Top of the page

Can parallel processing resurrect Moore's Law?

Forbes.com has published an interesting article about the end of Moore's Law in the computer processing world. In 1975, Intel's Gordon Moore wrote that computing power would double every year (later revised to every 1.5 years). This prediction held true for decades. Until now.

Bill Daily writes that in the last decade, Moore's Law has ceased to hold true because energy consumption scaling has not been able to meet the demands of increased processor performance. Daily sees this slowing down of innovation as a ""crisis"", but he offers a solution.

Bill Daily, who is Chief Scientist at NVIDIA, makes the case for parallel processing (using multiple CPUs and/or GPUs that run slower and more efficiently) as the answer to the end of Moore's Law. Incidentally (or not), his company is at the forefront of such technology. Nevertheless, he proposes a very sound solution to keep increasing CPU performance in the short term.

But, he mentions no other alternatives to parallel processing. I have briefly covered on the blog research that is being conducted about materials other than silicon that could be used to make integrated circuits (such as graphene). It's possible that these other approaches could also allow for the energy consumption scaling needed to resurrect Moore's Law.

So while CPU performance improvements have not been charging ahead at the rate seen in the '80s and '90s, there are good reasons to be hopeful that other technologies will emerge to continue the race.

Can parallel processing resurrect Moore's Law?

Forbes.com has published an interesting article about the end of Moore's Law in the computer processing world. In 1975, Intel's Gordon Moore wrote that computing power would double every year (later revised to every 1.5 years). This prediction held true for decades. Until now.

Bill Daily writes that in the last decade, Moore's Law has ceased to hold true because energy consumption scaling has not been able to meet the demands of increased processor performance. Daily sees this slowing down of innovation as a ""crisis"", but he offers a solution.

Bill Daily, who is Chief Scientist at NVIDIA, makes the case for parallel processing (using multiple CPUs and/or GPUs that run slower and more efficiently) as the answer to the end of Moore's Law. Incidentally (or not), his company is at the forefront of such technology. Nevertheless, he proposes a very sound solution to keep increasing CPU performance in the short term.

But, he mentions no other alternatives to parallel processing. I have briefly covered on the blog research that is being conducted about materials other than silicon that could be used to make integrated circuits (such as graphene). It's possible that these other approaches could also allow for the energy consumption scaling needed to resurrect Moore's Law.

So while CPU performance improvements have not been charging ahead at the rate seen in the '80s and '90s, there are good reasons to be hopeful that other technologies will emerge to continue the race.