Efficient implmentation of Python multiprocesssing Pool - python

I have two codes. One is pooled (multiprocessing) version of the other. However, the parallel version with even 1 processor is taking a long time whereas the serial version finishes in ~15 sec. Can someone help to accelerate the second version.
Serial
import numpy as np, time
def mapTo(d):
global tree
for idx, item in enumerate(list(d), start=1):
tree[str(item)].append(idx)
data=np.random.randint(1,4, 20000000)
tree = dict({"1":[],"2":[],"3":[]})
s= time.perf_counter()
mapTo(data)
e = time.perf_counter()
print("elapsed time:",e-s)
takes: ~15 sec
Parallel
from multiprocessing import Manager, Pool
from functools import partial
import numpy as np
import time
def mapTo(i_d,tree):
idx,item = i_d
l = tree[str(item)]
l.append(idx)
tree[str(item)] = l
manager = Manager()
data = np.random.randint(1,4, 20000000)
# sharedtree= manager.dict({"1":manager.list(),"2":manager.list(),"3":manager.list()})
sharedtree = manager.dict({"1":[],"2":[],"3":[]})
s= time.perf_counter()
with Pool(processes=1) as pool:
pool.map(partial(mapTo, tree=sharedtree), list(enumerate(data,start=1)))
e = time.perf_counter()
print("elapsed time:",e-s)

Related

Thread Pool Executor using Concurrent: no improvement for various number of workers

