Clock speed has been one of the most identifiable aspects of processor performance during the last decade. But that mindset needs to change as process technologies shrink and alternative methods of increasing performance should be examined, according to most members of a panel of industry experts at the International Solid State Circuits Conference (ISSCC).
"In the past, performance has tracked directly with frequency. That's clearly ending," manager of systems technology and microarchitecture at IBM, Philip Emma, said.
Building a high performance microprocessor involves a series of trade-offs, and there is no one accepted way to increase performance. Some designers increase the chip's clock speed, some companies choose to improve the bandwidth of interconnects, and some explore ideas like multithreading and multicore designs. Most chip companies use some combination of the three to deliver ever increasing levels of performance.
With the chip industry in the midst of a process technology generation leap, increasing concerns about power dissipation are making some chip designers wonder if the days of higher and higher clock rates are done. Concerns about power dominated an earlier ISSCC presentation of cutting-edge processors, where just about every presenter addressed the issue of managing power dissipation.
"If you push frequency further, you have to spend power. If we push frequency any further, we're not going to be able to cool these things," a fellow with Advanced Micro Devices, Alisa Scherer, said.
The lone cheerleader for increasing gigahertz was principal architect on the Pentium 4 project at Intel, Doug Carmean. Bringing higher and higher frequencies to its Pentium 4 chips has served Intel well for several years, and the company plans to bring the newest version of the Pentium 4 to 4GHz by the end of the year.
Increasing clock speed was still the best way to advance performance for certain applications on the desktop side of the world, like three-dimensional rendering, Carmean said.
Extracting performance from single threaded workloads was difficult, but it work4r, and frequency was the most easily understood aspect of performance, he said.
Most PC consumers didn't understand exactly what made their machine perform, but could be trained to recognise other aspects of chip performance, Scherer said, alluding to the various aspects of automobile performance that enter into a buying decision.
"For every dollar you spend trying to manage power, you could spend on marketing your overall performance," Scherer said.
For many applications in the server world, frequency was not the answer, vice-president and fellow with Sun Microsystems, Marc Tremblay, said.
Sun had just released its first dual-core processor, and was working on multiple core designs where each core runs several threads, he said.
By moving to multiple cores, chip designers could also reduce the footprint of the chip die while maintaining or improving performance, chief technology officer with Fujitsu's enterprise systems group, Hisashige Ando, said.
Multiple cores each running at reduced clock speeds could also save power, he said.