Computer Systems Design       


Computer Systems Design - overview

It's an exciting time for computer systems design at IBM. We have many opportunities to develop enhanced systems for emerging workloads, especially in the areas of cognitive computing and IoT. Meanwhile, the importance of improving performance and efficiency for more conventional server-class workloads remains as high as ever.

IBM has pioneered many seminal computing innovations. Our expertise throughout the hardware and software stacks has led to industry-leading full system solutions. Although the major focus of our expertise typically has been the design of high-performance servers and mainframes -- the IBM POWER and z Systems, for example -- IBM has aggressively embarked on new research efforts over the last few years in anticipation of major industry shifts in computing workloads. Given the industry-leading base in general purpose high-performance computing, new accelerated IBM systems are poised for meeting the next challenges in computing and adapting to customer needs.

Anticipating and addressing future computing needs also shows the way to many opportunities. With diminishing returns of technology scaling, as well as consolidation in chip manufacturing, systems design may now play an even greater role in product differentiation. And despite the increase in technology challenges, accelerator-based heterogeneous architectures are a key path for both attaining computing performance and customizing systems to unique customer needs. However, more diverse hardware functionality, faster design cycles, and nimble design teams are all factors in domain-specific acceleration. IBM has a proven track record in architectural innovation and in the development of efficient hardware design approaches (See Large Block and Structured Synthesis  and SynTunSys). Thus, we anticipate a renaissance in hardware design.

Over time, fundamental semiconductor technology innovations may also lead to significant computing improvements. IBM has been at the forefront of semiconductor technology research throughout the history of the IT industry. Going forward, we might see innovations such as 3D integration and quantum computing combined with accelerated systems to reach the unprecedented computing needs that the systems and enterprises of the future will require.

The Computer Systems Design PIC encompasses multiple sub-disciplines ranging across various aspects of computer system design. Designing a complete system often requires collaboration across a number of these disciplines. The following links provide more details about specific CSD-related projects.


Last updated on May 16, 2017