A new era of diversity computing has arrived as the world shifts to data-centric. As new workloads such as artificial intelligence, cloud data centers, the Internet of Things, next-generation networks, and autonomous driving continue to emerge, future computing innovations must build different types of computing architectures such as GPUs, FPGAs, AI chips, and vision processing chips to meet the needs of diverse workloads on the basis of current CPUs.
As a leader in computing, Intel proposes six pillars of technology that continue to expand product leadership, laying a solid foundation for shaping the future heterogeneous computing landscape and meeting diverse computing needs. Not only in the classical computing field, but also in the field of new technologies such as neuromorphic computing and quantum computing, Intel is also actively advancing.
In neuromorphic computing, Intel released the Loihi chip, a single chip based on a 14-nanometer process that integrates computing and storage in architecture designs, with 128 cores, each with 1,000 neuronal computing models. A single chip can simulate 13 000 neurons, support multiple learning modes, and support a way similar to the human brain.
“There seem to be a big gap in the human brain with 86 billion neurons, but Moore’s law continues to work, and encapsulation technology is advancing, reaching the number of neurons in the human brain is only a matter of time and engineering. Song Jiqiang said. In neuromorphic computing, Intel has established a global neuromorphic research community, INRC, and now has 75 organizations joining it.
Quantum computing is also a key investment area for Intel. In his view, quantum computing is mainly used to solve the classic computer’s indeterminate large-scale computing problems, such as password deciphering, biomedical, chemical and other combination explosion problems.
In the race to realize the function and potential of quantum computers, researchers have focused extensively on the manufacture of qubits and building test chips to demonstrate that a small number of qubits running in superposition can improve computing at exponential levels. Intel Research has released its first cryogenic control chip, codenamed “Horse Ridge,” which will accelerate the development of a full stack of quantum computing systems.
Because the current quantum computer operates in the millimeter temperature range, it is only a fraction of a degree higher than absolute zero. But the properties of the silicon spin qubit allow it to operate at 1 Kelvin or higher temperatures, which greatly reduces the challenge of cooling quantum systems. As research progresses, Intel’s goal is to have cryogenic control and silicon spin qubits work at the same temperature. Intel was able to leverage its expertise in advanced packaging and interconnect technologies to create a solution that integrates qubits and control devices into thin packages.
Song Jiqiang said that although the quantum computing industry is progressing very fast, but from the point of view of the practicality of quantum computing system, there are both theoretical difficulties, engineering difficulties, cost-effective and enforceable need to be achieved, there is still a long way to go. In his view, quantum computing will not replace classical computing, nor will it replace devices such as mobile phones, laptops, and smart homes; On the path of technological evolution, Song Jiqiang believes that the silicon-based spin electron method is more reliable.