Geek Pie Does Well in International ISC Competition

ON2018-07-10TAG: ShanghaiTech UniversityCATEGORY: Global

ISC-2018 Student Cluster Competition(ISC-SCC) announced the final results on June 27th local time at Frankfurt am Main, Germany. ShanghaiTech University team GeekPie_HPC won the Fan Favorite Team award. In addition, GeekPie_HPC team achieved the highest scores for HPCG Benchmark and AI application. Nearly three hundred teams from all over the world took part in the ISC-SCC this year, and 12 teams coming from nine different countries qualified for the ISC-SCC final. This GeekPie_HPC’s first time to participate at ISC. Made up five students ( Yuan Wei Jun, Liu Jian Zhong, Shen Ji Qi, Luo Huo Cong, Xie Zhi Qiang) the team was invited to the ISC-SCC final directly as the silver medal winner of ASC-2018.

The team from Tsinghua University won first place in the competition, Singapore Nanyang Technology University Team ranked second, and South Africa Center of High-Performance Computing Team took third.

ISC, SC, and ASC host the three largest high-performance computing challenges for students in the world. Teams have to build and test their HPC clusters under the 3KW budget and pursue higher scores of six applications that come from three major categories -- HPC benchmarks, HPC applications, and AI applications.

The HPC benchmarks are comprising of three commonly used benchmarks: HPL (the High-Performance Linpack Benchmark), HPCC (the High-Performance Computing Challenge Suite), and HPCG (the High-Performance Conjugate Gradients Benchmark).

The HPC Applications are GRID, which is a HPC for Lattice Quantum Chromodynamics (Lattice QCD) calculations, Nektar++, a spectral/HP element framework designed to support the construction of efficient high-performance scalable solvers for a wide range of partial differential equations, and a mystery application that is released after the submission of HPC benchmarks results. This year, the mystery application was Nek5000, a popular simulation tool for computational fluid dynamics.

The Deep Learning AI task was an image recognition model training, using TensorFlow.

GeekPie_HSC members: from left to right Yuan Wei Jun, Liu Jian Zhong, Shen Ji Qi, Luo Huo Cong, Xie Zhi Qiang