Papermaster, M. (2017, April 13). How computing will change amid challenges to moore’s law. Retrieved from https://techcrunch.com/2017/04/13/how-computing-will-change-amid-challenges-to-moores-law/
To explain the current pace of speed improvements in developing CPUs.
Discussing current increased demands in computing power.
The difference between processing power in the cloud versus processing power in the local machine.
Addressing how to increase productivity in programming for local machine processing.
Addressing how to increase physical processing within the local machine
In Papermaster’s How Computing will Change amid Challenges to Moore’s Law (2017)he discusses the challenges between the slowing of Moore’s Law and the increasing demand for local processing power in machines. Moore’s Law is the hypothesis that every two years technological advancements will enable CPUs to double in speed. Papermaster’s assertion is that this speed has begun to slow down leading to the need to find other ways to continue the same pace of technological innovation.
The first idea is that processing needs could be moved to the cloud rather than on local machines. But much of the innovation happening is in areas like self-driving cars, drones, robotics, augmented reality (AR) and virtual reality (VR). Papermaster includes an example of The Royal College of Medicine using VR and AR to create disruptive applications in the surgical area. Ultimately, cloud processing can’t be relied upon with those innovative areas because of the need for real-time data processing.
Papermaster argues that with the integration of software and hardware engineering, we can overcome the loses from the slowing of Moore’s Law. He argues that if we build specialized accelerators using both CPU and GPU processors, using innovative techniques such as 3D die stacking and stacked memory, and integrate these devices because of the smaller size of the processing units these focused hardware devices can be used through open-source frameworks in order to take advantage of many distinct functions.
He calls this new technique the Moore’s Law Plus, and he argues that by combining the on-going CPU speed increases with the innovative integration of hardware and software that we can maintain the same acceleration of technological progression.
Moore’s Law is no longer valid as a two year cycle, it takes longer to move processor size and speed forward.
Cloud computing is not responsive enough or fail-safe enough to provide the computational power needed in certain settings.
Key innovative areas that are demanding more processing power are AR and VR interfaces, as well as self-driving cars, drones, and robotics.
In order to continue at the same technological pace, businesses will need to use both software and hardware integration by pushing both areas into new directions.
There is a danger of this making software development more difficult, it is important to keep software programming effort as a key factor when making these integration decisions.
I think that Papermaster’s article, How Computing will Change amid Challenges to Moore’s Law was well written and takes an interesting look into what life after Moore’s Law looks like. He doesn’t make much of an argument for the slowing of Moore’s Law and just states it as a fact. I do think that could have been better analyzed. However, the assertion is correct.
Intel’s famous “tick-tock” development process was originally designed around Moore’s Law where the first year was the tick (process) and the second year was the tock (architecture). By the third year, technology would have improved and they would be on to a new process. However, due to the slowing of Moore’s Law, Intel has had to develop a new development process that allows them to release new processors even when they haven’t been able to make physical changes every two years like the used to (Bright, 2016).
The recognition of cloud computing but the rationale on why it is not enough for certain areas of technology such as self-driving cars, AR or VR was helpful. I found it interesting that the solution was not in one discipline but was actually cross-displine by melding both software and hardware innovation.
Some proof of Papermaster’s assertions can be found in Apple’s new iPhone 8 and iPhone X. In order to power the AR of the new iPhones as well as capitalize on battery life Apple built 6 cores into it’s System on Chip (SoC) A11 Bionic. It has two power efficient cores for general use that allows the iPhone to have great battery life as it is used normally. It also has four high performance cores that allow for heavy duty multithreading when dealing with intensive tasks such as AR. The A11 Bionic SoC also contains a motion co-processor that off-loads all the motion tracking data away from the generic processors into a processor designed for dealing with motion data (Altavilla, 2017).
I think that Papermaster’s article was a through analysis of the challenges that businesses are facing as Moore’s Law is slowly become less and less true. There are some areas that could have been expounded upon by providing proof of Moore’s Law slowing or some concrete examples of how the software and hardware integration can lead innovation at the same pace. However, the premise and conclusion were still true.
Altavilla, D. (2017, September 13). Apple A11 Bionic Processor Silicon Prowess may be the new iPhones’ Most Impressive Feature. Retrieved from https://www.forbes.com/sites/davealtavilla/2017/09/13/apple-a11-bionic-processor-silicon-prowess-may-be-the-new-iphones-most-impressive-feature/ - 2cf7358c5f82
Bright, P. (2016, March 23). Intel retires “tick-tock” development model, extending the life of each process.. Retrieved from https://arstechnica.com/information-technology/2016/03/intel-retires-tick-tock-development-model-extending-the-life-of-each-process/