True, i did not take IPC into account. However, both chips i mentioned are from the same year, so i'm not sure if they differ in IPC? Are there generally any differences in IPC between server and consumer grade CPU's? From my understanding, the e5 1620 is basically an i5/i7 from the same era in terms of structure, and just slightly modified, so i guess there wouldn't be any difference?
I can definitely see how having more cores/threads is gonna benefit most stuff today. IIRC back when Intel launched Haswell, the general consensus was that the i5 is the best gaming CPU, because games wouldn't benefit from the extra threads of an i7. I haven't actually kept myself updated in this regard, but i'm gonna assume newer games make much better use of all the cores/threads, seeing as how Ryzens outperform the i5 lineup (or does this come down to IPC?)
In this case it doesn’t. The overclock ability of the i5 could make the chip faster than the e1650 in some things. It would have to be a good overclock though. With them both sitting not overclocked the e1650 is faster. Is it possible that with enough overclock you could get the 4670k to run some stuff faster than the e1650? Sure. Not a whole lot though, and the problem the 4670k runs into is if it hits a need for more than 4 threads it starts creating overhead. Most modern games are 6 thread, because that’s what jaguar1 had available for games. That means that running those games the 4670k will have 2 threads worth of overhead all the time where the e1650 won’t. This is what killed the 5ghz single core chips and made the 2.4ghz duos crush them.