

In [1]:

    
%pylab inline
import sys
sys.path.append('..')
import trappy.wa









    



Populating the interactive namespace from numpy and matplotlib

Displaying Results

At the end of running an agenda, Workload Automation produces a summary of the results for the workloads. trappy can parse the results and tabulate it in a notebook. Additionally, an optional id argument an be passed which if supplied overrides the id in the results file.



In [2]:

    
results = trappy.wa.get_results("../tests")
results









    Out[2]:






  
    
      
      geekbench
      antutu
      egypt_offscreen
      thechase
      t-rex_offscreen
    
    
      
      power_allocator
      step_wise
      power_allocator
      step_wise
      power_allocator
      step_wise
      power_allocator
      step_wise
      power_allocator
      step_wise
    
  
  
    
      0
      3
      8
      5
      4
      652
      504
      491.615669
      242.052226
      1777
      2365
    
    
      1
      1
      4
      3
      9
      555
      2507
      NaN
      NaN
      397
      429
    
    
      2
      5
      2
      2
      7
      790
      325
      NaN
      NaN
      512
      424

In addition to displaying tabulated results, trappy can graphically plot individual benchmarks as well as all the benchmarks in the result.



In [3]:

    
results.plot_results_benchmark("antutu")



In [4]:

    
results.plot_results()

Combining Results

trappy allows combination of results across different agenda runs as well. This is useful if you want to compare / display results from different runs of Workload Automation



In [5]:

    
unconstrained = trappy.wa.get_results("../tests/unconstrained.csv", name="Unconstrained")
constrained = trappy.wa.get_results("../tests/constrained.csv", name="Constrained")

results = trappy.wa.combine_results([unconstrained, constrained])
results









    Out[5]:






  
    
      
      antutu
      egypt_offscreen
      t-rex_offscreen
      geekbench
    
    
      
      Unconstrained
      Constrained
      Unconstrained
      Constrained
      Unconstrained
      Constrained
      Unconstrained
      Constrained
    
  
  
    
      0
      2
      2
      864
      334
      185
      560
      6
      1
    
    
      1
      NaN
      NaN
      802
      242
      878
      872
      6
      1
    
    
      2
      6
      7
      47
      190
      262
      918
      1
      3
    
    
      3
      4
      4
      NaN
      NaN
      588
      5
      7
      2
    
    
      4
      6
      2
      NaN
      NaN
      559
      494
      9
      8
    
    
      5
      9
      8
      NaN
      NaN
      NaN
      NaN
      NaN
      NaN



In [6]:

    
results.plot_results()

	geekbench		antutu		egypt_offscreen		thechase		t-rex_offscreen
	power_allocator	step_wise	power_allocator	step_wise	power_allocator	step_wise	power_allocator	step_wise	power_allocator	step_wise
0	3	8	5	4	652	504	491.615669	242.052226	1777	2365
1	1	4	3	9	555	2507	NaN	NaN	397	429
2	5	2	2	7	790	325	NaN	NaN	512	424

	antutu		egypt_offscreen		t-rex_offscreen		geekbench
	Unconstrained	Constrained	Unconstrained	Constrained	Unconstrained	Constrained	Unconstrained	Constrained
0	2	2	864	334	185	560	6	1
1	NaN	NaN	802	242	878	872	6	1
2	6	7	47	190	262	918	1	3
3	4	4	NaN	NaN	588	5	7	2
4	6	2	NaN	NaN	559	494	9	8
5	9	8	NaN	NaN	NaN	NaN	NaN	NaN