Ideal scaling for benchmark plots

[jkoefinger]

Expected behavior

Ideal scaling should go through origin and first node

Actual behavior

Ideal scaling should goes through first and second node

Your version of MDBenchmark:

(run mdbenchmark --version on your machine)

[ksebaz]

@jkoefinger if you want to test, #193 now contains the new scaling projection. It looks like this. Is that as you intended?

[balazsfabian]

@ksebaz, #193 indeed produces the correct scaling