Pioneering Architect of Soviet-era Computing
Published on: 28th Jun 2014
One of the pioneers of supercomputing during the Soviet era, who also happens to be a former champion mountain climber, has been working in the computing industry for 60 years, including the last 10 at Intel, but he's not quite ready to retire.
Boris Babayan, based in Moscow, joined Intel in 2004 as a fellow and director of architecture within the Software and Services Group (SSG) when Intel acquired Elbrus MCST. He leads efforts in such areas as compilers, binary translation and security technologies.
Babayan, 80, began his student career in 1951 doing some of the world's earliest work in computer science, including inventing one of the ways that computers execute calculations: carry-save arithmetic, which is still used today.
Here are five facts about Boris Babayan's six decades as a computer architect-before and after the fall of the Iron Curtain.
He was the first computer science student in the Soviet Union
Babayan was 7 years old when World War II broke out. Although the brutal Nazi occupation of the Soviet Union did not extend below the Caucasus Mountains into Azerbaijan where Babayan and his family lived, it was nonetheless, as Babayan recalled, "a very, very bad situation." Not only was there not enough food, but "there wasn't good paper" for schoolwork, so Babayan and his fellow schoolchildren scribbled their facts and figures in between the lines of discarded newspapers.
In 1951 at age 18, Babayan moved to Moscow to attend the Moscow Institute of Physics and Technology, founded by two eventual Nobel Laureates, Pyotr Kapitsa and Lev Landau. There he became the first computer science student in all of Russia-a field so new it was originally called "machine mathematics."
Said Babayan, "I am absolutely sure that I was the first in Russia, but I am curious whether it was a university discipline anywhere else in the world before 1951."
After his first year of studies, he became an intern at the Institute of Precise Mechanics and Computer Technology, led by Sergey A. Lebedev, whose 1958 BESM-2 computer was used to calculate satellite orbits and the trajectory of the first rocket to reach the surface of the moon.
Babayan currently serves as chair of the Institute and is a member of the Russian Academy of Science.
He was a principal designer of a seminal line of Soviet mainframes
While American processor designs were sometimes named after rivers, Soviet designs were often named after mountains. Mt. Elbrus lent its name to a series of Soviet (and later Russian) high-end mainframe computers that Babayan began working on while at the Soviet-era Institute of Precision Mechanics and Computer Technology.
According to the Computer History Museum, the Elbrus systems were intended for military-industrial purposes: "The combination of real-time control and high-performance computing is an unusual one; the resulting requirements were high performance, high reliability, and ease of programming."
The first two generations of Elbrus featured not only superscalar architecture but also had "very new technology in support of high-level languages," according to Babayan. "The technical name was 'capability system' and it was very successful," he added.
For the Elbrus 3, Babayan led the entire development of this 16-processor supercomputer, which used a newer VLIW (very long instruction word) architecture, with a compiler that ensured instructions being executed in parallel were not dependent on one another.
After the fall of the Soviet Union, Russian computer scientists began to work with the United States. "Our technology success was very interesting for Western companies like HP and Sun Microsystems," says Babayan. Sun worked with Babayan to form the Moscow Center for SPARC Technology in 1992, which was acquired by Intel in 2004.
He received two of the Soviet Union's highest honors
In 1987, Babayan was awarded the Lenin Prize-the highest prize in the former Soviet Union-for his development of the Elbrus 2 supercomputer and its out-of-order superscalar architecture. Earlier, in 1974, he had been presented with the USSR State Prize at the Kremlin for his work on computer-aided design.
He explained, "Binary instructions used are sequential, but hardware is very parallel. So superscalar, at run-time, converts sequential notation to parallel notation to speed up execution." Babayan's Elbrus machine used this in 1978, and Intel began employing superscalar architecture circa 1995.
"In the Western world, Intel was first. After that, the golden age of Intel started, and they became ahead of everybody," Babayan remembered.
"But we built the first superscalar in 1978-many years before Intel. And for the second generation of this design I was awarded the Lenin Prize."
He was a champion mountain climber
In 1952, Babayan began mountain climbing, an interest he pursued until 1976 when he traded it in for downhill skiing. This was during the Iron Curtain period, which meant it was all but impossible to leave the USSR, so Babayan climbed within the borders of the Soviet Union.
Babayan successfully climbed the very steep, 2-kilometer high walls of the Caucasus Mountains, and won a silver medal in the USSR National Championship of 1957 for ascending the peak of 26 Bakinskikh Komissarov (elevation 22,421 feet or 6,834 meters) in the Pamir-a range of mountains formed by the junction of the Himalayas with several other ranges.
Years later, when the first Intel Itanium 2 processor, code-named "McKinley" was being jointly developed by HP and Intel, Babayan conspired with some Russian climbers mounting an expedition up Alaska's Mt. McKinley (elevation 20,322 feet or 6,194 meters). As a nod to his work on the '90s-era Russian Elbrus 2000 (E2K), "We gave them an Elbrus flag. They put this flag over Mt. McKinley, and we proved Elbrus was better than McKinley," Babayan chuckled.
He has no desire to retire
After 60 years in the computer industry, Babayan says he has plenty of projects he wants to bring to fruition and has no plans to retire.
"I have many technical ideas. It is my opinion that without me it is very unlikely that these ideas will be developed well. So I am trying to continue.
"Maybe if I stop working it will be bad," he boomed, laughing. "Better to work!"