Online GPU Platforms

Updated: July 31, 2023

Online GPU platforms are cloud-based services that provide access to powerful Graphics Processing Units (GPUs) over the internet. These platforms are designed to enable developers, researchers, and data scientists to perform computationally intensive tasks, such as deep learning, artificial intelligence, and data analysis, without the need to invest in expensive hardware. Users can rent virtual GPUs on a pay-as-you-go basis, allowing them to scale their computing resources based on their specific needs. Online GPU platforms offer convenience and flexibility, as users can access high-performance computing resources from any device with an internet connection. Additionally, these platforms often provide pre-installed frameworks and libraries, simplifying the setup process and enabling users to focus on their work rather than managing hardware and software configurations. Online GPU platforms have democratized access to powerful computing resources, enabling a wider range of users to harness the capabilities of GPUs for their compute-intensive tasks.

See also: Top 10 Public Cloud Platforms

2017. Google Cloud Platform cuts the price of GPUs by up to 36 percent



Google is implementing a price reduction for Nvidia's Tesla GPUs on its Compute Engine service, with savings of up to 36 percent. In U.S. regions, the cost of utilizing the slightly older K80 GPUs has been reduced to $0.45 per hour, while the more advanced and powerful P100 machines will now cost $1.46 per hour, both with per-second billing. By doing so, Google aims to attract developers who seek to execute their own machine learning workloads on its cloud platform. Additionally, various other applications, such as physical simulations and molecular modeling, can greatly benefit from the extensive number of cores provided by these GPUs.