I'm trying to implement a task in parallel using Concurrent. Please find below a piece of code for it:
import os
import time
from concurrent.futures import ProcessPoolExecutor as PE
import concurrent.futures
# num CPUs
cpu_num = len(os.sched_getaffinity(0))
print("Number of cpu available : ",cpu_num)
# max_Worker = cpu_num
max_Worker = 1
# A fake input array
n=1000000
array = list(range(n))
results = []
# A fake function being applied to each element of array
def task(i):
return i**2
x = time.time()
with concurrent.futures.ThreadPoolExecutor(max_workers=max_Worker) as executor:
features = {executor.submit(task, j) for j in array}
# the real function is heavy and we need to be sure of completeness of each run
for future in concurrent.futures.as_completed(features):
results.append(future.result())
results = [future.result() for future in features]
y = time.time()
print('=========================================')
print(f"Train data preparation time (s): {(y-x)}")
print('=========================================')
And now my questions,
Although there is no error, is it correct/optimized?
While playing with the number of workers, seems there is no
improvement in the speed (e.g., 1 vs 16, no difference). Then,
what's the problem and how can be solved?
Thanks in advance,
See my comment to your question. To the overhead I mentioned in that comment you need to also add the oberhead in just creating the process pool itself.
The following is a benchmark with several results. The first is a timing from just calling the worker function task 100000 times and creating a results list and printing out the last element of that list. It will become apparent why I have reduced the number of times I am calling task from 1000000 to 100000.
The next attempt is to use multiprocessing to accomplish the same thing using a ProcessPoolExecutor with the submit method and then processing the Future instances that are returned.
The next attempt is to instead use the map method with the default chunksize argument of 1 being used. It is important to understand this argument. With a chunksize value of 1, each element of the iterable that is passed to the map method is written individually to a queue of tasks as a chunk to be processed by the processes in the pool. When a pool process becomes idle looking for work, it pulls from the queue the next chunk of tasks to be performed, processes each task comprising the chunk and then becomes idle again. When there are a lot of submitted tasks being submitted via map, a chunksize value of 1 is inefficient. You would expect its performance to be equivalent to repeatedly issuing submit calls for each element of the iterable.
The next attempt specifies a chunksize value which approximates more or less the value that the map function used by the Pool class in the multiprocessing package would have used by default. As you can see, the improvement is dramatic, but still not an improvement over the non-multiprocessing case.
The final attempt uses the multiprocessing faciltity provided by package multiprocessing and its multiprocessing.pool.Pool class. The difference in this benchmark is that its map function uses a more intelligent default chunksize when no chunksize argument is specified.
import os
import time
from concurrent.futures import ProcessPoolExecutor as PE
from multiprocessing import Pool
# A fake function being applied to each element of array
def task(i):
return i**2
# required for Windows:
if __name__ == '__main__':
n=100000
t1 = time.time()
results = [task(i) for i in range(n)]
print('Non-multiprocessing time:', time.time() - t1, results[-1])
# num CPUs
cpu_num = os.cpu_count()
print("Number of CPUs available: ",cpu_num)
t1 = time.time()
with PE(max_workers=cpu_num) as executor:
futures = [executor.submit(task, i) for i in range(n)]
results = [future.result() for future in futures]
print('Multiprocessing time using submit:', time.time() - t1, results[-1])
t1 = time.time()
with PE(max_workers=cpu_num) as executor:
results = list(executor.map(task, range(n)))
print('Multiprocessing time using map:', time.time() - t1, results[-1])
t1 = time.time()
chunksize = n // (4 * cpu_num)
with PE(max_workers=cpu_num) as executor:
results = list(executor.map(task, range(n), chunksize=chunksize))
print(f'Multiprocessing time using map: {time.time() - t1}, chunksize: {chunksize}', results[-1])
t1 = time.time()
with Pool(cpu_num) as executor:
results = executor.map(task, range(n))
print('Multiprocessing time using Pool.map:', time.time() - t1, results[-1])
Prints:
Non-multiprocessing time: 0.027019739151000977 9999800001
Number of CPUs available: 8
Multiprocessing time using submit: 77.34723353385925 9999800001
Multiprocessing time using map: 79.52981925010681 9999800001
Multiprocessing time using map: 0.30500149726867676, chunksize: 3125 9999800001
Multiprocessing time using Pool.map: 0.2799997329711914 9999800001
Update
The following bechmarks use a version of task that is very CPU-intensive and shows the benefit of multiprocessing. It would also seem for this small iterable size (100), forcing a chunksize value of 1 for the Pool.map case (it would by default compute a chunksize value of 4), is slightly more performant.
import os
import time
from concurrent.futures import ProcessPoolExecutor as PE
from multiprocessing import Pool
# A fake function being applied to each element of array
def task(i):
for _ in range(1_000_000):
result = i ** 2
return result
def compute_chunksize(iterable_size, pool_size):
chunksize, remainder = divmod(iterable_size, pool_size * 4)
if remainder:
chunksize += 1
return chunksize
# required for Windows:
if __name__ == '__main__':
n = 100
cpu_num = os.cpu_count()
chunksize = compute_chunksize(n, cpu_num)
t1 = time.time()
results = [task(i) for i in range(n)]
t2 = time.time()
print('Non-multiprocessing time:', t2 - t1, results[-1])
# num CPUs
print("Number of CPUs available: ",cpu_num)
t1 = time.time()
with PE(max_workers=cpu_num) as executor:
futures = [executor.submit(task, i) for i in range(n)]
results = [future.result() for future in futures]
t2 = time.time()
print('Multiprocessing time using submit:', t2 - t1, results[-1])
t1 = time.time()
with PE(max_workers=cpu_num) as executor:
results = list(executor.map(task, range(n)))
t2 = time.time()
print('Multiprocessing time using map:', t2 - t1, results[-1])
t1 = time.time()
with PE(max_workers=cpu_num) as executor:
results = list(executor.map(task, range(n), chunksize=chunksize))
t2 = time.time()
print(f'Multiprocessing time using map: {t2 - t1}, chunksize: {chunksize}', results[-1])
t1 = time.time()
with Pool(cpu_num) as executor:
results = executor.map(task, range(n))
t2 = time.time()
print('Multiprocessing time using Pool.map:', t2 - t1, results[-1])
t1 = time.time()
with Pool(cpu_num) as executor:
results = executor.map(task, range(n), chunksize=1)
t2 = time.time()
print('Multiprocessing time using Pool.map (chunksize=1):', t2 - t1, results[-1])
Prints:
Non-multiprocessing time: 23.12758779525757 9801
Number of CPUs available: 8
Multiprocessing time using submit: 5.336004018783569 9801
Multiprocessing time using map: 5.364996671676636 9801
Multiprocessing time using map: 5.444890975952148, chunksize: 4 9801
Multiprocessing time using Pool.map: 5.400001287460327 9801
Multiprocessing time using Pool.map (chunksize=1): 4.698001146316528 9801

