Quantopian's community platform is shutting down. Please read this post for more information and download your code.
Back to Community
Speed comparison - Offline Zipline environment 35 times faster than Online Quantopian (is this true???)
Don Q
edited

Dear All,

When testing speed of my local zipline environment vs Quantopian environment, I found local zipline environment 35 times faster than Q server.

Lesson Learnt: handle_data is running every minute, to increase speed for backesting daily data, new function and scheduler has to be created as Dan mentioned below otherwise backtest is very slow.

Example of creating scheduled function: 

def initialize(context):

    attach_pipeline(make_pipeline(context), 'my_pipeline')  
    # Schedule the renamed `my_handle_data` to run once a day  
    schedule_function(my_handle_data, date_rules.every_day(), time_rules.market_close())

# Rename `handle data` so it doesn't run every minute  
def my_handle_data(context, data):
  • Offline Zipline test below takes 2min 20s to backtest 10 years (2006-1-1 : 2016-1-1)
from datetime import datetime  
st = datetime.now()  
from zipline import run_algorithm  
from zipline.api import *  
from zipline.pipeline import CustomFactor, Pipeline  
from zipline.pipeline.data import USEquityPricing  
from zipline.pipeline.factors import Returns, SimpleMovingAverage  
import zipline.pipeline.filters as Filters  
import pytz  
import pandas as pd  
import numpy as np


def initialize(context):  
    attach_pipeline(make_pipeline(context), 'my_pipeline')

def make_pipeline(context):  
    context.symbols = symbols("AAPL","AXP","BA","CAT","CSCO","CVX","DIS","GS","HD","IBM","INTC","JNJ",  
    "JPM","KO","MCD","MMM","MRK","MSFT","NKE","PFE","PG","TRV","UNH","UTX","VZ","WMT","XOM",  )  
    universe = Filters.StaticAssets(context.symbols)  
    mean_close_100 = SimpleMovingAverage(inputs=[USEquityPricing.close],window_length=100)  
    yesterday_close = USEquityPricing.close.latest

    return Pipeline(  
            columns = {  
            'mean_close_100' : mean_close_100,  
            'yesterday_close' : yesterday_close,  
                      },  
            screen = universe,  
            )  
def handle_data(context, data):

    context.output = pipeline_output('my_pipeline')  
    context.output['weight'] = np.where(context.output['yesterday_close']  > context.output['mean_close_100'], 1/len(context.symbols), 0.0)  
    for stock, analytics in context.output.iterrows():  
        if data.can_trade(stock):  
            order_target_percent(stock, analytics['weight'])

start = datetime(2006, 1, 1, tzinfo=pytz.UTC)  
end = datetime(2016, 1, 1, tzinfo=pytz.UTC)

results = run_algorithm(start=start, end=end,  
                        initialize=initialize,  
                        handle_data=handle_data,  
                        capital_base=10000,  
                        data_frequency = 'daily', bundle='quandl' )  
ed = datetime.now()  
print(ed-st)
  • Online Quantopian server test below takes 2min 20s to backtest 3.5 months (2006-1-1 : 2006-3-15)
from quantopian.pipeline.data.builtin import USEquityPricing  
from quantopian.algorithm import attach_pipeline, pipeline_output  
from quantopian.pipeline import Pipeline  
import quantopian.pipeline.filters as Filters  
from quantopian.pipeline.factors import Returns, SimpleMovingAverage  
import pandas as pd  
import numpy as np

def initialize(context):

    attach_pipeline(make_pipeline(context), 'my_pipeline')


def make_pipeline(context):  
    context.symbols = symbols("AAPL","AXP","BA","CAT","CSCO","CVX","DIS","GS","HD","IBM","INTC","JNJ",  
    "JPM","KO","MCD","MMM","MRK","MSFT","NKE","PFE","PG","TRV","UNH","UTX","VZ","WMT","XOM",  )  

    universe = Filters.StaticAssets(context.symbols)  
    mean_close_100 = SimpleMovingAverage(inputs=[USEquityPricing.close],window_length=100)  
    yesterday_close = USEquityPricing.close.latest

