Google has reduced the cost of NVIDIA Tesla GPUs running on Google Compute Engine virtual machines, making it more cost-effective for businesses to run high-performance instances on their cloud infrastructure.
The company's GPU virtual machines allow customers to create exact performance and cost VM configuration for their workload, using the right number of vCPUs, GPUs and memory for each application running. According to Google, this means businesses are never charged more than they need to pay for each individual application.
Google also supports the addition of up to 3TB of Local SSD to any GPU-enabled VM if storage is required.
"We've been seeing customers (like Shazam and Schlumberger) harnessing the scale of Google Cloud, and the power of NVIDIA Tesla GPUs to innovate, accelerate and save money," Chris Kleban, product manager at Google Cloud, said. "Today we're extending the benefits of GPUs by cutting the price of NVIDIA Tesla GPUs attached to on-demand Google Compute Engine virtual machines by up to 36%."
The price per K80 GPU attached to a VM is now $0.45 per hour, while each P100 costs $1.46 per hour for US regions. In Belgium (the only European region supporting P10 and K80 models), they will cost slightly more - $1.60 per P100 GPU and $0.49 per K80 GPU.
Google also announced it would be reducing the cost of preemptible local SSDs too, bringing the price down by 40% compared to on-demand local SSDs.
"We hope that the price reduction on NVIDIA Tesla GPUs and preemptible Local SSDs unlocks new opportunities and helps you solve more interesting business, engineering and scientific problems," Kleban added.
