Thanks to languages ​​like CUDA and OpenCL, they have gradually evolved to handle tasks that were previously processed by the CPU. Thus, these new servers can be perfectly used in areas such as image processing, bioinformatics, and Big Data, which consume significant server resources.
The OVH team has created three new servers to address this problem. The first, based on an NVIDIA GeForce GTX 970, allows for both accelerated graphics applications and fast calculations. For example, this server can process and encode videos using the card's 1,664 cores, perform 3D rendering, or develop and test CUDA or OpenCL applications.
For professionals in sectors such as design or engineering, the server based on the NVIDIA Tesla 60 graphics card allows up to 32 users to share the machine's graphics resources simultaneously. This graphics card enables users to take advantage of the flexibility offered by virtual desktop environments (VDI) while also providing powerful graphics capabilities for 3D applications (DirectX or OpenGL). The server is designed, for example, to create high-performance VDI infrastructures, remote workspaces for 3D applications, or even host graphically demanding online games.
The third server, the Tesla K80, was built around the card of the same name, which is currently the world's fastest GPU accelerator. Developed by NVIDIA for high-performance computing and machine learning, this card boasts a dual-GPU design and 4,992 CUDA cores, making it five to ten times faster than traditional CPU configurations. Furthermore, the server is compatible with VMware, Microsoft HyperView, and Citrix technologies.
OVH offers the option to add up to 256 IPs at no monthly cost (excluding installation fees). Each server enjoys a guaranteed 500 Mbps bandwidth and is protected by OVH's anti-DDoS technology. The M60 and K80 models are also connected to the 10 Gbps vRack.

More information or a quote