I am student of Computer Architecture and i was wondering how come apple dual core processor's performance working with 1.4 GHZ clock is more than 8 core android phone processors?
I searched for it and i came to know that apple is building its processors with a wider and more complex pipeline that also has a great deal of focus on maximizing memory bandwidth.
The most important aspects when it comes computing performance is memory bandwidth, latency and ability to execute threads in a more efficient way. This governs the instructions per cycle (IPC) that a processor can push put.
In simple words apple is focusing on complex pipelining to improve IPC.
So here are my some questions that i want to ask.
I am little bit confused with the current architecture of core i7 whether its pipeline structure is of 20 stages or 24 stages? Moreover if we want to improve the performance of current architecture of i7 can we achieve this by increasing stages of pipelining as apple is doing?
No doubt that the complex pipelining will result in more hazards like Structural and Data hazards but there are certain ways we can remove these hazards i.e. Forwarding ,Stalling.
IF somebody has better solution instead of increasing pipelining stages ?please share.
Also if some body has the document reference of current pipelining structure of i7 please share it with me...