UL Solutions is launching the Procyon AI Image Generation Benchmark this week, a new benchmark that the company says was designed to provide a consistent, accurate, and understandable workload for measuring the AI performance of high-end hardware, including powerful on-device AI accelerators, such as high-end discrete GPUs. This benchmark can be obtained through a site license that starts at $5,000 per year, according to its official website.
Features:
- Heavy test built around an image generation workload, using state-of-the-art neural networks.
- Designed to measure inference performance of powerful AI accelerators, such as discrete GPUs.
- Verify inference engine implementation and compatibility.
- Simple to set up and use via the UL Procyon application or via command-line.
- Test with multiple versions of the Stable Diffusion AI model.
- Compare up to 4 results side-by-side in the app.
Supported inference engines:
- Intel OpenVINO
- NVIDIA TensorRT
- ONNX runtime with DirectML
Site license includes:
- Annual site license for UL Procyon AI Image Generation Benchmark.
- Unlimited number of users.
- Unlimited number of devices.
- Priority support via email and telephone.
Inspect generated images/get detailed scores for each image generation batch:
UL on its new benchmark:
The benchmark includes two tests for measuring the performance of both mid-range and high-end discrete graphics cards. We also hope to add more AI image generation tests in the future to support other performance categories. The Stable Diffusion XL (FP16) test is our most demanding AI inference workload, and only the latest high-end GPUs meet the minimum requirements to run it. For moderately powerful discrete GPUs, we recommend the Stable Diffusion 1.5 (FP16) test.