Panda just got 7x F-F-Faster!
Probably the fastest full featured Angular App around ¯\_(ツ)_/¯
Today, we’re happy to announce that Panda is 7x faster. Speed is important to us because we want Panda to be a quick way for you to get information.
Huge Improvement on Load Time
By consuming far less RAM and CPU power.
So, where is the magic?
Panda is a Chrome Extension both running on new tab and background.
We booted Angular only once in background.
Hell yeah, we don’t load any of that third party libraries with every new tab. They are already loaded!
More details on how we improved the speed
There are a lot of things you have to be mindful of when making your web app faster. We have already taken advantage of minified resources, optimized images, browser caching and Angular template cache. However we still needed to do more to make Panda faster.
So what did we improve in this release? Some of them were obvious bottlenecks but for some, we had to dig deep and do a lot of experimentation. Below is a list of the improvements we made.
- Reducing the number of blocking HTTP requests when application is loading — making everything async.
- Using a content delivery network for both edge caching and reducing network latency on uncacheable endpoints.
- Putting multi-layers of cache — client’s local storage, built-in browser cache, CDN, in-memory solutions…
- Reducing the script size — removing environment based snippets at build time, replacing 3rd party libraries with lightweight alternatives. (Yes, you can survive without jQuery.)
- Cleaning a bulk of watchers to speed up ng-repeat rendering.
- Loading AngularJS app in the background and bootstrapping onto each new tab. This was quite a challange; we couldn’t find many examples. We had to read a lot of source code and discussions in the Angular core, fork many third parties to make them compatible with our app, build housekeeping tasks to make everything more stable. But this has by far the most noticable impact!
This is a little summary, but if you want to know more subscribe and wait for the upcoming technical article.
