System on a chip aka SoC definition
So, the most
of the processors which you see inside the smartphone and tablet are designed
using ARM architecture. One more thing is
that the operating system which is designed for these smartphones is optimized for
this system on chips. So, although on paper these smartphone processors don't
look powerful compared to the desktop processors, because of this software optimization, you hardly feel any lag when you operating these smartphones and
tablets. Apart from that, the advancement in manufacturing technology has
allowed us to integrate more and more transistors inside the same area. So, not
only in the smartphone but on desktop and laptop also, you will see that many
parts are integrated inside the same chip.
What's inside a System on a Chip
So, these
are some of the reasons why you can integrate all the components
inside a single chip. So, now as we know about what is the System on the Chip, let's find out what is inside this System on the Chip.
- Central Processing Unit (CPU) - It's called the brain of SoC. It handles most of the calculations coding-decoding for your smartphone's operating system and almost all other applications in your phone. CPU is the widely used type of processor. It is designed so well that it can handle a variety of different tasks even heavier ones. Apart from running your operating system and apps, it also works as a bridge between other processors for synchronizing data. The CPU performance can be enhanced or lowered down by changing the clock speed which is in (GHz), or the number of cores can also be done by changing the basic architecture to do more work with each cycle.
- Graphics Processing Unit (GPU) - Along with the CPU, the GPU is another component embedded into an SoC. In terms of design, due to their less general-purpose comparing with CPUs, they are very different. It handles all the graphics-related tasks, like visualizing an application's user interface & gaming purposes. As opposed to the CPUs, which can execute only one or two commands with each cycle, GPUs can calculate enormous operations(calculations) with each cycle. In the mobile SoC world, two major GPUs are widely used: Qualcomm's Adreno and ARM's Mali.
- Image Signal Processor (ISP) - It works to convert the data of your phone's camera into Images and Videos. While the editing of an image is often done by the CPU and GPU, there's a lot of processing performed on a camera sensor data, before even you saved it to your phone. An ISP is basically a DSP that handles focusing, demosaicing, sharpening, and noise reduction. In layman's language, it converts all your digital data from the camera sensor into a picture.
- Neural Processing Unit (NPU) - Like other processing units, like GPUs designed for graphics math and ISPs, are designed for image tasks, NPUs are specially designed to execute neural network and machine learning tasks rapidly and effectively CPUs. They have their own local memory caches also, to boost up execution without the usage of RAM. These neural networks frequently need operations that collect multiple pieces of insert data to cause a single output. This is totally different from math used by CPUs.
- Video Encoder/Decoder - As the name suggests it basically handles the power-efficient conversion of video files and formats.
- Modems - Next you see inside a System On Chip are the connectivity modems. And these modems ensures that user has all kind of connectivity like- Wi-Fi, FM radio, USB, Bluetooth, GPS, cellular radios, 4G LTE & 5G modems. It converts wireless signals into data in your phone. The 4G modem is situated inside the SoC while the early 5G modems are external, which means they have to attach to the SoC.
Good explanation
ReplyDeleteThanx man...keep reading and sharing
ReplyDelete