Multiprocessing in Python: Parallelize a for loop to fill a Numpy array

I've been reading threads like this one but any of them seems to work for my case. I'm trying to parallelize the following toy example to fill a Numpy array inside a for loop using Multiprocessing in Python:
import numpy as np
from multiprocessing import Pool
import time
def func1(x, y=1):
return x**2 + y
def func2(n, parallel=False):
my_array = np.zeros((n))
# Parallelized version:
if parallel:
pool = Pool(processes=6)
for idx, val in enumerate(range(1, n+1)):
result = pool.apply_async(func1, [val])
my_array[idx] = result.get()
pool.close()
# Not parallelized version:
else:
for i in range(1, n+1):
my_array[i-1] = func1(i)
return my_array
def main():
start = time.time()
my_array = func2(60000)
end = time.time()
print(my_array)
print("Normal time: {}\n".format(end-start))
start_parallel = time.time()
my_array_parallelized = func2(60000, parallel=True)
end_parallel = time.time()
print(my_array_parallelized)
print("Time based on multiprocessing: {}".format(end_parallel-start_parallel))
if __name__ == '__main__':
main()
The lines in the code based on Multiprocessing seem to work and give you the right results. However, it takes far longer than the non parallelized version. Here is the output of both versions:
[2.00000e+00 5.00000e+00 1.00000e+01 ... 3.59976e+09 3.59988e+09
3.60000e+09]
Normal time: 0.01605963706970215
[2.00000e+00 5.00000e+00 1.00000e+01 ... 3.59976e+09 3.59988e+09
3.60000e+09]
Time based on multiprocessing: 2.8775112628936768
My intuition tells me that it should be a better way of capturing results from pool.apply_async(). What am I doing wrong? What is the most efficient way to accomplish this? Thx.
Creating processes is expensive. On my machine it take at leas several hundred of microsecond per process created. Moreover, the multiprocessing module copy the data to be computed between process and then gather the results from the process pool. This inter-process communication is very slow too. The problem is that your computation is trivial and can be done very quickly, likely much faster than all the introduced overhead. The multiprocessing module is only useful when you are dealing with quite small datasets and perform intensive computation (compared to the amount of computed data).
Hopefully, when it comes to numericals computations using Numpy, there is a simple and fast way to parallelize your application: the Numba JIT. Numba can parallelize a code if you explicitly use parallel structures (parallel=True and prange). It uses threads and not heavy processes that are working in shared memory. Numba can overcome the GIL if your code does not deal with native types and Numpy arrays instead of pure Python dynamic object (lists, big integers, classes, etc.). Here is an example:
import numpy as np
import numba as nb
import time
#nb.njit
def func1(x, y=1):
return x**2 + y
#nb.njit('float64[:](int64)', parallel=True)
def func2(n):
my_array = np.zeros(n)
for i in nb.prange(1, n+1):
my_array[i-1] = func1(i)
return my_array
def main():
start = time.time()
my_array = func2(60000)
end = time.time()
print(my_array)
print("Numba time: {}\n".format(end-start))
if __name__ == '__main__':
main()
Because Numba compiles the code at runtime, it is able to fully optimize the loop to a no-op resulting in a time close to 0 second in this case.
Here is the solution proposed by #thisisalsomypassword that improves my initial proposal. That is, "collecting the AsyncResult objects in a list within the loop and then calling AsyncResult.get() after all processes have started on each result object":
import numpy as np
from multiprocessing import Pool
import time
def func1(x, y=1):
time.sleep(0.1)
return x**2 + y
def func2(n, parallel=False):
my_array = np.zeros((n))
# Parallelized version:
if parallel:
pool = Pool(processes=6)
####### HERE COMES THE CHANGE #######
results = [pool.apply_async(func1, [val]) for val in range(1, n+1)]
for idx, val in enumerate(results):
my_array[idx] = val.get()
#######
pool.close()
# Not parallelized version:
else:
for i in range(1, n+1):
my_array[i-1] = func1(i)
return my_array
def main():
start = time.time()
my_array = func2(600)
end = time.time()
print(my_array)
print("Normal time: {}\n".format(end-start))
start_parallel = time.time()
my_array_parallelized = func2(600, parallel=True)
end_parallel = time.time()
print(my_array_parallelized)
print("Time based on multiprocessing: {}".format(end_parallel-start_parallel))
if __name__ == '__main__':
main()
Now it works. Time is reduced considerably with Multiprocessing:
Normal time: 60.107836008071
Time based on multiprocessing: 10.049324989318848
time.sleep(0.1) was added in func1 to cancel out the effect of being a super trivial task.

