UL Solution, the developer of 3DMark, announced the launch of Procyon AI text generation on September 9, local time.benchmarkingThe AI gas pedal hardware's text generation capabilities are judged by synthesizing large-language AI models using a variety of different parameter scales.
The Procyon AI Text Generation Benchmark currently supports local NVIDIA, AMD, and Intel GPUs via the DirectML Common API, as well as their own GPUs via Intel's OpenVINO (Note:Independence and integrationGraphicsall support(but there are differences in video memory / RAM capacity requirements).
The benchmarkingFour models with varying parameter scales are now included(Phi-3.5-mini, Mistral-7B, Llama-3.1-8B, and Llama-2-13B), which allows for a relatively fair comparison of the performance of AI GPUs in diverse large-language model text generation use cases.
UL Solution says the Procyon AI Text Generation Benchmark is easier to use, reduces the size of the installation compared to testing directly with a full AI model, and provides a detailed picture of how system resources are utilized when running text generation loads.