In [1]:

    

%matplotlib inline 

import pandas as pd 
import seaborn as sns


    

In [8]:

    

sns.set_style('whitegrid')
sns.set_context('poster')


    

In [16]:

    

data = pd.read_csv('results.csv')


    

In [17]:

    

data["latency (μs)"] = data["latency (ns)"] / 1000


    

In [9]:

    

g = sns.barplot("clients", "latency (μs)", data=data, palette="Blues_d")
g.set_title("RTT Message Latency to Single Server (total linearizeability)")
g.set_ylabel("mean latency (μs)")
g.set_xlabel("# concurrent clients")


    

    
Out[9]:





<matplotlib.text.Text at 0x1139622b0>






    

In [18]:

    

data = data.sample(frac=0.1)


    

In [28]:

    

data = data[data["latency (μs)"] < 5000.0]


    

In [29]:

    

data.describe()


    

    
Out[29]:






  

    

      

      
clients
      
latency (ns)
      
latency (μs)
    
  
  

    

      
count
      
335048.00000
      
3.350480e+05
      
335048.000000
    
    

      
mean
      
14.96151
      
1.335117e+06
      
1335.117352
    
    

      
std
      
6.46556
      
8.530676e+05
      
853.067565
    
    

      
min
      
1.00000
      
2.027580e+05
      
202.758000
    
    

      
25%
      
9.00000
      
6.907792e+05
      
690.779250
    
    

      
50%
      
15.00000
      
1.120420e+06
      
1120.420000
    
    

      
75%
      
21.00000
      
1.756270e+06
      
1756.270250
    
    

      
max
      
25.00000
      
4.999442e+06
      
4999.442000
    
  

In [33]:

    

g = sns.boxplot("clients", "latency (μs)", data=data, palette="Blues_d")


    

In [10]:

    

data = pd.read_csv('throughput.csv')


    

In [11]:

    

data.describe()


    

    
Out[11]:






  

    

      

      
clients
      
messages
      
latency (ns)
      
bytes
    
  
  

    

      
count
      
169.000000
      
169.000000
      
1.690000e+02
      
1.690000e+02
    
    

      
mean
      
17.307692
      
20244.869822
      
2.967853e+10
      
9.583974e+05
    
    

      
std
      
6.137318
      
7913.115783
      
3.324709e+08
      
3.789484e+05
    
    

      
min
      
1.000000
      
10666.000000
      
2.881337e+10
      
4.997270e+05
    
    

      
25%
      
13.000000
      
14087.000000
      
2.962512e+10
      
6.634730e+05
    
    

      
50%
      
19.000000
      
17111.000000
      
2.972315e+10
      
8.082660e+05
    
    

      
75%
      
23.000000
      
27398.000000
      
2.979560e+10
      
1.301017e+06
    
    

      
max
      
25.000000
      
40395.000000
      
3.224504e+10
      
1.923333e+06
    
  

In [12]:

    

data['throughput (msg/sec)'] = data['messages'] / (data['latency (ns)'] / 1000000000)


    

In [15]:

    

g = sns.barplot("clients", "throughput (msg/sec)", data=data, palette="Blues_d")
g.set_title("Write Throughput to Single Server (total linearizeability)")
g.set_ylabel("throughput (msg/sec)")
g.set_xlabel("# concurrent clients")


    

    
Out[15]:





<matplotlib.text.Text at 0x11041ef98>