    return Pipeline(  
            columns = {  
            'mean_close_100' : mean_close_100,  
            'yesterday_close' : yesterday_close,  
                      },  
            screen = universe,  
            )  
def handle_data(context, data):

    context.output = pipeline_output('my_pipeline')  
    context.output['weight'] = np.where(context.output['yesterday_close']  >                 context.output['mean_close_100'], 1/len(context.symbols), 0.0)  
    for stock, analytics in context.output.iterrows():  
        if data.can_trade(stock):  
            order_target_percent(stock, analytics['weight'])
1 response
Dan Whitnable

The online Quantopian environment always runs in data_frequency='minute' mode. It appears this comparison is running the offline Zipline code with data_frequency='daily'. That would account for a big performance impact. Maybe try changing the zipline run_algorithm method to 'minute'. Like this.

results = run_algorithm(start=start, end=end,  
                        initialize=initialize,  
                        handle_data=handle_data,  
                        capital_base=10000,  
                        data_frequency = 'minute', bundle='quandl' )

Or, as an alternative, in the online Quantopian environment, schedule the handle_data code to run only once a day. This isn't exactly the same because, in the background, all the order logic will still execute every minute. But it should give a better comparison. Simply rename the handle_data function and then schedule that function to run once a day. Something like this

def initialize(context):

    attach_pipeline(make_pipeline(context), 'my_pipeline')  
    # Schedule the renamed `my_handle_data` to run once a day  
    schedule_function(my_handle_data, date_rules.every_day(), time_rules.market_close())

# Rename `handle data` so it doesn't run every minute  
def my_handle_data(context, data):

Maybe try with those changes and see how each performs.

Disclaimer

The material on this website is provided for informational purposes only and does not constitute an offer to sell, a solicitation to buy, or a recommendation or endorsement for any security or strategy, nor does it constitute an offer to provide investment advisory services by Quantopian. In addition, the material offers no opinion with respect to the suitability of any security or specific investment. No information contained herein should be regarded as a suggestion to engage in or refrain from any investment-related course of action as none of Quantopian nor any of its affiliates is undertaking to provide investment advice, act as an adviser to any plan or entity subject to the Employee Retirement Income Security Act of 1974, as amended, individual retirement account or individual retirement annuity, or give advice in a fiduciary capacity with respect to the materials presented herein. If you are an individual retirement or other investor, contact your financial advisor or other fiduciary unrelated to Quantopian about whether any given investment idea, strategy, product or service described herein may be appropriate for your circumstances. All investments involve risk, including loss of principal. Quantopian makes no guarantees as to the accuracy or completeness of the views expressed in the website. The views are subject to change, and may have become unreliable for various reasons, including changes in market conditions or economic circumstances.

About Quantopian
Careers
Community
Events
Documentation
Academia
Lectures
Status
Facebook
Twitter
LinkedIn
YouTube
Terms of Use
Privacy Policy

The material on this website is provided for informational purposes only and does not constitute an offer to sell, a solicitation to buy, or a recommendation or endorsement for any security or strategy, nor does it constitute an offer to provide investment advisory services by Quantopian.

In addition, the material offers no opinion with respect to the suitability of any security or specific investment. No information contained herein should be regarded as a suggestion to engage in or refrain from any investment-related course of action as none of Quantopian nor any of its affiliates is undertaking to provide investment advice, act as an adviser to any plan or entity subject to the Employee Retirement Income Security Act of 1974, as amended, individual retirement account or individual retirement annuity, or give advice in a fiduciary capacity with respect to the materials presented herein. If you are an individual retirement or other investor, contact your financial advisor or other fiduciary unrelated to Quantopian about whether any given investment idea, strategy, product or service described herein may be appropriate for your circumstances. All investments involve risk, including loss of principal. Quantopian makes no guarantees as to the accuracy or completeness of the views expressed in the website. The views are subject to change, and may have become unreliable for various reasons, including changes in market conditions or economic circumstances.