Originally Posted by Derp
C2Q to Nehalem was a significant upgrade. Happy customers.
Nehalem to Sandy was excellent all around, much low power consumption, higher IPC and more clock speed head room. Again, happy customers.
Then sandy to ivy was mediocre. The IPC gain was partially offset by the lower overclocks because of the stupid tim + glue used.
Now we have Ivy to Haswell and it's the same mediocre snail's pace of improvements.
Basically, from January 2011 (release of sandy) to this day Intel has parked the bus. People are angry/disappointed and they should be. Haswell is on our radar because what the hell else are we going to buy? Piledriver? Please.
I'm not sure that they can make another big push forward without a major revision to the Core microarchitecture, if not replacing it entirely. The jump from Nehalem to Sandy may well have been the last big leap forward they could accomplish, and all they can do now is tweak it to get another few percent here or there, or reduce the power draw.
All designs reach their limit. When Intel was making NetBurst processors, they went from Willamette all the way to Cedar Mill, but then Tejas, the next version, was apparently so power-hungry that it would have made Bulldozer look like a "green" product. That was the end of the line. I think they're now getting near the end of the line with the Core processors.
AMD has an opportunity here. They still have a fairly new microarchitecture with lots of room to improve. Sure wish they'd bring out a real competitor with Steamroller and make Intel quit re-hashing the same old tech in the performance segment. It wouldn't surprise me if they don't have the successor to Core ready already, but they're going to sit on it until AMD puts out a product that makes them roll it out.