starmap, starmap_async python

I want to translate a huge matlab model to python. Therefor I need to work on the key functions first. One key function handles parallel processing. Basically, a matrix with parameters is the input, in which every row represents the parameters for one run. These parameters are used within a computation-heavy function. This computation-heavy function should run in parallel, I don't need the results of a previous run for any other run. So all processes can run independent from eachother.
Why is starmap_async slower on my pc? Also: When i add more code (to test consecutive computation) my python crashes (i use spyder). Can you give me advice?
import time
import numpy as np
import multiprocessing as mp
from functools import partial
# Create simulated data matrix
data = np.random.random((100,3000))
data = np.column_stack((np.arange(1,len(data)+1,1),data))
def EAF_DGL(*z, package_num):
sum_row = 0
for i in range(1,np.shape(z)[0]):
sum_row = sum_row + z[i]
func_result = np.column_stack((package_num,z[0],sum_row))
return func_result
t0 = time.time()
if __name__ == "__main__":
package_num = 1
help_EAF_DGL = partial(EAF_DGL, package_num=1)
with mp.Pool() as pool:
#result = pool.starmap(partial(EAF_DGL, package_num), [(data[i]) for i in range(0,np.shape(data)[0])])
result = pool.starmap_async(help_EAF_DGL, [(data[i]) for i in range(0,np.shape(data)[0])]).get()
pool.close()
pool.join()
t1 = time.time()
calculation_time_parallel_async = t1-t0
print(calculation_time_parallel_async)
t2 = time.time()
if __name__ == "__main__":
package_num = 1
help_EAF_DGL = partial(EAF_DGL, package_num=1)
with mp.Pool() as pool:
#result = pool.starmap(partial(EAF_DGL, package_num), [(data[i]) for i in range(0,np.shape(data)[0])])
result = pool.starmap(help_EAF_DGL, [(data[i]) for i in range(0,np.shape(data)[0])])
pool.close()
pool.join()
t3 = time.time()
calculation_time_parallel = t3-t2
print(calculation_time_parallel)

Python multiprocessing calculations significantly slower than sequential calculations

i was playing with multiprocessing in python. I'm trying do distribute calculations on arrays to multiple CPU cores. In order to do that I'm forking as many processes as multiprocessing.cpu_count() returns and I'm passing subsets of the array to the processes (by partitioning the array indices). The array is operated on as a shared memory object.
However, for varying array sizes I cannot experience any runtime improvements. Why is that?
This is just a toy example, I'm not trying to achieve something with this calculations.
import multiprocessing as mp
import numpy as np
import time
import sharedmem
def some_function_mult(q, arr, index, width):
q.put((sum(arr[index:index+width])/np.amax(arr[index:index+width])**2)/40)
def some_function(arr, index, width):
return sum((arr[index:index+width])/np.amax(arr[index:index+width])**2)/40
def main():
num = mp.cpu_count()
n = 200000000
width = n/num
random_array = np.random.randint(0,255,n)
shared = sharedmem.empty(n)
shared[:] = random_array
print (shared)
queue = mp.Queue()
processes = [mp.Process(target=some_function_mult, args=(queue, shared, i*width, width)) for i in xrange(num)]
start_time = time.time()
for p in processes:
p.start()
result = []
for p in processes:
result.append(queue.get())
for p in processes:
p.join()
end_time = time.time()
print ('Multiprocessing execution time = ' + str(end_time-start_time))
print (result)
result = []
start_time =time.time()
for i in range(num):
result.append(some_function(random_array, i*width, width))
end_time = time.time()
print ('Sequential processing time = ' + str(end_time-start_time))
print (result)
if __name__ == '__main__':
main()

Is this the right way to use multiprocessing queue with python?

I am trying to speed up my code by using multiprocessing with Python. The only problem I ran into when trying to implement multiprocessing was that my function has a return statement and I needed to save that data to a list. The best way I found using google was to use queue as "q.put()" and retrieve it using "q.get()". The only issue is that I think i'm not utilizing this the right way because when I use command prompt after compiling, it shows i'm hardly using my cpu and I only see one Python process running. If I remove "q.get()" the process is super fast and utilizes my cpu. Am I doing this the right way?
import time
import numpy as np
import pandas as pd
import multiprocessing
from multiprocessing import Process, Queue
def test(x,y,q):
q.put(x * y)
if __name__ == '__main__':
q = Queue()
one = []
two = []
three = []
start_time = time.time()
for x in np.arange(30, 60, 1):
for y in np.arange(0.01, 2, 0.5):
p = multiprocessing.Process(target=test, args=(x, y, q))
p.start()
one.append(q.get())
two.append(int(x))
three.append(float(y))
print(x, ' | ', y, ' | ', one[-1])
p.join()
print("--- %s seconds ---" % (time.time() - start_time))
d = {'x' : one, 'y': two, 'q' : three}
data = pd.DataFrame(d)
print(data.tail())
No, this is not correct. You start a process and wait for the result through q.get at once. Therefore only one process running at the same time. If you want to operate on many tasks, use multiprocessing.Pool:
import time
import numpy as np
from multiprocessing import Pool
from itertools import product
def test((x,y)):
return x, y, x * y
def main():
start_time = time.time()
pool = Pool()
result = pool.map(test, product(np.arange(30, 60, 1), np.arange(0.01, 2, 0.5)))
pool.close()
print("--- %s seconds ---" % (time.time() - start_time))
print(result)
if __name__ == '__main__':
main()

Categories

Resources