Does size matter? University in Barcelona gains 11.1 petaFLOP supercomputer, as Intel touts its HPC credentials
Chinese PC maker Lenovo has implemented what it is calling ‘the world’s largest next-generation Intel based supercomputer.’
The 11.1 petaFLOP supercomputer dubbed ‘MareNostrum 4’ , will be located at the Chapel Torre Girona at the Polytechnic University of Catalonia, in Barcelona.
It is certainly not the most powerful supercomputer in the world (it is ranked 13th in the top 500 supercomputer list), but it will allow scientists to conduct research into the human genome, as well as biomechanics and weather forecasting.
Lenovo said that the MareNostrum 4 is powered by more than 3,400 nodes of Lenovo’s next-generation servers.
These servers contain Intel Xeon processors, and the system is apparently interconnected with more than 60 kilometres (37 miles) of Intel Omni-Path Technology 100 Gb/s network cabling.
Lenovo said this latest supercomputer is the third HPC solution that it has installed at the Partnership for Advanced Computing in Europe (PRACE). This, claimed Lenovo, solidifies its position as the fastest growing supercomputer company in Europe.
“The fast delivery, installation and optimization of the MareNostrum 4 system at BSC, showcases Lenovo’s end-to-end, high-performance computing strength,” said Kirk Skaugen, President of Lenovo Data Center Group.
“Building on our 25 years of history in x86 server computing and our number one position in x86 server customer satisfaction and reliability, our goal at Lenovo is to be the largest supercomputing company on earth helping solve humanities biggest challenges through the rapid advancement of technology and innovation,” Skaugen added.
Indeed, Lenovo backed up this claim after it revealed that it has 92 entries in the TOP500 supercomputer list.
China for example currently has the world’s fast supercomputer with its Tianhe-2, which delivers 33.86 petaflops.
That said, the United States is hitting back.
Last week the US government handed $258m over to a number of US tech firms to develop an exascale supercomputer.
America’s fastest supercomputer, the Cray Titan at the Oak Ridge National Laboratory, is thought to be only capable of 15 petaflops, and is ranked in third position behind two systems in China.
The US however has five of the 10 fastest computers in the world, and it does retain a global leadership in the actual application of high performance computing to national security, industry, and science.
HPC At Intel
Intel meanwhile used the 2017 International Supercomputing Conference to tout its high-performance computing and AI capabilities.
The chip maker said that its upcoming Intel Xeon Scalable processors will offer “significant per-core performance enhancements and other advancements”.
And when these chips are integrated with the Intel Advanced Vector Extensions 512, the HPC platform will be able to generate 2X FLOPs/clock-cycle peak improvements.
“Intel AVX-512 combined with improvements in cores, cache and memory, delivers up to 2.27x more performance than today’s Intel Xeon processor E5 v4 (formerly codenamed Broadwell), and up to 8.2x more double precision GFLOPS/second when compared to a 4-year old Intel Xeon processor E5 family in the installed base,” said the firm.
The company also said it has made significant advancements in the Top500 list of supercomputers with 110 new systems including three systems based on the upcoming Intel Xeon Scalable processors.