In [ 45 ]: plt.figure(figsize=( 8 , 4 ))
plt.plot(strikes, option_values_seq, ‘b’, label=‘Sequential’)
plt.plot(strikes, option_values_par, ‘r.’, label=‘Parallel’)
plt.grid(True); plt.legend(loc= 0 )
plt.xlabel(‘strikes’)
plt.ylabel(‘European call option values’)
Figure 8-3. Comparison of European call option values
Performance Comparison
With the help of the perf_comp_func function, we can compare the performance a bit
more rigorously:
In [ 46 ]: n = 50 # number of option valuations
func_list = [‘seq_value’, ‘par_value’]
data_list = 2 * [‘n’]
In [ 47 ]: perf_comp_data(func_list, data_list)
Out[47]: function: par_value, av. time sec: 0.90832, relative: 1.0
function: seq_value, av. time sec: 5.75137, relative: 6.3
The results clearly demonstrate that using IPython.parallel for parallel execution of
functions can lead to an almost linear scaling of the performance with the number of cores
available.
