HBM1 wont be faster than it is.
Is this meant to be tautology? It is what it is!
Overclocking doesn't count.
. . . until an OEM launches a product at those same clockspeeds and puts a warranty on it, right? If someone wanted to sell a video card with 300/600 mhz HBM, they'd have it yesterday. HBM will be supplanted by HBM2 fairly soon so arguing overmuch about it is rather pointless. But it's obvious that the technology is not limited to 250 mhz.
HBM2 will double the clock, still be far behind in latency.
Still behind what? GDDR5? Again, Macri disagrees with you.
And no, its still not better latency than DDR2/3/4.
Mostly I look at the average memory latency of an entire platform. Sadly I don't have Sandra kicking around right now, but Aida64 shows me that the latency of my DDR3-2400 CL 10 system (74.2 ns) to be inferior to that of an A64 3200+ running DDR400 CL2.5 on an old ASRock 939S56-M (72.4 ns). In terms of cycle latency, that A64 is doing a lot better, since it was only clocked at 2 GHz for the test. And it's only CL 2.5! CL 2 would have yielded superior results, as would higher clockspeeds.
The top latency entry is an x2 6400+ @ 3.2 GHz running dual DDR2-800 CL4 (55.2 ns).