This is a completely useless, stupid, illiterate work!
You're comparing two different OS! It is the OS that distributes the resources to the applications! And it's up to the OS to decide which processor state to use for a given task, and the states for 45W and 15W CPUs are completely different! Notebookcheck is really going down.

This comparison is useless. You can't compare these things while being on different OSes. Comparing Windows 7 to Windows 10 here.

Exactly, i7-3520M is 35W... just forgot this fact, because for years now it's 15W 2/4 CPU everywhere

Very nice comparison, thanks for contributing. However, assuming that the wattage numbers you're supplying when specifying the CPU models refer to the TDP of each model, it should be 35W for the i7-3520M.

Source: https://ark.intel.com/products/64893/Intel-Core-i7-3520M-Processor-4M-Cache-up-to-3_60-GHz

(or just look at Notebookcheck's list of benchmark results for mobile processors)

Another intresting thing on 15W vs. 45W is up to X1.5 more FPS on World of warcraft for ... 15W model. 45W meets watts/temperatures limitations on 12.5" HP 2570p ultraportable

Last night i make this tests comparing:
i7-3820QM 3.7GHz 8M cache 4/8 45W (QM)
i7-3520M 3.4GHz 4M cache 2/4 15W (M)
on HP 2570p ultraportable

*Winrar 5.31 multy*
QM: ~7200 @ 27W 1.041V 77C
M: ~3600 @ 17W 1.056V 66C
x2 perf., x1.6 watts

*Winrar 5.31 single*
QM: ~1250 @ 19W 1.051V 72C
M: ~1100 @ 13W 1.056V 60C
x1.14 perf., x1.46 watts

*x264/aac 1080 compression*
QM: 12FPS @ 37W 1.036V 87C
M: 7.2FPS @ 24W 1.041V 77C
x1.7 perf., x1.5 watts

So i clearly see 45W QM is up to x2 faster (on MT load), but prefer 15W for heat/noise consideration. On ST load there is wattage overkill on the QM model.

My plan is to wait for 4/8 15W notebooks... hope they be almost same performance as an older 45W models