Galaxy S9 Should Fight For Samsung's New Hardware
One of the keys to Apple’s new facial recognition system is in the dedicated silicon and the AI techniques being used to compare faces to data stored locally on the handsets. Intensive processing is always better hard-coded in a chip than left to software.
Using custom hardware is route that Apple has exploited heavily with the iPhone. After the constraints CPU choice and development speeds with the Mac computers, Apple ensured the same would not be true in its smartphone devices. The Axxx system on chips lie at the heart of the iOS ecosystem. These chips can be developed in tandem with the operating system, safe in the knowledge that there will be no other relationship.
That’s not the case with Android. That operating system has to work with a wide range of hardware - off the top of my head you have Qualcomm’s Snapdragon, Samsung’s Exynos, and Huawei’s Kirrin range as compatible system on chips. Then you have the various manufacturers, countless chip providers, alternate screens and interfaces… Android needs the extra power so that the compatibility layers in software can reach equivalent benchmarks in iOS devices.
Which is where the reports of Samsung looking at developing dedicated AI processing engines in the CPU cores of its smartphones is exciting. The South Korean company already uses AI in its Android offerings through the Bixby digital assistant and smart speakers. That uses AI techniques to process natural language but this uses a fair bit of cloud-based processing.
Dedicated hardware for AI processing in any Galaxy handset would bring more of the work into the handset, would increase the efficiency of the whole package, and open up more possibilities that need more processing power.
Now read more about Samsung’s futuristic folding phone, the Galaxy X…
">Samsung’s Bixby could be getting a little more horsepower in the future, as Samsung is reported to be working on improving the AI potential of its smartphone hardware. Given the increased focus on AI in mobile devices and the advantages that dedicated hardware offer, this is the right track to be on but the South Korean company has some catching up to do.
One of the keys to Apple’s new facial recognition system is in the dedicated silicon and the AI techniques being used to compare faces to data stored locally on the handsets. Intensive processing is always better hard-coded in a chip than left to software.
Using custom hardware is route that Apple has exploited heavily with the iPhone. After the constraints CPU choice and development speeds with the Mac computers, Apple ensured the same would not be true in its smartphone devices. The Axxx system on chips lie at the heart of the iOS ecosystem. These chips can be developed in tandem with the operating system, safe in the knowledge that there will be no other relationship.
That’s not the case with Android. That operating system has to work with a wide range of hardware - off the top of my head you have Qualcomm’s Snapdragon, Samsung’s Exynos, and Huawei’s Kirrin range as compatible system on chips. Then you have the various manufacturers, countless chip providers, alternate screens and interfaces… Android needs the extra power so that the compatibility layers in software can reach equivalent benchmarks in iOS devices.
Which is where the reports of Samsung looking at developing dedicated AI processing engines in the CPU cores of its smartphones is exciting. The South Korean company already uses AI in its Android offerings through the Bixby digital assistant and smart speakers. That uses AI techniques to process natural language but this uses a fair bit of cloud-based processing.
Dedicated hardware for AI processing in any Galaxy handset would bring more of the work into the handset, would increase the efficiency of the whole package, and open up more possibilities that need more processing power.
Now read more about Samsung’s futuristic folding phone, the Galaxy X…
Post a Comment