Nvidia Tesla M10 GPU Puts Four Mid-level Maxwell GPUs on a Single Card
Nvidia has unveiled the Tesla M10 GPU which is designed to power a lot of virtual desktops and support cloud-powered gaming. The GPU offers system administrators the ability to put up to 64 virtual users per board and 128 per server with two boards.
The Tesla M10 is basically a quad mid-level Maxwell on a single card, with each GPU paired with 8GB of GDDR5 RAM for a total of 32GB. Which means, the new graphics card actually adds high-density GPU resources to Nvidia’s existing Tesla product line. Key specifications of Nvidia Tesla M10 are as under:
- GPU: Quad Mid-level Maxwell
- CUDA Cores: 2560 (640 per GPU)
- Memory Size: 32 GB GDDR5 (8 GB per GPU)
- H.264 1080p30 streams: 28
- Max vGPU instances: 64
- Form Factor: PCIe Dual Slot
- Power: 225W
- Thermal: Passive
Aimed at the enterprise world, the M10 coupled with the latest version of GRID can support virtual applications such as Citrix’s XenApp and virtual PCs running Windows, which are performance hungry to carry out professional graphics work.
So this will basically help engineers and designers work remotely without compromising the GPU and compute power they need to run these apps.
The latest Grid software will be available starting today, while Tesla M10 hardware will hit the shelves in August.