Primate Labs has introduced an innovative Geekbench AI benchmark to evaluate the performance of artificial intelligence on various devices. This cross-platform test allows you to compare the effectiveness of AI algorithms on processors, video cards and neural processors.
Geekbench AI Benchmark
New AI Evaluation Tool
The creators of the popular Geekbench benchmark, Primate Labs, have released the first version of a specialized test for assessing the performance of artificial intelligence - Geekbench AI. The tool builds on previous Geekbench ML development focused on machine learning.
Features of AI testing
Artificial intelligence algorithms demonstrate different effectiveness on different hardware. Geekbench AI, being a cross-platform test, allows you to get a variety of results. In this regard, the rating table is divided into three categories: processors, graphics accelerators and neural processing units (NPU).
Assessment methodology
Geekbench AI uses proprietary algorithms to measure the speed of the latest processors. This figure can be compared between devices in each group, although it is technically possible to compare even a smartphone with a powerful 450W graphics card.
The Challenges of AI Testing
Primate Labs notes that testing AI is significantly more complex than evaluating GPUs. It's important to measure more than just raw computing power. Over time, GPU manufacturers have introduced new frameworks and API support, which should also be taken into account in the benchmark. Geekbench AI focuses on different subtests, but it's difficult to compare them to gaming tests that take into account scaling and frame generation.
Score Categories
Geekbench AI scores will be presented in three categories: full precision, semi-precision, and quantized scores. This solution aims to provide greater insight into how developers or OEMs interpret tests by providing more metrics, similar to how CPU testing focuses on single- and multi-threaded metrics.
Availability and Application
Using Geekbench AI, users can identify the best performing hardware for AI tasks. The official rating may not update as quickly as new devices are added. The Geekbench AI test is available as a free trial.
Glossary
- Primate Labs - benchmark company, creator of Geekbench
- Geekbench AI - a new benchmark for assessing AI performance
- NPU - neural processor, a specialized chip for AI tasks
- GPU - graphic processor, used, among other things, for AI calculations
- API - application programming interface, a set of tools for software development
Links
Answers to questions
What is Geekbench AI and who developed it?
What device categories does Geekbench AI rank?
How does Geekbench AI rate AI performance?
What types of scores does Geekbench AI provide?
How can I access Geekbench AI?
Hashtags
Save a link to this article
Discussion of the topic – Geekbench AI: A Revolution in Artificial Intelligence Testing
Primate Labs has released Geekbench AI, a universal cross-platform benchmark for testing artificial intelligence performance on various devices.
Latest comments
8 comments
Write a comment
Your email address will not be published. Required fields are checked *
Ethan
It's cool that Geekbench now tests AI too! 🤖 I wonder how my computer will cope with these tests.
Sophie
Yes, Ethan, that's really cool! What's especially interesting is that they divided the results into categories. I think this will help to better understand the performance of different devices in AI tasks. 🧠💻
Diego
I agree with Sophie! It’s also cool that the test is cross-platform. It will be possible to compare the performance of a laptop and a smartphone. 📱💻 Has anyone tried it yet?
Lena
I haven't yet, Diego, but I really want to try! I wonder how my new phone will handle the NPU. 🤔 What do you plan to test on?
Giovanni
All this, of course, is interesting, but why does the average user need these AI tests? Again a marketing ploy to sell new devices. It would be better if they focused on real performance in everyday tasks.
Ethan
Giovanni, I understand your skepticism, but AI is already everywhere - from photo processing to voice assistants. This test can help you choose a device that will better cope with such tasks. 📸🗣️
Anna
I agree with Ethan! Giovanni, AI is the future. I think such tests will help developers create more efficient AI applications for different devices. This benefits everyone! 🚀
Diego
Anna, great point! 👍 By the way, I have already downloaded Geekbench AI and tested my PC. The results are impressive! Lena, be sure to try it on your phone, it will be interesting to compare.