Generally speaking, companies don’t like to announce any ‘flops’ in their performance. That is, unless you are in the world of supercomputing, where FLOPS are a good thing indeed.
FLOPS stands for “floating-point operations per second” and the bigger your number, the better.
In fact, FLOPS can get so large they need a prefix so that people don’t need to say things like “this machine has a capacity of five quintillion FLOPS”.
So before we get to Alibaba Cloud’s new supercomputer – launched last week – here is quick run-down of the lingo.
FLOPS are often described as gigaFLOPS, teraFLOPS, petaFLOPS and exaFLOPS (Or GFLOPS, TFLOPS, PFLOPS and EFLOPS for short). In the future we will also welcome machines capable of zettaFLOPS and yottaFLOPS.
Alibaba Cloud’s new supercomputer cluster in Hebei has a maximum speed of 12 EFLOPS, which means it is capable of processing 12,000,000,000,000,000,000 floating point operations a second.
Put simply this means that Alibaba’s cluster in Zhangbei is the most powerful non-quantum computer in the world. It is faster than Google’s Oklahoma-based cloud cluster (9 EFLOPS) and much faster than the world fastest institutional supercomputer, the Frontier at the US Department of Energy’s Oak Ridge Laboratory (1.1 EFLOPS).
Which isn’t to say that Alibaba’s new machine cluster is the fastest in the world – even supercomputers are outclassed by quantum computers, whose computing prowess is measured in qubits, not FLOPS.
So what is Alibaba’s new cluster going to be used for? Like Google’s Oklahoma cluster, it is available for rent. If a company is working on a complex product that needs huge amounts of data analysis, they may want to use a supercomputer to simulate complex scenarios. Such modelling could drastically quicken their product’s time to market.
Similarly, if an academic institution wants to analyse a massive dataset or apply artificial intelligence to a set of texts, it might benefit from access to enhanced computing power.
Alibaba has said its supercomputer is likely to be used for AI-related work such as designing and programming autonomous vehicles.
Alibaba’s Feitian intelligent computing platform, which underpins the new cluster, has already been used by XPENG Motors – of which Alibaba is a shareholder – in the planning for the construction of a major support facility for autonomous driving, for instance.
Intelligent computing is dependent on deep sources of processing power and energy that normal computers can’t handle. “Intelligent computing power is actually a costly resource. In contrast to general-purpose computing, intelligent computing requires massive data to train AI models, and computing power is wasted in data migration, synchronisation and other links,” Cai Yinghua, president of Alibaba Cloud Intelligence Global Sales said in a press release.
Alibaba claims the Feitian system massively reduces these losses, making its supercomputer capacity worth the rental fee.
In addition, the new centre at Zhangbei has tried to maximise its energy efficiency and adopt greener ways of powering and cooling its machines.
If a client requires more than 12 EFLOPS of capacity, Alibaba Cloud already has another 3 EFLOPS ready for access at another intelligent computing centre in Inner Mongolia which can be run in tandem.
“Computing power, or the ability to process data, is widely regarded as a vital foundation for bolstering the development of the digital economy and a new engine to unleash the potential of data as a factor of production,” ChinaNews.com wrote.
Alibaba says that companies in China are starting to adopt artificial intelligence in much greater numbers and that they already have a higher usage rate for AI techniques than the global average.
China’s total computing power scale now ranks second in the world, adds the Ministry of Industry and Information Technology.
© ChinTell Ltd. All rights reserved.
Sponsored by HSBC.
The Week in China website and the weekly magazine publications are owned and maintained by ChinTell Limited, Hong Kong. Neither HSBC nor any member of the HSBC group of companies ("HSBC") endorses the contents and/or is involved in selecting, creating or editing the contents of the Week in China website or the Week in China magazine. The views expressed in these publications are solely the views of ChinTell Limited and do not necessarily reflect the views or investment ideas of HSBC. No responsibility will therefore be assumed by HSBC for the contents of these publications or for the errors or omissions therein.