Most “slowness” is not caused by CPUs performance but by network latency and differences in data scale combined with inadequate indexing or poorly chosen algorithms.
A 1MB empty database on localhost will run well on anything!
What’s needed in my experience is network simulators that can add latency, and full sized databases with real data.
I'd also add that a lot of the inefficiencies are a direct result of client demands. Not in the sense that they want it to be slower, but they refuse to budge on the requirements and do not want to adapt their business process. We bog down our validation code with dozens and dozens of edge case processing rules that could be bypassed if the client simply changed their own processes, but alas, no.
Most “slowness” is not caused by CPUs performance but by network latency and differences in data scale combined with inadequate indexing or poorly chosen algorithms.
A 1MB empty database on localhost will run well on anything!
What’s needed in my experience is network simulators that can add latency, and full sized databases with real data.