Django generating test data incredibly slow

Question

Django 1.8, Postgres 8.4.20

I'm trying to prepare some performance test data in DB - generating 100 000 Django Users as well as other model instances. Here is my code:

for n in range(100000):
    # Generate User with one related Profile
    user = User.objects.create(username=str(random.random()))
    UserProfile.objects.create(custom_id=str(random.random()), user=user)

    # For each User generate a Journey with 10 related Visits
    journey = Journey.objects.create(user_id=user.id)
    for i in range(10):
        journey.visit_set.add(Visit(custom_id=str(random.random())))

    # Also for each User generate a Progress with 100 related Challenges
    progress = Progress.objects.create(user_id=user.id)
    for j in range(100):
        progress.challenge_set.add(Challenge(custom_id=str(random.random())))

It works as expected, the only issue is that it is incredibly slow. On a VPS with 1 core and 1 Gb RAM it took 1 hour to generate just 4000 users with corresponding related entries.

The consumed CPU is about 10%, memory 200mb and the load average floats around 1.00.

How can I speed this up?

Answer 1

As it was suggested in comments, I used bulk_create and launched my script 10 times in parallel. This increased the speed to about 100 User entries per second, which was fine in my case.