What is amazing for the 'A12X Bionic' processor on the new iPad Pro?

Apple's new iPad Pro , which appeared on November 7, 2018, has evolved a lot, including bezelless full screen and the introduction of a new Apple Pencil . The SoC , the heart of such iPad Pro, is equipped with the 7-nm process " A12X Bionic ", which has become a hot topic as the processing performance as a computer has dramatically improved.

Apple walks Ars through the iPad Pro's A 12 X system on a chip | Ars Technica
https://arstechnica.com/gadgets/2018/11/apple-walks-ars-through-the-ipad-pros-a12x-system-on-a-chip/

Apple has also appealed at the new product launch that the graphics performance of A12X Bionic is very high and GPU performance appeared 1000 times as the first iPad released eight years ago, as a home game machine in 2016 It is said to be equivalent to Xbox One S. In addition, the CPU performance is also higher than the 92% notebook PC on the market, boasts that it achieves 90% faster performance than the A10X processor mounted on the previous model iPad Pro .

There are 8 cores in the A12X CPU, 4 core with focus on performance, 4 with core focused on efficiency. The difference from the conventional Apple processor is that "you can activate all 8 cores at once". Belonging to Apple hardware department of Anand dermis said, by the "performance controller with custom design, was now able to use all eight of the core at a time. This capability is A10X processor in a large load is applied work It's improved by 90% than that. "

Technology-based media Ars Technica conducted a benchmark test and compared the performance of iPad Pro with the MacBook Pro, the result was that the new iPad Pro outperformed the performance of most MacBook Pros. It is said that only the latest 15-inch MacBook Pro equipped with the 8th generation Intel Core i9 CPU blew out the performance surpassing the new iPad Pro, but this alone also shows how A12X is better. Shinpi said, "Usually this kind of performance is only realized on huge machines with fans, but with a very efficient design processor, even with a 5.9 mm thin iPad Pro, You can get the performance of it. "

Also, the A12X GPU has seven cores, one more than A10X. Mr. Shimpi said "About the first 7-core GPU designed independently, it is faster and more efficient than the A 10 X GPU so that A 12 X has twice the graphics performance compared to A 10 X It is said. " Mr. Shimpi also emphasized that this performance was realized without a fan.

The GPU performance of iPad Pro is prominent as a mobile terminal, but it is still not much more than a notebook PC. However, it still has excellent GPU performance as a mobile terminal, repeatedly mentioning that Apple has the same performance as the small version Xbox One S of Xbox One that appeared in 2016 It is.

In addition, the A12X has a neural engine, which provides high performance when processing related to machine learning is executed on CPU or GPU. The Apple processor before A11 installed in iPhone X does not have the neural engine itself, and Apple also says that "A12X's most advantage is a neural engine loading".

For many users it is difficult to understand what kind of benefits the speedup of machine learning by neural engines brings. Actually, the neural engine enhances face recognition function, enables Face ID installed in iPad Pro, enhances sensor of TrueDepth camera, features extraction in photo application, augmented reality application, etc. I have brought it to iPad Pro.

Apple focuses on machine learning on local devices to enhance user privacy. Although it is obvious that the efficiency of machine learning improves if you use huge amount of data on cloud services, Apple seems to focus on machine learning model on the device from the viewpoint of latency problem due to data transfer and user privacy is.

Philippe Schiller , senior vice president of marketing for Apple, said that instead of having Apple improve its processor's competitiveness in the process of competing with other companies, he said, "We concentrate on ourselves, I tried the best way I could do "and said that he succeeded in boosting performance dramatically. Siller said that for several years before the processor was actually released, it has taken close communication within Apple's organization and speeded up development within the organization.