The boundary between the supercomputing center and the cloud is becoming more and more blurred, and more and more cases show that the efficiency of cloud computing in all aspects even exceeds that of the supercomputing center. Take the cooperation between cloud computing service provider Cloudam Cloud and a domestic medical university as an example to specifically analyze the difference between supercomputing and cloud computing.
For users of high-performance computing, the supercomputing center is usually the first choice, but with the continuous improvement of cloud computing power service providers, it even exceeds the supercomputing center in efficiency. Cloud computing power service providers similar to Cloudam cloud use can provide The one-stop computing power service has installed some commercial software for high-performance computing users. More importantly, the price is almost the same as that of the supercomputing center, which is very friendly to users with small-scale computing resource needs.
For example, Cloudam Cloud once cooperated with a TOP2 university, and the school's medical school drug discovery project also encountered a computing resource bottleneck problem. The project needs to use multiple targets for full compound screening. However, the university’s current HPC platform can support only 2000 cores, which is obviously not enough. In addition, the project also faced the problem of tight time and limited computing resources.
If the school chooses a supercomputing center for the project, it will not only need to queue up, but also cannot integrate a large amount of heterogeneous data, which means that the efficiency of research and development will be reduced, and more importantly, the supercomputing center will not provide it with deployment solutions. , Which means that the school needs additional operation and maintenance personnel.
By comparison, the school finally chose the cloud E computing power platform for unified resource management. Through the cloud E platform, users can directly use the browser to call the pre-deployed Virtualflow screening tool on the platform to achieve multi-stage screening of targets. During the operation, the user does not need to worry about the hardware configuration, and can remotely operate and filter using a clear and easy-to-understand graphical interface. In order to ensure the safety of users, Cloud E has signed data security and confidentiality agreements with customers, and the cloud strictly guarantees the security and privacy of user input data and final screening results.
Through the Cloud E one-stop deployment solution of CLOUDAM cloud, high concurrency and high availability work was realized, which greatly shortened the original project process, and finally can be completed in one month. And after accessing CLOUDAM cloud cloud E computing power platform, the overall operation and maintenance cost is reduced to 1/5 of the original.
Cloudam Cloud Cloud E provides a one-stop solution for high-performance computing for the biotechnology industry, integrating heterogeneous resources, allowing various software and frameworks to run on the CPU and GPU nodes of Cloud E with high performance, realizing massive CPU cores Data virtual screening service, top-level GPU model training tasks. The automated workflow also makes the entire use process simple and clear. Users only need to focus on professional content, and no longer have to worry about the configuration and deployment of hardware devices.
Through this case, it is not difficult to find the advantages of cloud computing power service providers. Although supercomputing centers are a good choice for scientific researchers at this stage, as cloud computing power service providers gradually become market-oriented, cloud computing power The computing platform is very promising.