Benchmarking 4x ARM Cortex-A7 CPU and

Benchmarking 4x ARM Cortex-A7 CPU and

4x ARM Cortex-A53 for Multimedia Systems using JPEG Compression

Introduction: Now a days more and more research has been carried out on the field of embedded system. So, lots of new technologies and new board have been coming in the market from different vendors. With the increasing technology and growing demand for performing different unique tasks, the embedded computer takes the place of the traditional general purpose computer machine. Now a days the embedded systems are designed such a way that they are very much capable of doing specific task or talks with greater accuracy and efficiency. ARM is a very old vendor for designing different kind of CPUs. Two of their most popular CPUs are Cortex A7 and Cortex A53. Both consists of four cores. This research paper basically has shown

Get Help With Your Essay
If you need assistance with writing your essay, our professional essay writing service is here to help!
Essay Writing Service

Comparison of the performance between cortex A7 and Cortex A53 by using JEPG compression method using execution time as a parameter of performance. Both of these processors support ARM BIG. Little technologies. The main advantages of this technology is the have 2 power efficient cores that usually handles the low power background management tasks. But the rest of the two processors are power hungry and usually responsible for doing bigger tasks. Raspberry pi is a popular embedded device. It is a single board computer. Now a days raspberry pie is used in various kind of real time systems.  There are many versions of raspberry pie available in the market. Two of the most used raspberry pie visions are raspberry pie version 2 and the raspberry pie version 3. The first one has used ARM cortex-A7 CPU and the Raspberry Pie 3 has used 4x ARM cortex as CPU. JEPG image compassion is one of most common compression techniques which has been used here. The benchmarking was done on the basis of execution time. The execution time can be further divided into different modules such as Image reading, Display Image, discrete Cosine transformation, quantization and decoding. Thus, we have come to the conclusion that which of these models performs better than the other one.

Configuration of the Experiment: This paper has been used the raspberry pie 2 and raspberry pie 3 for the experiment. In the raspberry pie module, the JEPG compression algorithm is done using OPENCV. And the OpenCV in raspberry pie is supported by python programing language.

Also OpenCV is responsible for proving all the libraries that are responsible for implementing a successful JEPG image compression technique which use the raspberry pie’s CPU using python

Programming language. So, the entire system works like this- the input image – > JEPG compression module on raspberry pie – > Output or the compressed image. If we want to describe how the compressed image can be made or how the JEPG compressed algorithm works, we can divided the whole system into several point. Let’s discuss those below:

All it starts with the input raw images.

Then the color transform technique is done.

Thus, the entire image is divided into 8X8 blocks.

Next step is discrete cosine transformation. It basically uses the Fourier Transformation techniques which is a important topics for signal and image processing.

Next step is quantization and encoding. Encoding puts special character or pattern for efficient transmission or storage of data.

 The encoding which results in JEPG compressed image data.

Now for obtaining the original like image form it decode the JEPG compressed image data. Dequantization technique is also applied here.

For getting the continuous for of image inverse discrete transformation is applied on it.

 Then up sampling is applied.

For retrieving original like color, color transformation is applied.

After the JEPG compression technique each image is compared with each other in respect of image quality and compression execution time.

System Description:

Raspberry Pie: Raspberry is most popular on chip embedded computer which used arm processor as CPU. The Raspberry pie 2 has ARM cortex A7 and the raspberry pie 3 has ARM cortex A53. Raspberry pie is a small RISC (Reduced Instruction Set Computer) which makes it capable of carrying out several embedded tasks.

Cortex A7: Cortex A7 is an ARM designed processor which is famous for its low energy consumption. It has been built with 8 stage pipeline which has used cortex A5 chip. Thus, makes it extremely energy efficient CPU. It’s L2 cache is also responsible for the low energy consumption and low latencies. It has also got an improved 128bit AMBA bus. But the CPU is 32 bits. So, the ram capacity for this one is limited to around 4GB.Anything greater than 4GB won’t be recognized by this CPU.

Cortex A53: It supports both 32bit and 64bit which makes it to have more RAM. Cortex A53 has improved performance with low power consumption which make it extremely popular and demanding around the world. These features make it suitable for different embedded devices.

Results: In this paper they have used RGB images which are in PNG data format. The results are collected after the image compression is done. After observing the results, it is very clear that Cortex A53 is far more superior than cortex A7 in terms of execution time. In every steps like: reading images, displaying them, DCT transformation, quantization and encoding, the cortex A53 take less time than cortex A7. The total execution time of cortex A53 is 45% less than the cortex A7. So, cortex A53 is more powerful than cortex A7.

Conclusion: In this paper they haver tried to show us the overall performance comparison of two most popular processor of this days the cortex A53 and the cortex A7. The have used different sizes of images for this experiment. After the experiments it can be clearly concluded that the cortex A53 in raspberry pie 3 is superior than the cortex A7 in raspberry pie 2. This experiment has generated numerous datasets which are very important in understanding behavior and characteristics of this two processors.

References:

Benchmarking 4x ARM Cortex-A7 CPU and 4x ARM Cortex-A53 for Multimedia Systems using JPEG Compression, Waleed Khan, Nasru Minallah.

https://community.arm.com/tags/Armv7/Armv7_2D00_A

https://www.arm.com/products/silicon-ip-cpu/cortex-a/cortex-a53

https://developer.arm.com/ip-products/processors/cortex-a/cortex-a7

https://www.arm.com/products/silicon-ip-cpu/cortex-a/cortex-a7

https://en.wikipedia.org/wiki/ARM_big.LITTLE

Raspberry Pi 3 vs Pi 2: What’s the difference?