New GPU configurations
We are pleased to inform that two new configurations containing GPUs have been introduced to our standard offer and new dedicated family of GPU virtual machines has been created. New configurations are the answer to a growing demand for GPU processing.
- ds.3xlarge.4gpu is a VM on a Dedicated Server (DS.VM) equipped with four GeForce 2080 TI graphics cards, 496 GB RAM, 48 Vcore's and two, fast NVMe disks with a capacity of 1920 GB and low access times granting high performance.
- gpu.medium is a Virtual Machine (VM) equipped with one GeForce 2080 Ti graphics card, 118 GB RAM and fast SSD network storage
Due to high interest, in the first days of sales these configurations are available only in orders processed through the sales department: firstname.lastname@example.org. We can also create custom configurations containing 1, 2, 4, 8 or 10 graphics cards in one DS server.
GPU SERVERS AVAILABLE ON THE PLATFORM
Graphics Processing Units (GPU) are recently the biggest source of increse in computing power. They were developed as a specialized tool for processing of images mainly in gaming. But currently they are applied in a growing number of CPU heavy applications: not only related to image processing but also to artificial intelligence (AI), deep learning and others.
In the Earth Observation domain (EO), graphics processors are used, among others for radar interferometry. Currently, software development is focused on gaining more and more benefits from the use of calculation accelerators. We install graphic processors in Virtual Machines on a Dedicated Server (DS). These are virtual cloud machines that fill a full physical server that cannot contain any other virtual machines. These physical machines are additionally equipped with one or many GPU cards to which the user gets access through the "passsthrough" mechanism. Thanks to this, the hypervisor overhead for computational operations is virtually eliminated. In addition, these machines are equipped with very fast NVMe storage enabling appropriate supply of the system with data for calculations.