Primate Labs, the company behind the widely-used Geekbench benchmark, has announced the release of Geekbench AI 1.0. This new benchmark suite is specifically designed to evaluate the performance of machine learning, deep learning, and AI-centric workloads across different platforms. The release follows years of dedicated development and extensive feedback from customers, partners, and the AI engineering community.
Geekbench AI 1.0 introduces a novel testing methodology, mirroring real-world AI applications and ensuring cross-platform compatibility. This approach is a testament to Primate Labs’ commitment to building benchmarks that accurately reflect the needs of developers and users.
Primate Labs acknowledges the complexities of measuring AI performance, stating on its website that “it’s hard to determine which tests are the most important for the performance you want to measure, especially across different platforms, and particularly when everyone is doing things in subtly different ways.” The company goes on to explain that they have developed their tests in close collaboration with software and hardware engineers across the industry, ensuring that their benchmarks represent the real-world use cases that AI applications are designed for.
Drawing parallels to the challenges of evaluating graphics cards in the past, Primate Labs highlights the complexities of the AI landscape. Just as GPUs in the 90s and 2000s offered diverse features and support for various software frameworks, AI workloads today present a complex web of hardware and software considerations. With a multitude of AI frameworks, hardware accelerators, and specialized libraries, selecting the right platform for AI applications can be a daunting task. This is where Geekbench AI comes in, offering a standardized approach to measuring and comparing AI performance.
You can download Geekbench AI 1.0 now and start benchmarking your AI systems. This release is a significant step forward in the field of AI benchmarking, providing developers and users with a reliable and relevant tool to evaluate the performance of their AI applications and hardware.
