There are quite a few industries whether it is deep learning and AI or for artists and VR developers where top-tier talent gets awesome perks, such as having the best workstations. For those who have jobs that depend on it, the NVIDIA Quadro GV100 is the GPU to get. It has a full 32GB of HBM2 memory, like the latest NVIDIA Tesla V100 32GB edition. It is based on the newest NVIDIA Volta architecture, and it is designed to operate in workstations.
NVIDIA Quadro GV100 Key Specs
The NVIDIA Quadro GV100 is based on Volta, an architecture we already know quite a bit about. Differentiating the GV100 from lesser GPUs, like the NVIDIA Titan V, are a few key features.
First, the NVIDIA Quadro GV100 has 32GB of HBM2 memory onboard with a 4096-bit interface. That is a significant jump from the previous top-end Volta desktop card, the Titan V which had 12GB of memory and a 3072-bit memory interface. Like other Volta architectures (save the “baby Volta” 150W cards) it has 5120 CUDA cores and 640 Tensor cores. Those tensor cores are a major hardware accelerator for NVIDIA’s AI efforts.
Beyond this, the new GV100 supports NVLink with 200GB/s of bandwidth between two GPUs.
Since it is meant for the workstation market, these are actively cooled GPUs with 250W maximum power consumption and four DisplayPort 1.4 outputs for driving multiple high-resolution monitors.
The new Quadro GV100 with NVIDIA RTX Technology is destined to find its way into systems from HP, Dell and Lenovo systems for those working on ray tracing graphics.
NVIDIA Quadro GV100 Key Specs
Here are the key specs of the new NVIDIA Quadro GV100 GPUs:
Thes are certainly high-end GPUs that are similar to the Tesla V100 GPUs that have been refreshed for GTC 2018.
Final Words
Capabilities and hardware like this do not come inexpensively. We expect these GPUs to have a hefty price tag associated with them. At the same time, when you are spending large dollar amounts on top-tier deep learning and AI researchers, getting them the hardware they need to succeed can be worth the few thousands of extra dollars in cost.
