Google recently announced price cut of up to 36 percent for NVIDIA Tesla GPUs that are offered via its on-demand Google Compute Engine virtual machines. With this, in US regions, the cost for using K80 GPU’s is $0.45 per hour while for P100 machines, it will be $1.46 per hour.
Google Compute Engine offers powerful GPUs that can be added to users’ VM instances. These GPUs can be used to accelerate specific workloads on VM instances like machine learning and data processing.
Scientists, engineers and artists require access to large parallel computational power. Hence, Google offers VMs with GPUs that have the capacity to boost performance, and by connecting them to NVIDIA Tesla K80 and P100s GPUs, Google has managed to bring maximum flexibility to the developers. A number of applications like molecular modelling and physical stimulations will benefit from the hundreds of cores available on the GPUs.
Google, also as a bonus, announced lowering the prices of preemptible local SSDs by nearly 40 percent as compared to on-demand Local SSDs.
Google said in its blog, “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.”
Google has targeted the price cuts at developers, who can now run their own machine learning workloads in the cloud without worrying about the cost constraints.
The announcement has come a few days before AWS’s annual conference that will kick off on 27th November in Las Vegas, where AWS is expected to make a number of revelations related to AI and Machine Learning.
It will be interesting to see whether AWS also announces some more price cuts or not. Stay tuned for the latest updates.