A simple web app for comparing cpus. No bloat, just quick and easy comparison.
This project was build with backwards compatibility in mind, so that even 12-year-old cpus can be compared.
The data is scraped from each manufacturer's website. The data is then cached in a redis database, to ensure blazing fast response times. If the data is older than required by the app, it will be updated.
The selected cpus are saved in the url, so that you can share your comparison with others.
Simply select the manufacturer and the model of the cpu you want to compare. The data will automatically fetch, after the bar vanishes or after pressing enter.
After both cpus are loaded (indicated by the green color), the comparison table will be displayed.
The app always tries to fetch the data from the redis cache first. If the data is not in the cache, it will be fetched from the manufacturer's website, using the approaches described below.
Getting the data for intel cpus was not that hard. I simply used their internal search API, that is used for search on
their own website.
The only con of this approach is, that by simply searching a phrase like Core, it can return a random cpu.
Getting the data for AMD cpus was a bit more tricky. AMD does not have an internal search API, but I discovered
a page which has all of their cpus embedded in the html
file itself. I wrote a simple script that extracted the names and their ids that can be used to get directly to their
specs page.
This is better than the earlier solution because we can save ~4 seconds of response time.
The biggest problem is that AMD doesn't like being fetched by a server. In the end I ended up spinning my own instance of Browserless and using it to fetch the data. This is not ideal and adds lots of delay ( approximately 6-7 seconds), but that's better than nothing.
Overall the AMD data is not as accurate and is very unreliable. For example, they can have two release dates for the same cpu on one page.