Comment on page
Computing Power at Our Fingertips: The Cornerstone of An Intelligent Society
Photo credit — Automation Engineer
The future demand for computing power will be enormous. The undergoing smart society process involves using such key technologies as Artificial intelligence (AI), blockchain, Internet-of-Things (IoT), and Augmented Reality (AR)/Virtual Reality (VR) in different scenarios to boost productivity, improve people’s lives, and make the society more efficient. Wide application of these technologies is bound to engage more computing units, driving computing power to grow steadily. Specifically, by 2030:
- AI will penetrate all industries and will carry a computing power of 16,000 EFLOPS (combined computing power of 160 billion Qualcomm Snapdragon 855 NPU2 embedded in today’s smartphones);
- Blockchain technology underpinning crypto currency and other areas will engage a computing power of more than 5,500 EFLOPS (combined computing power of 1.3 billion units of AntMiner V9);
- IoT will connect all appliances in factories and households, and require 8,500 EFLOPS of computing power deployed on the cloud and the edge (combined computing power of 7.9 billion chips used in high-end IoT edge devices);
- AR/VR/Metaverse, when playing to its full potential, will have more than 3,900 EFLOPS of computing power (combined computing power of 2.1 billion units of SONY PS4 consoles).
To meet all the above demands, computing power in the future should be available Anywhere, Anytime, at Any Capacity, and in Any Object.
To better understand computing power and effectively measure it, researchers have put together a Computing Power Evaluation Model. It measures the computing power of all typical computing units. By calculating and comparing the computing powers of 27 countries, the research concludes that a country’s computing power is tightly coupled with its level of economic development.
According to the 2021–2022 Global Computing Power Index Assessment Report jointly released by Inspur and IDC on March 17, 2022, the proportion of the digital economy is expected to reach 41.5% by 2025, driven by the continuous development of the global digital economy. At the same time, the national computing power index and GDP trends show a notable positive correlation. Every 1% increase in the computing power index of 15 key countries will lead to 3.5‰ growth in the digital economy and 1.8‰ growth in GDP. This trend is expected to continue.
In addition, when the computing power index of a country reaches 40 or above, the driving force of the country’s GDP growth will increase by 1.5 times for every 1% increase in the computing power index. This increase jumps to 3 times when the country’s computing power index reaches 60 points or above, boosting economic growth in a more significant way.
Along with silicon-based chips, computing power has evolved through three stages, namely, “single-core,” “multi-core,” and “networked.”
The computing power of single-core chips will peak at 3nm. Moving down from 3nm would be too difficult to manufacture and too expensive to commercialize. Multi-core chips seem to be the solution — but not a perfect one. As more cores stack up, dissipation effects and rising power consumption will make the trade-off of more cores for more computing power uneconomical — the tipping point is at 128 cores.
When the capacity of single-core and multi-core chips is exhausted, “networked” computing power will come in to fulfill the demand gap. Meanwhile, to break free from the limitations posed by the existing cloud technology and bandwidth cost, deploying computing power distributedly on the edge will be a must to address the demand-supply mismatch. Eventually, an ideal networked, self-organized, and ubiquitous computing power deployment architecture will take form.
While striving to develop computing power and harvest its advantage, there are three challenges that should be addressed.
· Unaffordable computing power. As new and experimental AI technologies demand significantly more computing power, the cost will rise exponentially with the traditional cloud models. Potential solutions may include spreading out the heavy-capital data centers to crowdsourcing devices and incentivizing private capital to build computing infrastructure.
· Surging power consumption. Power consumption of data centers accounts for 3% of global carbon emissions and is estimated to grow at 6% annually. The industry should encourage deploying energy-conserving and green computing power infrastructure, ensuring sustainable development of the smart society.
· Security threats. While building and operating a smart society, the internet community should be able to monitor and prevent potential security risks. Feasible solutions may include: 1) establishing a community-owned sovereign network to allow participants around the globe to secure the network, 2) establishing a public-funded supervising authority for computing power security and allowing third-party supervision, and 3) establishing a trustless accountability mechanism to regulate market participants.
Over the past 3 years, the exaBITS team has committed to delivering new and innovative ways to meet the demand for computing. With these efforts, we hope to provide the world with more abundant and affordable computing power, create long-term value for our customers, and contribute to society.
In the next 30 years, AI will become deeply integrated into all industries, driving deep changes in each and every one. It will bring brand-new experiences to every home and every individual, fundamentally changing how we work and live. To create a bright future with ubiquitous computing power, everyone in the industry needs to go all in.
exaBITS is ready and willing to work with all of our partners in the computing industry and make “ubiquitous computing power” a reality.