The BBVA Foundation announced this Wednesday that it has awarded the Frontiers of Knowledge Award in Information and Communication Technologies to John Hennessy (Huntington, New York, 1952, currently at Stanford University) and David Patterson (Illinois, United States, 1947, University of California at Berkeley), for, according to the institution, “founding as a new scientific area computer architecture, the discipline that designs the brain of every computer system, its central processor ”.
“Professors Hennessy and Patterson are synonymous with the creation and formalization of computer architecture,” says the minutes of the BBVA Foundation, according to the press release issued today. “Before his work, the design of computers – and in particular the measurement of their performance – was more an art than a science, and professionals lacked applicable principles to conceptualize and evaluate it. Patterson and Hennessy first built a conceptual framework that brought a robust approach to measuring the performance, energy efficiency and complexity of a computer to the field. “
The scientific contributions to computer architecture by Hennessy and Patterson, says the BBVA Foundation, have also had a didactic aspect, reflected in a textbook, Computer Architecture: A Quantitative Approach, which three decades after its appearance is still considered the Bible of the discipline in universities around the planet, having reached six editions in which its content has been updated.
Hennessy and Patterson have not only made extraordinary contributions to computer architecture as researchers and educators, projected to several generations of computer scientists through their teaching activities, but they have also promoted technological innovation and the business development of their ideas. . Both laureates created RISC, an architecture on which the design of central processors is based and today it is “at the heart of virtually all data center servers, desktops, laptops, smartphones and embedded computers. [en televisores, automóviles y dispositivos del Internet de las Cosas]”, According to the minutes.
RISC, the acronym for Computers with Reduced Instruction Set, was developed in the eighties based on a concept that both laureates advocated, and that Hennessy, contacted after the ruling, has summed up in the phrase: “The simpler, more efficient”. It was a principle completely opposite to those then prevailing among computer designers, and it was precisely this going against the current that prompted the two winners to collaborate.
A systematic and reproducible method
For Patterson, a common thread in their work is the effort to apply a systematic and reproducible method to their research. Thus they contributed to the formalization of the area of computer architecture, thus they created RISC and thus they wrote their book: “We design processors just like we design books: doing experiments and rehearsing,” he explained by videoconference after hearing the jury’s decision, whose members have been jointly appointed by the BBVA Foundation and the CSIC.
Proof of its success is the fact that RISC technology, and its principles of maximum efficiency, are now present in 99% of processors in all types of systems, and make possible from the lightness of laptops to the long duration of the battery of smartphones. Regarding his textbook as an instrument that has defined the pillars of a discipline and that plays an essential role in its transmission, Hennessy has assured that “one of the greatest joys of my life is to see that there are students from all over the world who they appreciate our work ”.
“The work of Hennessy and Patterson has had a profound and lasting impact,” the minutes conclude. “They conceived the scientific field of computer architecture, pushed for a systematic and quantitative approach to system performance, promoted a class of processors with reduced instruction sets that has transformed the way the industry builds computers; and they have made transformative advances for computer reliability and large-scale system consistency. “
From ‘dark art’ to science ‘within everyone’s reach’
In the 1980s, before the award winners, “each company had its own way of designing processors, based on the intuition of a few experts,” recalls Ron Ho, director of Silicon Engineering at Facebook and secretary of the jury. “It was like witchcraft, a series of difficult methods to transmit. Hennessy and Patterson changed that, created a framework, defined parameters that made it possible to compare different systems in efficiency and speed. Computer architecture stopped being a dark art to become science, a systematized knowledge that can be transmitted. They democratized – made available to everyone – the knowledge necessary for computer design ”.
Both awardees recalled the effort to put the fundamentals of discipline black on white: “Both John and I were teachers, and we love teaching. We decided to write a book because it frustrated us that none of the books we had helped us pass on what we knew to our students, ”recalls Patterson.
Hennessy, for his part, adds: “Computer architecture was taught descriptively in universities. It was as if you were visiting a museum like the Prado, admiring individual works. We didn’t like this approach at all, we wanted something based on important metrics like performance and cost. So we started writing with this goal in mind, so that the field would take on a scientific and engineering focus, rather than one based on mere descriptions. 30 years have passed and the book is available in a dozen languages, used by hundreds of thousands of students around the world ”.
Simple is effective
The RISC processor, born in the early eighties, is also the result of a conceptual change. Computer designers of the day assumed that the processor would be faster if it had to execute few instructions, no matter how complex. Hennessy and Patterson demonstrated with their RISC prototype that if the processor executed more, but simpler, instructions, computer programs would execute much faster.
“RISC is mostly about efficiency,” Hennessy explains. “The key idea is: the simplest is the most effective. Think of a literary essay with very complex words and elaborate sentences. It is very difficult to read quickly. In contrast, an essay with simple terms reads quickly. This is what RISC does: it uses very simple words that run fast. This allowed us to achieve a breakthrough in the performance of computers, which today has achieved great advantages in terms of efficiency and energy use ”.
From initial rejection to use on almost all devices
The good results of RISC, however, initially were not enough to convince the community. “It was too disruptive, counterintuitive technology: if you make things simpler, the computer will go faster,” Hennessy recalls. Patterson agrees: “At first, RISC was incredibly controversial. John and I were alone on one side of the debate. But years later this was reversed. We identified the formulas that explained why RISC was more efficient, and this led to the creation of startups. John created his own company [MIPS Technologies] and I collaborated with an existing one [Sun Microsystems], which brought highly successful products to market. This changed the opinion of the community ”.
The appearance in the last decade of devices that are both smaller and more powerful, such as the smartphones and tablets, made the benefits of RISC technology even more apparent. More energy efficiency translates into longer battery life and lower cost.
Beyond their joint work to formalize computer architecture and design RISC technology, the jury also noted that both honorees have contributed other lasting innovations. Patterson “created a field of study around the reliability of computers,” the report notes, while Hennessy “worked on the development of distributed shared memory multiprocessor systems.” Taken together, the jury believes that these ideas have laid the foundation for modern data centers, databases, and internet search engines as well. All these systems and tools require the continuous support of highly reliable large-scale computers, which could not have been developed without the multiprocessors created, in turn, with the aid of the innovative vision of the two winners.
A new ‘golden age’ for computer architects
The race towards even further miniaturization of processors is not over, and that means, for the winners, that the challenges for computer architecture will be even greater in the short term.
“Moore’s Law [el número de transistores en un procesador se duplica aproximadamente cada dos años] it has pushed the technology to build increasingly efficient computers, but it is coming to an end, “says Patterson. However, people still want much faster computers. This means that computer architects are going to have to figure out how to design computers that will not have higher transistors, but will still improve performance. The field may reach a new ‘golden age’ ”.
Hennessy, for his part, believes that one of the greatest current challenges for the discipline is imposed by the rise of artificial intelligence (AI): “The demand for AI performance is increasing dramatically, we will have to rethink how to design computers to do highly complex tasks, such as machine learning, efficiently. This will bring exciting innovations, and represents a great opportunity for young people to make really important contributions. “