yes, to 5% noises
Please do not use noise limit for performance tests like this.
There are two ways you can measure performance:
1)
Use the same scene on computer A and B.
Set a specific pass limit (for example 50 passes).
Let it render for 50 passes on computer A and B.
Compare the resulting render times. The computer with shorter render time has better performance.
2)
Use the same scene on computer A and B.
Start rendering and observe the ray/s value in the Stats tab of the VFB.
The computer with higher rays/s value has better performance.