For some time now, processors have been offering cores specifically focused on AI tasks. They are called NPUs and their capacity is measured in TOPS.
We see, in principle, that all processors on the market offer the same power in TOPS. At first, we might tend to think that in this task everyone will offer more or less the same performance. But this is not the case and to make a proper comparison, the Geekbench AI tool was launched.
Measuring a CPU’s AI performance is easy with Geekbench AI
First of all, you should know that this tool is not something new. Primate Labs has been offering Geekbench ML software for some time to measure or evaluate the AI capabilities of processors. In reality, this is just a small name change, which makes it more user-friendly.
The term ML (Machine Learning), although more accurate, is not user-friendly for the end user. That’s why they changed it to AI (Artificial Intelligence). Come on, pure marketing.
It is important to know that AI algorithms can work very differently on different platforms. Since this software is cross-platform, users may see different scores depending on the system.
To provide accurate measurements and relevant information, the test results are divided into three groups: CPU, GPU, and NPU. And, in reality, NPUs are a completely unnecessary addition. The GPU, especially the one integrated into processors, already offers good capabilities for AI, due to its design focused on parallel computing.
In short, what NPUs offer is high inference speed with low power consumption. Moreover, these cores are an essential requirement to use Microsoft’s Copilot+ that was recently included in Windows 11.
What is quite interesting explained by Primate Labs is that AI benchmarks are more complex than GPU tests. It is not only the raw processing power that needs to be measured. This tool displays three parameters, such as:
- Total precision
- Average accuracy
- Quantified scores
The idea is to provide a better understanding to the user so that they can make better choices based on the metrics. It would be something similar to what happens with processors, which show performance in a single core and in all cores.
This tool will undoubtedly evolve as the different models are updated and become more complex. Currently, AI algorithms are constantly evolving and are still in fairly early stages of development.
If you’re interested, you can check out the official Geekbench AI ranking, but for now, only performance data, smartphones and tablets (mostly Apple) appear. We imagine that now that the tool is a little more known, data from desktop processors will appear.
