Posting a super pi score in the era of multicore processors is like driving your Porsche on the narrow village road going 20Km/h... (I know yup posted it to prove how "fun" that Opty was,yet Super Pi is fail from perf. evaluation POV when you count multicore MPUs).
Also comparing Bloomfield and Istanbul is valid since Bloomfield is still very high end with only 1 Westmere hexa core being faster. Bloomfield design relies massively on added logic(SMT aka Hyperthreading) which adds ~20% on the die size while giving similar perf. increase on average in poorly coded multithread-capable applications .Each core inside Bloomfield is MASSIVELY larger than one in Deneb/Istanbul,thing 15.3mm2 Vs 24.5mm2,that's staggering 60% more added core logic area. Deneb is actually doing great compared how much smaller its cores are.So Istanbul/Thuban with 50% more cores actually levels the field in core logic area investment .The only real downside in AMD's Deneb is quite large sram cell size and this is why Deneb and Bloomfield take up similar die size in total(while very cores in Deneb are,like i previously said 60% smaller and bound to be slower in MT applications).

As for your wet dream of 6-10 cores on desktop let's see how that turns out.I heard not so soon.