As the latest 7Nm design, the machine learning performance of Cortex-A77 has increased 35 times over that of two years ago.
As for Mali-G77 GPU, based on the new Valhall architecture, the performance improves by 40%, efficiency improves by 30%, performance density improves by 30%, and machine learning performance improves by 60%.
It is worth mentioning that the Mail-G77 GPU can also work with the new Mali-D77 display processor.
Finally, ARM released a new machine learning (ML) processor, which is part of the Project Trillium computing platform and the Neural Network (NN) software framework.
According to ARM, the idea is to add a dedicated neuroprocessing unit (NPU) to the chipset to increase energy efficiency by two times and memory compression technology by three times.
If all goes well, we are expected to see the latest technologies mentioned above on our flagship smart phones (or ARM laptops, etc.) in 2020.