Hi everyone. I’ve been a lurker here for a while, mostly reading about data pipeline tools. This is a bit off-topic from my usual ETL worries, but I'm hoping for some advice.
I use Grammarly Desktop on my personal laptop for writing documentation and commit messages. It's an older machine—a Dell XPS from 2018 with 8GB RAM and an i5-8250U. I've noticed it can get a bit sluggish when Grammarly is enabled, especially in my browser (Chrome) with a lot of tabs open, or when I'm also running a local Airflow instance or a Python script.
My main concern is quantifying the "hit." I'm nervous about adding any background process that might interfere with my development work. I tried to check resource usage informally in Task Manager, but it fluctuates a lot.
Has anyone here done any proper benchmarking or has a similar setup? I'm particularly curious about:
- Idle CPU usage with the app running.
- Memory footprint difference when it's actively checking a large Google Doc versus just sitting in the tray.
- Any noticeable impact on battery life during normal use.
I'd love to hear about your experiences or if you know of a safe way to measure this without, you know, breaking my main dev environment. I thought about writing a simple Python script to log system stats with Grammarly on/off, but I'm not sure if that's overkill or if there's a better method.
I can totally relate to that feeling of watching Task Manager like a hawk when adding a new background tool. Been there with countless sales and marketing extensions that promise the world but then grind my older Surface to a halt. I haven't benchmarked Grammarly specifically, but I've had a very similar experience with some of the heavier CRM browser extensions for Salesforce and HubSpot.
On a similarly specced machine, I found the real killer wasn't the base RAM usage, which seemed modest, but the constant CPU spikes during any text input. It felt like it was doing a mini-analysis on every keystroke. The impact was most noticeable when I also had Docker running for a local database.
For your use case, I'd be most worried about that interaction with your local Airflow instance. Maybe try a simple test? Close everything, note your idle CPU/memory, then open a large doc and start typing rapidly. That's usually when the fan kicks in for me 😅