Big Ideas Lab: Supercomputing at Lawrence Livermore National Laboratory

Episode Release Date: September 24, 2024
Host: Rob Neely
Guest Speakers: Michael McCoy, Becky Springmeier, Edward Teller
Podcast: Big Ideas Lab by Mission.org

Introduction: The Dawn of Supercomputing

The episode opens with a historical recount of the inception of supercomputing at Lawrence Livermore National Laboratory (LLNL). Rob Neely sets the stage by transporting listeners to July 1960, where a team of dedicated engineers and scientists are assembling the Univac Livermore Advanced Research Computer (UNIVAC LARC). This colossal machine, weighing 135,000 pounds and standing eight feet tall, symbolized a revolutionary leap in computational power and marked the beginning of the supercomputing era.

Rob Neely (00:13):
"This was the Univac larc. It was the most advanced computer of its time... marked the dawn of a new era in computing. The era of supercomputers."

Early Computing Efforts at LLNL

Becky Springmeier provides insights into the early use of computers at Livermore Labs, emphasizing their primary mission centered around the nuclear weapons stockpile. Before electronic computers, human "computers," often women performing manual calculations, were integral to the lab’s operations.

Becky Springmeier (04:09):
"We're a nuclear weapons lab, and the mission here has always been rooted in the nuclear weapons stockpile... computers were used to give the weapons designers of the time insight into really how to design these new weapons."

Michael McCoy reflects on the limitations of early computers like the Univac, highlighting their rudimentary memory and the cumbersome nature of programming in machine language.

Michael McCoy (05:29):
"It was a univac, a Remington Rand machine probably weighed tons with tubes and must have been almost impossible to use... you had to do it in something called machine language, was very cumbersome."

Transition to Supercomputing: The ASCII Initiative

The late 1950s to early 1960s marked significant advancements in computing power at LLNL. Supercomputers like the Univac LARC began to distinguish themselves from commercial counterparts through enhanced speed and specialized capabilities tailored to national security needs.

Becky Springmeier (08:58):
"...the term supercomputer began to emerge. Supercomputers really was its own line and was very specialized for the kind of mission we do here."

In 1995, the Accelerated Strategic Computing Initiative (ASCII) was launched, uniting three national laboratories to develop high-confidence simulation capabilities essential for maintaining the nuclear stockpile without physical testing.

Rob Neely (13:50):
"The primary goal of ASCII is to create the high confidence simulation capabilities needed to integrate fundamental science experiments and archival data into the stewardship of the actual weapons in the stockpile."

Michael McCoy outlines the three critical advancements necessary for ASCII's success: technological progress in computing hardware, the evolution of simulation codes, and the development of robust infrastructure.

Michael McCoy (14:48):
"There are three things that have to happen... our codes... and the infrastructure. If you don't advance all three of those together simultaneously, and in a coordinated way across all of these disciplines, you are a dead duck."

Milestones: From Blue Pacific to Sequoia

The journey of supercomputing at LLNL is marked by significant milestones:

1. Blue Pacific (1998):
   Installed in the fall of 1998, Blue Pacific achieved a 3.9 teraflop performance and enabled the first three-dimensional simulation of an exploding nuclear weapon just a year later.

   Becky Springmeier (20:59):
   "We are now breaking the speed barrier when it comes to computing power and computing time."

2. ASCII White:
   Despite initial setbacks due to the system's complexity and reliability issues, ASCII White eventually secured the number one position for the fastest supercomputer in the world in November 2000, boasting a 12 teraflop performance.

   Michael McCoy (21:52):
   "They're so complex... it was going to be a monumentally complex and difficult thing to address this problem."

3. BlueGene and Sequoia:
   LLNL developed the BlueGene series, with Sequoia achieving a 16.32 petaflop performance in June 2012, making it the world's fastest and most energy-efficient supercomputer at the time.

   Michael McCoy (24:44):
   "We blew the Earth simulator out of the water with this 360 teraflop computer... suggested that there were other approaches consuming far less power that should be followed."

The Push Towards Exascale Computing

As LLNL advances, the focus shifts to exascale computing—the next monumental step, targeting computational speeds of 10^18 operations per second. The lab's upcoming supercomputer, El Capitan, is poised to achieve this milestone by late 2024.

Becky Springmeier (32:32):
"Exascale is the next grand step in the supercomputing evolution... that's what's cutting edge, that's what's complicated, and that's what some of the things we do best here at Livermore."

The transition to exascale involves overcoming significant challenges, including power consumption and the complexity of coordinating millions of operations simultaneously. Livermore's approach involves massively parallel computing, where thousands of processors work in unison to handle intricate simulations.

Becky Springmeier (31:24):
"Imagine you've got a really powerful gaming system... and replicate it 10,000 times and put that all in a very dense packaging on a very high-speed network..."

Philosophy and Persistence in Supercomputing

Central to LLNL's success in supercomputing is a philosophy of relentless persistence and a commitment to big ideas despite technical, political, and fiscal challenges. Michael McCoy emphasizes that scientific discovery is an ongoing journey, with each computational leap unlocking new possibilities for national security, medical research, and more.

Michael McCoy (34:04):
"Scientific discovery is not about the destination. It's about the journey... There's no done. We're never done... Science is like a human life. It's a journey. You're only done when you're dead."

Edward Teller echoes this sentiment, highlighting the continuous evolution of computing power and its integral role in advancing scientific understanding.

Edward Teller (26:15):
"When we were trying to get to terascale computing, that was a big deal... they get more and more capable."

Conclusion: An Endless Pursuit of Excellence

The episode concludes with a reflection on the unending nature of scientific and technological advancement. Despite surpassing previous records, LLNL remains committed to pushing the boundaries of what's possible in supercomputing. The lab's dedication ensures that it remains at the forefront of innovation, continually contributing to national security and broader scientific endeavors.

Michael McCoy (35:23):
"The journey towards unparalleled computational heights is far from over... the team at Livermore are hard at work in the pursuit of the next big achievement."

Rob Neely (35:23):
"The journey towards unparalleled computational heights is far from over... Lawrence Livermore National Laboratory invites you to join our diverse team..."

Listeners are encouraged to engage with LLNL's mission and explore career opportunities, underscoring the lab's role as a beacon of innovation and scientific exploration.

Notable Quotes:

• Michael McCoy (00:03):
  "It is the fast, reliable and tireless performance of a variety of arithmetic and logical operations that gives the computer its great utility and power."

• Michael McCoy (02:48):
  "If you're like us, with a mission which is protecting the security of this country... you are a backpacker walking on a ridge with two chasms."

• Michael McCoy (10:24):
  "We have learned to make war by unlocking the atom. To make peace, we must limit our use of that power."

• Michael McCoy (18:37):
  "I don't think anybody argued that we could do something different at 100 teraflops. That wasn't the hard part. The hard part was, how did we get to 100 teraflops?"

• Michael McCoy (34:04):
  "Scientific discovery is not about the destination. It's about the journey... Science is like a human life. It's a journey. You're only done when you're dead."

Final Note:
This episode of Big Ideas Lab provides a comprehensive look into the evolution of supercomputing at Lawrence Livermore National Laboratory. From the early days of the Univac LARC to the cutting-edge exascale computing with El Capitan, the narrative showcases the lab's pivotal role in advancing computational technology for national security and scientific discovery. Through expert insights and historical anecdotes, listeners gain a profound understanding of the challenges and triumphs in the realm of supercomputing.