Original plan was to upgrade to i7 Ivy Bridge. But I underestimated it getting a price cut. It was only around 2019 that I was able to secure a used i7-3770 for roughly $135.Wow, talk about an unbalanced system!
Wow, I guess the Ivy Bridge i7s held value pretty well. I wonder what a 3770k would have cost around then.Original plan was to upgrade to i7 Ivy Bridge. But I underestimated it getting a price cut. It was only around 2019 that I was able to secure a used i7-3770 for roughly $135.
Another reason I held off on the upgrade was that my Asus laptop had a i7-4770HQ.
I wonder what a 3770k would have cost around then.
When I tried firefox this week for a day.... it bloated to 20GB+ if left unchecked or until a tab crashed and then it would reclaim the RAM. The FF RAM issue reminds me of Chrome a while ago that had a similar issue though the tabs never crashed it just consumed to the max amount it could and kept running.
RAM size vs performance is usually a step function. If you aren't using 16 GB actively for program/data, going to 32GB won't make much difference. (Exception: very file-data intensive stuff using gigabytes of file data that the OS can cache in 32GB but not in 16GB. That's not a typical office use case, but might occur more often in content creation with numerous large active projects.)
I only need the 32GB on my primary computer when I'm doing multiple test runs in the background. I put 16GB into the cluster boxes, for cost reasons, and now I regret it; I never expected to load the cluster the way I do now on occasion, and I could probably cut 3TB TPCH load times by 1/3 or more with the added memory on each node.
So, the answer is "it depends". If you need 32GB you probably need it pretty badly. If you don't need it, adding it usually won't make a big difference.