In theory: The Pentium G3220 and G3258 are effectively the same processor - they're the same version of the Haswell architecture. The only difference is the multiplier - it's set lower on the G3220 than the G3258, and the latter is unlocked. Altering the multiplier on the G3258 to match the G3220 will result in the same performance from both. At least, that's the theory - I haven't actually done any underclocking like this. How closely does the theory match up to the reality?
Yes basically. But for such thing there would be no good reason to. Only good reason to underclock is to try and make less power use, so you would need to go further by lowering the voltage and etc as low as possible while still being stable... then you basically just got your standard model to be like a low wattage S or T series.
Cool, thanks for that - just had that strange feeling when something seems too simple but makes complete sense. I'm aware there's no performance, stability or power consumption rationale for underclocking this chip in this manner, it's being done purely for the sake of a short-term comparison.
It should be noted the reason the G3220 is set lower is most likely because it failed in-house testing at the higher rate. So to avoid having to toss an otherwise perfectly good processor into the recycle bin for a total loss, Intel brands it with the lower spec moniker and still turns a profit (or at least not a total loss). Note this is a common, industry-wide practice. RAM and drive makers do something similar all the time.