I am running a few websites on a VPS and I am planning to upgrade it. The thing is that when you go up in ram you automatically end up with more but slower cores.
Example:
- 6GB ram = 2x2.4 GHz @ 100%
- 8GB ram = 4x2.4 GHz @ 60%
Now obviously 4x 0.6=2.4 meaning I "gain" 40% compute power. However a single page visited by a single user is not multithreaded. So I kind of expect calculation time needed to server a page to increase by the same percentage.
So I guess I am expecting to be able to handle 40% more traffic BUT my website will load 40% slower for everyone due to the loss of 960 Mhz per core. Now it is a VPS so I could try for a few days. But I am unsure how to consistently test what is best. Tests I have used in the past have been inconsistent over time to say the least...
Is there anyone with a great idea how I can make an informed decision about this?
[link][4 comments]