Nodejs optimizations with c++ addons

Let's make service using node js to create ‘QR codes’(CPU intensive task) to understand how can we use C++ for optimizations.

Results for generating 10000 Qr Codes:

C++ single thread:2960.125ms

Nodejs: 19180.051ms

Node js code to generate QR codes:

Following steps for C++:

Step 1: package.json

We need to add “gypfile” flag as true and dependencies.

Step 2: Create DataProcessingAsyncWorker.h (addon folder)

Header file that we will need further.

Step 3: Create DataProcessingAsyncWorker.cc (addon folder)

Here we write processing logic.

Step 4: Create Addon.cc (addon folder)

Here we have c++ function (ProcessData) which we can call from javascript and which will execute our main processing logic on a separate thread by adding it to worker queue.

Step 5: Install C++ dependencies

Step 6: Create binding.gyp

Step 7: Npm install

Npm install will build C++ files and it can be used in node js as follows:

And we are done with code walkthrough!

Every request will be added to a new thread. Let's try with 2 concurrent requests for 50,000 and 40,000 codes.

Here we can see parallel processing of 2 requests on separate threads(Multithreading).

Conclusion

We saw how easy it is to optimize node js for CPU intensive tasks using c++.

Source code: https://github.com/riddheshganatra/qrcodenodecpp