Quad-core can help with multi-threaded applications and processes. So, in general, more active threads can be processed rather than the kernel and CPU having to context switch more in the dual-core CPU case. More context switches mean it can take longer to get work done because the CPU and kernel are working more to swap in/out threads and processes that all want a slice of the CPU's compute time.
As for the shared L3 cache sizes, there are usually L1-3 caches for CPUs. Each is typically increasingly larger in size. CPUs look to on-chip cache to see if frequently accessed instructions/data/memory addresses are present. If they are, then its much faster to retrieve stuff from on-chip cache than having to leave the CPU and retrieve data/memory addresses from the relatively slower system RAM. Thus with larger L3 cache, fewer accesses to system RAM are required. Databases, mail servers, and other apps that can store working sets in RAM can benefit from larger L3 caches.
At a glance, the 2014 and 2012 FSB (Front Side Bus) and system RAM are operating at 1600 MhZ. I'm not up to speed on the LPDDR3 and DDR3 RAM, so LP type could have lower latency and other benefits. Not sure how much of a noticeable benefit LP vs non-LP RAM provides.
I've generalized much, and am going from memory (rusty in this subject area), so apologies if I missed any of the finer nuances.
10-24-2014 10:50 PM
