Nvidia has announced a line of “Nvidia-Certified Systems. This line consists of a series of systems for companies that want to get started with artificial intelligence at different scales.
To build the systems, Nvidia has partnered with Dell, HP Enterprise, Gigabyte, Insput and Super Micro. Nvidia’s guidelines focus on providing optimal performance for tasks such as machine learning and data analysis.
The certified systems are available for both the HGX and EGX server platforms, for hyperscale computing and edge computing respectively. They can be equipped with Nvidia’s A100 GPU, which excels in AI applications. The A40-GPU, which is more suitable for graphics tasks, is also available. Furthermore, the systems are equipped with Nvidias fast Bluefield or Mellanox network adapters.
Rapidly growing popularity
Nvidia is clearly trying to capitalise on the fast-growing popularity of artificial intelligence. Last year, the company introduced the A100 GPU, specially developed for machine learning. In October, Nvidia proudly published a press release that the A100 had set a new record in the MLPerf benchmark. Since then, the company has announced several devices featuring the A100 GPU, including the compact DGX Station A100, which featured four GPUs.
Adel El Hallak, who is in charge of GPU Cloud at Nvidia, says that AI has become mainstream. Customers are looking for guaranteed functionality, performance, scalability and security. “Doing AI at scale is difficult and up until now has been a do-it-yourself program,” he says. With Nvidia-Certified Systems, the company wants to offer an easily deployable solution.
In total, Nvidia has announced 14 systems. These are:
- Dell EMC PowerEdge R7525 and R740
- Gigabyte R281-G30, R282-Z96, G242-Z11, G482-Z54 and G492-Z51
- HPE Apollo 6500 Gen10 System and ProLiant DL380 Gen10 Server
- Inspur NF5488A5
- Supermicro A+ Server AS -4124GS-TNR and AS -2124GQ-NART