NVIDIA L4 GPUs. 5-second cold starts. Scale to zero when idle.
Deploy your model, get an endpoint, pay only for compute time. No GPU provisioning or infrastructure management required.
Try Free
$300 Free Credits to Build on Google Cloud
New to Google Cloud? Get $300 in credits to explore Compute Engine, BigQuery, Cloud Run, Gemini Enterprise Agent Platform, and more.
Start your next project with $300 in free Google Cloud credit. Spin up VMs, run containers, query petabytes in BigQuery, or build agents with Gemini Enterprise Agent Platform. Once your credits are used, keep building with 20+ always-free tier products including Compute Engine, Cloud Storage, GKE, and Cloud Run functions. No commitment required—just sign up and start building.
...The performance of a typical bicubic upscale operation is typically in the order of 10 times faster than via the standard AWT mechanism using RenderingHints.VALUE_INTERPOLATION_BICUBIC.
The library depends on Aparapi (https://code.google.com/p/aparapi/). Any GPU supporting OpenCL should be supported (including low-end GPUs without double-precision fp support), and there is automatic fallback to pure java implementations for all operations in the event that there is no OpenCL support available.
A high-level API allows for very easy use from AWT/Swing, whilst the low-level array-based API allows integration with any imaging framework (e.g. from Android or SWT).