Says Prof. Avi Friedman, University of Rochester and the Technion, who will participate in the ChipEx 2013 conference
One of the most important guests at the conference ChipEx 2013 To be held on 30/4-1/5 at the conference center of the David Intercontinental Hotel in Tel Aviv will be Prof. Eby Freidman.
Friedman is a professor of electronics and computer engineering at the University of Rochester and also a visiting professor at the Faculty of Electrical Engineering at the Technion and serves as a consultant to several companies. We interviewed him ahead of the conference.
"My specialization is in designing circuits for high performance. In this field I developed a series of commercial products when I worked in the industry for 12 years and published over 400 articles, about ten patents and 16 books during over 20 years in academia. I specialize in clocks on chips and energy supply, XNUMXD integration and fast circuit design and interconnection.
From your point of view, what are the chip industry's biggest challenges in the next five years?
Prof. Friedman: "There are several challenges. A few of them:
- We are seeing the end of the ability to go down even further to the nano level for chip manufacturing. Development of new technologies in the post-CMOS era, and the end of the race for better, faster and cheaper chips that drove the high-tech world for several decades.
- The elimination of the celebration in everything related to the energy consumption and efficiency of both the mobility and the integration in the systems
- Heat management
- Integration of systems and applications in heterogeneous technologies, what is known as the information fusion problem.
- There are many other challenges and their mutual influence can also be debated.
- Can we get new processors designed to help the device run for weeks or more without charging?
- Prof. Friedman: My initial answer is no. However, for some applications where we are dealing with the nature of the processor, a longer time between loads may be possible. There are also new architectures such as Apple Silicon that may help. Recently, significant work is also being done in the field of on-chip energy production that could improve the situation.
What can you tell about your collaboration with the Technion?
I have been working closely with the Technion for over a dozen years. I advise graduate students (both master's and doctoral students), I have taught several courses, given many seminars, conducted workshops on important specific topics in the field of high voltage, and I also jointly manage with Prof. Avinoam Kolodani the Research Center for Advanced Circuits (ACRC).
7 תגובות
Friedman is right that the existing silicon technology is coming to a halt in the field of miniaturization, reducing power consumption and increasing the difficulty of heat dissipation. This may be the main reason for Intel's failure in tiny laptops (cell phones and tablets mainly). Today, tiny laptops with Intel processors are hardly used.
Even the lithium batteries that are used today in mobile phones in general and tiny mobile phones in particular are quite stuck in terms of their development. The result is that laptops in general and tiny laptops in particular - operate without charging for only a few hours, this is a very annoying situation.
The tiny portable computers did not overcome the problems that Friedman pointed out, but rather increased the efficiency of the programming structure of the processors to a certain extent (almost all the tiny computers today use the programming structure of the ARM company which gives permission to use its structure to chip manufacturers, probably in exchange for royalties). How much ARM's efficiency will help is difficult to predict, in Friedman's opinion this efficiency is limited (meaning this efficiency will stop for a while).
As for Plurality's ability to offer a greater improvement than ARM's, this is a theoretical claim that has yet to stand a practical test. In this matter, ARM was stuck for 20 years (it seems to me) without being able to implement its product (and the theoretical research that preceded ARM products even preceded ARM products by decades); Only when a mobile phone market was created did they get out of the dead end where they were stuck for 20 years.
You can contact Professor Tzachi Birak and Ran Ginoser, both from the Technion, they know the technology very well and even expected a lot from it.
Odelia
So where is the technology today? Where can you see it working??
The Plurality company from Israel had/has technology that makes this article completely irrelevant.
Both in terms of performance and power.
The company is in liquidation due to the problematic conduct of the CEO who joined hands with a group of investors, which caused the rest of the investors from Israel and abroad to stop investing.
The technological results were extremely impressive. You can ask Professor Ginoser Ran and Prof. Birak Tzachi to prove my words
I've been seeing the "end of the race for faster chips" prognosticators for twenty years, and so far their prediction has yet to materialize. In my uneducated opinion (I haven't published dozens of articles and hundreds of books in the last twenty years) there is still a lot of space down there, and also a human desire to utilize it for uses that may seem completely strange to us at the moment, and also those who may profit hundreds of billions from an investment of tens of billions in the development of the subject.
I didn't understand, is he talking about the end of Moore's Law?
Because from other articles I read on the subject, I actually understood that the trend of increasing the computing power of digital chips relative to the price (cost/benefit ratio of electronic chips) is expected to continue for at least another 2-3 decades, if not beyond that.
For example, one of the trends is a shift to three-dimensional chips that are supposed to accelerate computing power, or a combination of cognitive chips, which IBM has already presented a prototype of recently.
The person sounds interesting including the topic but the article failed to take off unfortunately