Q&A: Cadence's Vucurevich On Processing Power's Continued Importance
All current gaming hardware features specialized chips, but we don't often think of the importance of the architecture design at the chip level. One man who is acutely aware of the importance is Ted Vucurevich, Chief Technology Officer at chip design software specialist Cadence Design Systems.
Last October, Cadence conducted a roundtable
with personnel from high-profile chip manufacturers AMD, Intel, NVIDIA, and IBM, discussing the future of gaming processors with respect to both the console and PC arenas.
Gamasutra recently sat down with Vucurevich to drill down into Cadence's role in the whole process. As CTO, Vucurevich is responsible for driving research and development at Cadence, and in his prior role as the company's chief architect he helped develop strategies and technology initiatives for system-on-a-chip based design, software interoperability, design methodology development, and Internet-based electronic system design.
Vucurevich shares his vision of gaming hardware's future – including "heterogeneous computing platforms" and the democratization of interactive entertainment.
What is Cadence’s importance to the games industry?
Ted Vucurevich: The gaming industry has advanced to where we are today because console designers have been able to deliver twice the functionality at half the cost per function every 18 months for over 20 years. To be able to accomplish this, designers have had to rely on sophisticated design software provided by companies like Cadence. In effect, we enable the features and functions in the game computers and consoles to be created.
What have been most important technology developments for the games industry recently?
TV: Without a doubt the two most important developments have been the introduction of multi-core architectures in the CPU and the introduction of general computing capability in the GPU.
Where do you see it developing in future?
TV: The major manufacturers are working on what I would call a “heterogeneous computing platform." In simple terms this means that the boundary between general computing such as what is done in the CPU and specialized computing which currently is done in things like GPUs and audio processors is blending. What this means to game developers is the opportunity to improve multiple functional areas in their software such as AI, physics, real world interface (the Wii controller is a first-generation example) and, of course, graphics. A good example of this is Intel’s tera-scale project
High-end PC games have, at least over the past 10 years, driven technology uptake by game developers, but is that tailing off due to piracy?
TV: The chip development industry would undoubtedly be crippled if piracy substantially impacts economic return to developers, but I believe one trend for high-end games will be more of the subscription model as opposed to individual software unit sales.
This approach can limit piracy losses, and when Blizzard starts to introduce its content on consoles, I believe we will see an acceleration of this way for developers protect their investment.
What do you see as differentiating the current generations of consoles?
TV: First off, the 360 and PS3 are about a half-generation apart in my opinion. That being said, from a consumer level, the only real differentiation is all related to software.
On one hand, most of the developers have taken a multi-platform support approach and we essentially get non-differentiated game play, visuals, and audio. For the significantly smaller number of developers and titles which are console specific, the differentiation tends to be more in game play, storyline, and in general quality of the software titles rather than the hardware platform it is running on.
Of course at a "spec" level there is differentiation but turning that into a value proposition at the consumer level is exceptionally challenging.
Indeed, the Wii has changed a lot of things –- do you see future developments concentrating more on interface rather than processing power?
TV: Definitely! Nintendo was exceptionally successful in introducing a more "natural" way for people to interface with the virtual world, and as a result took a huge step in "democratizing" the use of consoles for interactive entertainment. This offers content providers a broader pallet of concepts that can be explored which moves the software from "just games" to "entertainment," which broadens the total market opportunity.
As far as processing power goes, I expect that consumer demand for real time interfaces will eventually drive their own set of performance and scaling requirements which will increase over time. Real time position tracking, tele-presence and latency will all drive more demand on processing power. If we look a bit further into the future at some of the amazing wireless sensors and actuators under development, you can envision a point where we are able to provide sensory immersive entertainment environments.
So you don’t think we’ve reached the limits of requirement for processing power for gaming?
TV: Only in the most traditional way of looking at processing power for gaming. Heterogeneous, distributed computing architectures will allow us to naturally interface the virtual world and to, in a multi-sensory way, interact with it. In my opinion, this will offer exciting (and profitable) opportunities to consume all the "processing power" we can offer for years to come.