Genetic research institute introduces Dell fusion architecture to interpret the mysteries of life

Relying on the advanced sequencing and testing technologies, efficient information analysis capabilities, abundant biological resources, and multi-disciplinary integrated biological research system of BGI, this company provides innovative biological research services for scientific researchers and application services for the public in the fields of biotechnology such as medicine, agriculture, and environment. To further enhance the computing capacity of the HPC platform, BGI introduced the Dell PowerEdge FX integrated architecture solution to upgrade and expand the biological science computing HPC platform. The total peak computing capacity of the platform was increased to 288.5 Tflops, and the TCO of the HPC platform was reduced by 20%.

Customer background requirements

 With the development of gene sequencing technology and the improvement of sequencing capabilities, the cost of sequencing is significantly decreasing. On the basis of improving computational accuracy, BGI strives to make the price of biological sequencing even lower. And continuously enhancing the computing power of the HPC (High Performance Computing) platform is the inevitable path to achieve this strategic goal.

Scheme value

The Dell EMC PowerEdge FX integrated architecture solution was introduced, and the total peak performance of HPC reached 288.5 Tflops.

To further enhance the computing capabilities of the HPC platform, BGI introduced the Dell PowerEdge FX integrated architecture solution and upgraded and expanded the original platform. After the project was completed, the total peak computing capacity of BGI’s HPC platform reached 288.5 Tflops, the total memory capacity reached 67.22 TB, and the total storage capacity reached 35.09 PB. Overall, the level of the platform was far ahead of that of other companies in the industry.

The efficient integration architecture enables the continuous improvement of scientific research computing capabilities.

BGI adopted a converged architecture solution combining Dell EMC PowerEdge FX and PowerEdge FC630. Four FC630 physical nodes were integrated in the 2U FX2 chassis. Each PowerEdge FC630 was equipped with two Intel® Xeon TME5-2670 v3 processors and 160G DDR4 memory. In the future, it can be expanded to a memory capacity of 768GB. The storage part includes two 2.5-inch hard disk drives or up to eight 1.8-inch solid-state drives. Due to the higher-end disk speed requirements, some FC630 servers were equipped with PowerEdge Express Flash NVMePCIe solid-state drives this time. In addition, BGI adopted the Intel Lustre system, which has outstanding performance, supports large-capacity storage, and has mature and high-reliability technology.

The efficient integration architecture has enabled the scientific research computing capabilities of BGI to continuously improve. The efficient research computing capabilities have also facilitated the realization of the fully self-owned intellectual property, internationally advanced-level gene sequencing system – BGISEQ-500, which enables “one-click sequencing”. More specifically, it can complete the entire process from DNA samples to data analysis results within 24 hours, with the accuracy of personal genome testing reaching an extremely high level of 99.99%.

Efficient management enables easy expansion, reducing TCO by 20% and promoting sustainable development in biological research.

In terms of scalability, the PowerEdge FX2 not only can accommodate redundant four-port direct-through 1 GbE or 10 GbE I/O modules, but also comes with two FN I/O Aggregator network aggregation nodes that offer more network functions. Moreover, its unique modular design provides Huada Genomics with flexible combinations of servers, storage, and networks. The extraordinary flexible configuration and expansion capability of the FX offer Huada Genomics a continuous development ability.

This has significantly enhanced the high-performance computing capabilities of BGI, greatly simplified the complexity of HPC management, strengthened the management ability, improved operational efficiency, and reduced costs by a large margin. Compared to rack-mounted servers of the same scale, the overall IT total cost of ownership (TCO) has been reduced by nearly 20%, the usage space has been saved by 30%, energy consumption has been reduced by approximately 40%, and network latency has reached the millisecond level. The HPC system of BGI has achieved efficient and ecological development.

Furthermore, with the continuous improvement of BGI’s biological scientific computing capabilities, its research work has achieved a leading position. The average citation rate of its papers has reached 61 times per article, which is 8.37 times higher than the national average of 6.51 times per article and 5.22 times higher than the global average of 9.8 times per article.

客户评价

The Dell EMC converged architecture solution utilizes the internal network, effectively avoiding the complexity of using external switching equipment and additional wiring. The onboard LAN network on the PowerEdge FX2 motherboard can provide four connections for each half-width computing node. Meanwhile, Dell’s FN IO Aggregator has achieved a 10GB high-speed channel and multiple external 10GB connections between all nodes.

——Yang Liping, the HPC platform architect of BGI (Biomolecular Genomics Institute)

“Dell EMC has provided us with a Java-based terminal simulation application management program. Through this program, we can quickly understand the status of the servers managed by CMC on each FX2 node, including CPU resource usage, CPU temperature, memory consumption, I/O speed, system average power, etc. At the same time, for the iDRAC GUI of the startup module and the remote application, we have obtained a unified operation interface, which effectively improves our management efficiency.”

——Li Yan, Senior Engineer of the Bio-computing HPC Platform at BGI