On Oct 16, 2011, at 10:09 PM, Mark Humphreys wrote: > On Sun, Oct 16, 2011 at 6:49 PM, Robar Philip <[email protected]>wrote: > >> Notice though that its 8 cores are not a win for builds, whereas the build >> test is one of the few tests where the core i7 2600K really separates itself >> from the rest of the pack—being almost 30% faster than its nearest >> competitor and 50% faster than the rest of the pack. >> >> Does anyone have any idea as to why the core i7 is such a win for builds? >> >> Phil > > Hyper-threading, which was reintroduced with the Core i7 / Nehalem chips, > combined with techniques like branch-execution prediction and higher clock > speeds make the Intel offering more powerful; but, of course, more > expensive. > > http://en.wikipedia.org/wiki/Hyper-threading
Ah, I knew that my Core i5 2500S didn’t have hyper-threading, but I didn’t realize that none of Core i5s have it. I guess Intel was smart enough to know that a hyper-threaded i5 would perform so closely to an i7 that it would remove much of the incentive to buy an i7. It amazes me that the CPU market is such that it makes sense for Intel to make 39 different Sandy Bridge desktop CPUs*. And if the myriad of variations in speed, turbo mode, L3 cache, graphics performance (normal and turbo) and power usage were’t enough to numb your mind, then a little digging at CPU World will show you that Intel turns off features like virtualization support for various models—so you can’t have your cake and eat it too. Phil * http://en.wikipedia.org/wiki/Sandy_Bridge#Desktop_processors _______________________________________________ OpenIndiana-discuss mailing list [email protected] http://openindiana.org/mailman/listinfo/openindiana-discuss
