There are several key trends coming in Computer Architecture that calls for brand new technologies and approaches.

1. Application Advancements:

With Applications that handles Big Data (Volume, Velocity and Variety), need for scalable computer architecture is increasing. Also, recent years have seen the emergence of data intensive computing, embedded computing. mobile computing and rich multimedia applications, a revolution in computer architecture is a must to multiply the speed of processing.

2. Not only "Speed"

Performance doesn't say "Speed" only. A Computer system needs to be reliable, secured and energy-efficient to say that it is a high performing system. All these metrics needs to be revisited to adjust to the recent trends in application advancements and emerging needs of the computing industry.

3. Parallel Programming

With the advent of GPU Computing and High Performance Computing, I believe a change is required in the basic model of Load, Store and Compute Architecture. Multi-core programming and Multi-processing programming have lot of challeneges to be solved. Still there are gaps in building a reliable and high performance distriibuted systems, which Computer Architecture could potentially solve.

4. Instruction Level Parallelism

http://www.cccblog.org/2010/07/08/what-now-in-instruction-level-parallelism-research/

5. Excerpt from CRA Conference

The grand challenge is to make parallelism pervasive and parallel

programming mainstream in order to enable software to make effective use

of the widely available parallel hardware and continue the performance

improvement trend of the past several decades. Parallel software is the key

to unleashing the performance potential of parallel machines. The goal is to

reach the point where, when we say ‘computers’ we naturally mean parallel

computers, and when we say ‘programming’ we naturally mean parallel

programming. The goal is to make parallel programming accessible to the

average programmer