How TeamEPCC Broke the LINPACK Record at ISC’14

Print Friendly, PDF & Email
(Left to Righ) Chenhui Quan, Konstantinos Mouzakitis, Emmanouil Farsarakis (Manos) and Georgios Iniatis.

(Left to Righ) Chenhui Quan, Konstantinos Mouzakitis, Emmanouil Farsarakis (Manos) and Georgios Iniatis.

Over at the EPCC Blog, team mentor Xu Guo writes that four MSc in HPC students achieved the highest-ever LINPACK score in the Student Cluster Competition at ISC’14.

The team benchmark score was 10.14TFLOP/s (or 3.38TFLOP/s per kW), the first time a student team has broken the 10 TFLOP/s barrier within the competition’s 3kW power limit.  If submitted, such a system would rank roughly #4 on the Green500 list.

Our system was designed with the Linpack benchmark as a first priority, as the “Highest Linpack” award is the only one where results can be compared from year to year. We chose to incorporate NVIDIA K40 accelerators in our system, as the data we collected on benchmarks showed that they provide very high flops per watt. We decided that since the majority of computation would be taking place in the GPUs, we would eliminate as much overhead as possible, having an equal amount of CPUs and GPUs in the final configuration.

According to TeamEPCC, their secret weapon was liquid cooling.

By cooling the CPUs and GPUs using the CoolIT technology, we were able to remove multiple fans from the system. In the end, only 4 of the original 20 fans were left on each server. The heat exchange consumed only about 90W in total and this was further reduced by deactivating some of the heat exchange’s fans, after carefully investigating the feasibility of doing so without putting the system at risk.

Read the Full Story.