The Gene Institute computing platform aims to achieve large-scale agricultural bioinformatics computing as its central task, promote the formation of a research system for storing, integrating, and mining massive agricultural bioinformatics big data, and lay the foundation for deep mining of agricultural bioinformatics big data. It promotes discipline construction with two centers: supercomputing and big data. Currently, the platform has a computing capacity of 500TFLops and a storage capacity of 16PB, mainly CPU computing, and is deployed with Intel and domestic ARM architectures, The total number of cores exceeds 15000. The platform is preparing to build a new phase of supercomputing cluster. By 2024, the platform's CPU computing capacity will reach 1500TFlops and storage capacity will be 40PB.
| Construction time | CPU computing power | GPU computing power | Storage capacity | Funding source |
|---|---|---|---|---|
| May 2014 | 35T | / | 1P | Self raised funds |
| July 2016 | 15T | / | 1.5P | Shenzhen Key Laboratory of Agricultural Phenotype and Utilization |
| April 2019 | 10T | / | 3P | Self raised funds |
| May 2020 | 90T | / | 0.5P | Self raised funds |
| September 2021 | 350T | 150T | 10P | Shenzhen Institute of Basic Research |
(1)The platform's computing resources are configured in a ladder configuration, and large memory node partitions such as 6T, 3T, and 2T can meet the needs of complex genome assembly; Ordinary node partitioning can meet many applications such as gene data and phenotype data analysis.
(2) The platform provides AI computing power and the most mainstream AI deep learning algorithm frameworks such as TensorFlow, Caffe, and Pytorch, which can respond to many complex intelligence challenges.
(3) The platform has more than 40 domestic CPU chip (Huawei Kunpeng ARM architecture) servers, and more than 50 commonly used bioinformat software have been migrated and deployed.
(4) The platform has Sentieon and gtx gene data analysis software resources, which can significantly accelerate the analysis process such as BWA GATK
(5) The platform has access to external computing resources with high cost-performance AMD, domestic HaiGuang, and other CPU architectures as a supplement, capable of meeting concurrent of tens of thousands of cores.
The computing platform implements the principle of open sharing, independent accounting, and establishes a cost accounting and service fee management mechanism to provide cluster construction and support, data analysis, bioinformatics database development, high-performance computing, computing acceleration, and massive storage services for biological research in and outside the institute.