This high-performance computer broke the record for calculating pi

The high-performance computer performed the calculation of pi with an accuracy of 62,831,853,071,796 digits.

Swiss research team High-performance computers claim to have added a new figure of 12.8 trillion feetWith accounts reaching a total of 62.8 billion numbers.

Based on the Center for Analysis, Visualization and Data Simulation (DAViS) at the University of Applied Sciences in Graubünden, high-performance computers performed Pi calculations with a resolution of 62,831,853,071,796 digits. Timothy Marican Last year it was 50 trillion decimal places.

Before Timothy Mullican, the trophy was the exclusive property of Google. The team discovered more than 31.4 trillion feet in 2018.

The Swiss team achieved this in just over 108 days. This is 3.5 times faster than Timothy Marican, who set the previous record at 303 days and is currently awaiting verification before being included in the Guinness Book of Records. At a later time, the full number will be announced, but the researchers announced that the last 10 digits of pi are now 7817924264.

Le Style Chudnovsky

For most people, pi is described as the ratio of the circumference to the diameter of a circle, often abbreviated as the first number, so that only memory can remember the furthest from math lessons: 3.1415. for several centuries. Mathematicians tried to calculate the number of bezoles as accurately as possible. However, pi is known as the irrational number and cannot be expressed with absolute precision. Therefore, it is not important to find the actual use. Instead, computing has become an unofficial standard for high-performance computing, giving scientists the opportunity to compete against each other.

“We wanted to achieve many goals in this measurement attempt,” says Heiko Rolke, president of DAViS. “By preparing and executing the calculations, we were able to accumulate a lot of knowledge and improve the process. This is a study that is currently collaborating on a large IT project in the areas of data analysis and simulation. Especially useful for our partners.”

DAViS researchers used a well-known algorithm called Chudnovsky’s formula. It was developed in 1988 and is considered the most efficient method for calculating pi. The Google team and Timothy Mullikan also used Chudnovsky’s algorithm. This algorithm was run in 2009 using another popular computer program, y-cruncher, designed specifically by American developer Alexander Lee to calculate pi.

According to the Swiss team, one of the main challenges was the amount of memory required to perform such large calculations. High-performance DAViS computers have two AMD Epyc7542 processors and 1 TB of RAM, which is not enough to store all the required numbers. So I used y-cruncher to move the numbers to an additional 38 hard drives with a total storage space of 16TB, freeing up a lot of RAM on the hard drives.

strong possibility

During operation, the temperature of the computer and disk can reach 80 ° C, therefore, the system is placed in a server rack with constant air cooling to prevent overheating. This contributed more than half of the total 1,700 watts of electricity that scientists estimate is necessary to make a full calculation. However, the system will be 153^Huh. Location Sur Green500.list..

It is unlikely that the additional 12.8 trillion feet will be immediately used for commercialization, and the results reflect the scientific innovation and performance of high-tech computers. For example, Chudanovsky’s formula is known for its complexity. By implementing an algorithm, scientists are finding that if the time and resources required to compute a number increase faster than the number itself, physical obstacles become more difficult to overcome as the number of computations increases.

For Swiss researchers, this new achievement reflects the capabilities of high-performance computing systems and their potential in other areas of research. “The calculations show that we are ready for extensive use of data and computing power in research and development,” said Thomas Keeler, project leader at the University of Applied Sciences in Graubünden. “This account also identified weaknesses in the infrastructure, such as a lack of spare capacity.”

DAViS supports the use of high performance computing in machine learning. For example, a project called Translatoria is building computer tools for translation from Romanian, which is spoken primarily in the now endangered Swiss state of Grisons.

The computing center is also studying the application of DNA sequence analysis in the study of allergies and asthma, which also requires a high-performance computer system. This new record will help pave the way for future practical applications.

sauce: ZDNet.com