-9-1 default 2599.968, 2599.968, 2599.968, 1535.539, 2599.968, 7933.69, 7933.784, 2599.968, 2599.968, 2599.968, 2599.968, 2599.968, 2600.062, 2599.968, 4286617.642, 7123.971 16 Dell Inc. (PowerEdge R630) Debian 8.2
S2 b23d17d1154faa86 1.98 0.99 -O3 -fno-caller-saves -ftracer -fno-tree-vectorize --param max-completely-peel-times=26 -O3 2 2 44 83 milepost-codelet-mibench-consumer-jpeg-c-src-jchuff-codelet-9-1 default 2599.968, 7933.69, 4291654.843, 7933.69, 7934.065, 2298.659, 2599.968, 2599.968, 2599.968, 2599.968, 2599.968, 2599.968, 2599.968, 7933.69, 2599.968, 2599.968 16 Dell Inc. (PowerEdge R630) Debian 8.2
S3 441bb413d241f5ba 1.93 0.34 -O3 -flto -fno-rerun-cse-after-loop -O3 4 1 58 93 milepost-codelet-mibench-consumer-lame-src-takehiro-codelet-16-1 default 7933.877, 3827.346, 7933.69, 4290125.111, 7933.877, 4289628.611, 4291370.392, 2599.968, 2599.968, 2599.968, 2599.968, 2599.968, 2599.968, 2599.968, 7933.784, 4292707.267 16 Dell Inc. (PowerEdge R630) Debian 8.2
S4 34283a3731468f5c 1.92 0.34 -O3 -fno-caller-saves -flto -O3 1 2 16 56 milepost-codelet-mibench-consumer-lame-src-takehiro-codelet-16-1 default 3027.846, 7934.252, 4288447.551, 2599.968, 2599.968, 2599.968, 7934.252, 2599.968, 2599.968, 2599.968, 2599.968, 2599.968, 2599.968, 2599.968, 7150.781, 76.69 16 Dell Inc. (PowerEdge R630) Debian 8.2
S5 a614a1caf6600e21 1.90 0.34 -O3 -flto -fno-reorder-blocks -O3 2 2 47 85 milepost-codelet-mibench-consumer-lame-src-takehiro-codelet-16-1 default 7521.096, 4287921.611, 7933.69, 2599.968, 7933.784, 6672.843, 7933.784, 2599.968, 2599.968, 2599.968, 2599.968, 2599.968, 2599.968, 7933.596, 4291968.707, 2599.968 16 Dell Inc. (PowerEdge R630) Debian 8.2
S6 79a38dd2474235b8 1.85 0.34 -O3 -fno-gcse -flto -O3 2 2 51 88 milepost-codelet-mibench-consumer-lame-src-takehiro-codelet-16-1 default 2599.968, 7933.69, 7688.25, 4287771.798, 7933.877, 2533.784, 2599.968, 2599.968, 2600.062, 2599.968, 2599.968, 2599.968, 2599.968, 2599.968, 5624.343, 6139.781 16 Dell Inc. (PowerEdge R630) Debian 8.2
S7 f3d101524537a75c 1.80 1.00 -O3 -fno-tree-loop-optimize -funsafe-math-optimizations -O3 1 2 63 95 milepost-codelet-mibench-consumer-jpeg-c-src-jchuff-codelet-9-1 default 2599.968, 2599.968, 7933.221, 2599.968, 7933.221, 2599.968, 7933.034, 2600.062, 2599.968, 2599.968, 2599.968, 2599.968, 2599.968, 4287771.705, 7932.752, 6096.852 16 Dell Inc. (PowerEdge R630) Debian 8.2
S8 445abdec73af67b8 1.61 1.04 -O3 -ffloat-store -fno-math-errno -O3 1 4 56 92 milepost-codelet-mibench-consumer-lame-src-quantize-codelet-7-1 default 2599.968, 1483.784, 2599.968, 7933.877, 2599.968, 7933.971, 2599.968, 2599.968, 2599.968, 2599.968, 2599.968, 2599.968, 2599.968, 7895.63, 5503.593, 4288861.923 16 Dell Inc. (PowerEdge R630) Debian 8.2
S9 58005df63b3af967 1.59 0.80 -O3 -fno-predictive-commoning -funroll-loops -O3 1 1 65 96 milepost-codelet-mibench-consumer-lame-src-quantize-codelet-7-1 default 2599.968, 2599.968, 2599.968, 3520.147, 2599.968, 6218.269, 7933.596, 2600.062, 2599.968, 2599.968, 4293412.005, 2599.968, 4293441.517, 2599.968, 4292467.736, 2599.968 16 Dell Inc. (PowerEdge R630) Debian 8.2
S10 69e74842ae43beec 1.59 0.80 -O3 -funroll-loops -O3 2 1 67 96 milepost-codelet-mibench-consumer-lame-src-quantize-codelet-7-1 default 7045.036, 2599.968, 7932.752, 2599.968, 7932.659, 7932.284, 7932.659, 2599.968, 2599.968, 2599.968, 2599.968, 2599.968, 2599.968, 8360.627, 7932.659, 2599.968 16 Dell Inc. (PowerEdge R630) Debian 8.2
S11 ac1ec024e69b7cb9 1.59 0.82 -O3 -fno-tree-ch -funroll-all-loops -O3 6 1 62 94 milepost-codelet-mibench-consumer-lame-src-quantize-codelet-7-1 default 4286904.244, 7933.971, 4290851.861, 7933.877, 6231.675, 7933.971, 2599.968, 2599.968, 2599.968, 2599.968, 2599.968, 2599.968, 2599.968, 2599.968, 4291650.33, 7933.784 16 Dell Inc. (PowerEdge R630) Debian 8.2
S12 c59f284824abdc89 1.42 1.08 -O3 -fno-merge-constants -fsingle-precision-constant -O3 2 0 70 98 milepost-codelet-mibench-office-rsynth-src-nsynth-codelet-5-1 default 2599.968, 7934.534, 7934.534, 7934.44, 4292238.33, 7934.534, 2599.968, 2599.968, 7297.781, 2599.968, 2599.968, 2599.968, 2599.968, 4291701.33, 5460.75, 4286721.829 16 Dell Inc. (PowerEdge R630) Debian 8.2
S13 4f38908a30da60a2 1.37 1.05 -O3 -ffast-math -fno-tree-ch -O3 3 4 34 73 cbench-consumer-jpeg-d decode image-jpeg-0012 data.jpg 2599.968, 2599.968, 7932.096, 2599.968, 7932.096, 2599.968, 7932.096, 2599.968, 2599.968, 2599.968, 2599.968, 2599.968, 2599.968, 2599.968, 4287943.267, 4288833.236 16 Dell Inc. (PowerEdge R630) Debian 8.2
S14 dd7940288e32f819 1.34 0.85 -O3 -fno-crossjumping -ftracer -O3 1 0 5 46 milepost-codelet-mibench-telecomm-adpcm-d-src-adpcm-codelet-1-1 default 7933.221, 7933.221, 7933.315, 7933.315, 7933.221, 4287771.705, 2599.968, 2599.968, 2599.968, 2599.968, 2599.968, 2599.968, 2599.968, 8073.19, 7933.221, 4289147.864 16 Dell Inc. (PowerEdge R630) Debian 8.2



[ 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