IBM's Chip Trail

A Big Blue exec sees power consumption continuing as a central concern.
Author:
Publish date:

Pimple-faced video gamers and anal-retentive system administrators both owe a debt of gratitude to Lisa Su. A 10-year veteran at IBM (SUNW) - Get Report, Su managed the PowerPC processor business, which makes the chips that run the back-end technology at many corporations, and was influential in developing the new Cell processor at the heart of Sony's (SNE) - Get Report upcoming PlayStation 3 game console.

As head of the semiconductor research and development center at IBM, Su is in the unique position to drive the semiconductor roadmap both at Big Blue and throughout the industry. She recently chatted with

TheStreet.com

TheStreet.com: What are some examples of markets and applications where the Cell processor's multicore architecture is going to have an impact?

Lisa Su

: We feel like the Cell processor has a lot of different applications. It may start first in the consumer area, but there are some higher-end applications as well, and that's where we're looking to expand and innovate.

Certainly, with the Cell processor we'll start with the games market. We also see it in some fairly high-end applications, for instance, weather forecasting. Being able to do lots and lots of computations in parallel is a good thing, and that's what you can do in multicore technology: Doing sort of complex simulations, financial simulations, a whole bunch of stuff like that.

We also see the capability, once you have a framework like this, to apply the same principles to multiple markets. So, you can see at the higher end you may use more Cell processors and at the lower end you may use fewer processors, but the framework is the same.

Power consumption has been a key driver in microprocessor design recently. On the portable side, will new battery technologies alleviate some of the need for microprocessors to focus on power consumption?

I think power will continue to be one of the places where we require a lot of innovation.

I would love to have better battery technology, but I think that this perhaps will not move forward quite as fast as we would like it to do. I would say we need innovation on all sides: semiconductors, cooling, packaging, system processor design, all of us have to be focused on power.

Now that the industry is moving to this multicore architecture, what does the design road map look like? Is it just about increasing the number of cores per socket?

The number of cores is one thing. I think also we like to look at just how much work you can do in a given piece of silicon area. So, from that standpoint, there are a lot of efficiencies that you want to get on the process technologies side. What we want to do is continue to keep shrinking the chips and put more devices on a given chip.

How is the business itself changing for companies in the microprocessor industry?

I think as technology is getting more complicated and as processor design is getting more complicated, there are actually advantages that will come into play for some of the more integrated companies.


The reason I say that is the design and technology is so hard to pull apart now, that you need constant communications and constant tradeoffs. You can't look at it as sort of I'm going to do my little piece, somebody else is going to do there little piece and we're going to put it together at the end.

So, that's how I see the technology affecting the business models. I think the closer the ties between the design teams and the semiconductor technology teams, the more successful the products are going to be. That's going to be a key differentiator. We call it a holistic approach.