Cloud price wars continue, this time it’s with a focus on GPU pricing.
Everyone likes saving a bit of money and Google tends to enjoy giving businesses a bit of a price cut now and then.
The advertising giant’s cloud computing arm, Google Cloud is cutting the price of using its Tesla GPUs from Nvidia for customers using the Google Compute Engine.
The prices for the GPUs attached to on-demand Google Compute Engine virtual machines is to drop by up to 36% for users in US regions, with each K80 GPU priced at $0.45 per hour while each P100 costs $1.46 per hour.
The company said that the lower priced GPUs, along with the Custom VM shapes and Sustained Usage Discounts, provide up to an “addition 30% off of instance pricing,” meaning that users can run highly parallelised compute tasks on GPUs “at a great price,” according to Google.
Google, which is clearly getting into the festive spirit nice and early, is also reducing the price of pre-emptible Local SSDs by almost 40% compared to on-demand Local SSDs.
With the company already citing users such as scientists, artists, and engineers, and use cases such as deep learning, physical simulation, and molecular modelling, it hopes that lower pricing will open up more areas for use.
The company said: “We hope that the price reduction on NVIDIA Tesla GPUs and pre-emptible Local SSDs unlocks new opportunities and helps you solve more interesting business, engineering and scientific problems.”
With Amazon Web Services about to hold its annual conference in Las Vegas, AWS re:Invent, it’ll be worth watching its rivals to see just how much of the cloud giant’s thunder they attempt to steal.
Last year the leader in the public cloud market held nothing back, releasing dozens of new products and services and frankly making the rest of the public cloud players look a little bit slow and cumbersome.