capnode / algoloop

Windows desktop algo trading with QuantConnect Lean engine.

Algoloop trading application

Install prebuilt software and get started.

Introduction

Algoloop is an open-source algorithmic trading application with a Windows desktop frontend to QuantConnect Lean trading engine.

• Local algorithm backtest execution
• Algorithms in C# and Python
• Algorithm optimization
• Market data from multiple providers
• Windows desktop user interface

Information

More information and user documentation can be found here.

Contribution

Contributions are welcome:

• Live trading execution
• Full integration with QuantConnect web services
• ...

Contact

[email protected]

algoloop's Issues

It it possible to use Chinese stock data to backtest?

It it possible to use Chinese stock data to backtest?

does it support cryptocurrency trading with Binance?

Replacement for LeanReader

I see you're using LeanReader. The problem i see here (mybe i'm wrong), it will read only a single (zip) file. To read the market data across multiple zip files, you need another setup.

Here's what i've created:
https://gist.github.com/arakis/2120bd826066a440142a3a20649ff4f9

Slow performance when running backtest optimization

Expected Behavior

Utilize as much as possible of available cpu cores for useful execution

Actual Behavior

When running backtest optimization with many parallel threads, the cpu utilization is decreased. Studies with the Visual Studio tool Concurency Analyzer shows that the cpu cores are occupied with thread syncronization and garbage collection. This behaviour increases by the number of backtest threads.

Potential Solution

The current solution uses app domain to isolate backtest threads. All backtest execution is in the same process and uses the same garbage collector which blocks runing threads during garbage collection operation. Find out if separation by processes shows better perfomace.

Reproducing the Problem

Run backtest simulation with multiple threads. The cpu usage is not linear by the number of threads.
Use the tool Concurrency Visualizer to visualize this problem,

System Information

Windows 10

Checklist

• I have completely filled out this template
• I have confirmed that this issue exists on the current master branch
• I have confirmed that this is not a duplicate issue by searching issues

• I have provided detailed steps to reproduce the issue

Common Repository vs Submodule for QuantConnect.Lean

What are the reasons to extend QuantConnect.Lean instead of having them as git Submodule / keep them as separate Repository?

Maybe the reasons are to modify the source, but as i see, most of the features have an Interface, you they can extended / cusotmized.