Comparison of Smartphones Hardware Trends

Abstract:
Mobile device becomes very important part of our daily life. Upcoming technologies
break the concept of limitation in terms of Architecture, memory and performance. This paper
deals with the growth of mobile Architecture, mobile memory, mobile performance and
additional features like accelerometers and compasses. As time passes mobile processors are
growing rapidly. Communication, performance, and low-power operation are affecting the design
and deployment of mobile processors over the years. To make a smartphone work like a
computer manufacturers make phones with processor like ARM11, CortexA8, Snapdragon,
Tegra and Armada etc. Choosing flash memories for mobile phones will immensely help
manufacturers to achieve their goals i.e. more flash memories with less cost. Mobile RAMs
are optimized to reduce the battery usage and increase the performance of the smart phones.
Mobile storage is an important factor for mobile developers because smart phones should have
enough memory to handle installing, running the application and switching from one application
to another. Identifying the problems and finding a solution to implement accelerometer and
electronic compasses in smart phones without depending of Wi-Fi infrastructure is very complex.
Noisy phone sensors and human movement makes it tough to deal with the implementation of
the compass. This paper also describes the comparisons of mobile phones with respect to the
operating system, Graphical processing unit, Control processing unit, Memory size, Storage and
Battery life.

1. INTRODUCTION:
As mobile phones become more versatile, they are used for various purposes.
Smartphone hardware consists of processors, Random Access Memory (RAM), Energy
consumption and additional real-time features like accelerometers, etc. Processors that are made
for mobile phones are restricted in terms of power, cost and real time computational
requirements. New generation high performance processor chips can overcome those limitations.
Performance related aspects are a major concern in todays society due to the rapid growth of
usage of Smartphones in day to day life. Memory size greatly increases in every upcoming
release of new smartphones [5]. Size and weight of the mobile phone affects the battery size.
Modern high end devices like smartphones are performing computer-like functionalities, which
squeezes out the energy from the small size battery and deepens the need for effective energy
management [1]. Therefore, this paper mainly concentrates on the comparisons of hardware
components of smartphones evolving in the 21st century.
This paper is organized as follows: Section 2 describes the processor chip evolution in
Smartphone history. Section 3 and Section 4 depict different processors and the memory
evolution over the decade, respectively. Section 5 discusses additional features of Smartphones
and finally, Section 6 summarizes all of the smartphone attributes.
2. Smartphone Processor Architecture:
There are many distinctions in mobile processors based on trends, technologies and
performance.

2.1 Digital signal Processor (DSP) Architectures

There are two types of Digital signal Processor based on release and timeline. They
are Traditional DSP Architectures and Modern DSP Architecture. Traditional DSP Architecture is
designed as single standalone integrated circuits (ICs). They can be programmed to be used in
wireless handsets for digital cellular telephony. Modern DSP Architectures are evolved for
Smartphones, to perform parallel processing and to have more computational power due to more
chip fabrication [4].
2.2 System on Chip (SOC) Architecture
Smartphone device architecture becomes simpler by using System on Chip. DSP hybrid
(DPA with SOC) architecture versions can handle the responsiveness in a device to improve
performance. SOC architecture provides the flexibility of having high performance with low
power design. To speed up the operations of Smartphones, many companies have implemented
instruction set customizations. Most of the Smartphones today, use a single or dual core SOCs
[4].
3. Smartphone Processors Types
3.1 ARM (Advanced RISC Machines) Processor
Advanced RISC Machines processors are used in Smartphones because of less power
consumption and better performance. It is basically a 32-bit instruction set on Reduced
Instruction Set Computing (RISC) architecture. The first ARM architecture version is ARMv4
used for low-end Smartphones, followed by ARMv6, ARMv7, ARMv8. ARMv8 is the latest
version, which contains a hardware floating point unit providing speed improvement [4]. A

single core ARM cortex v8 was clocked with 1.4GHz which is reasonable in the year 2011, but
in 2014 ARM processors were clocked with 3GHz. ARM processors with 64 bit perform twice as
fast compared to 32 bit ARM processors [4].
3.2 Qualcomm Snapdragon Processors
Snapdragon comes under the family of SOC architecture provided by Qualcomm. It has
the same features as ARM, but with an included feature utilizing Single Instruction, Multiple
Data (SIMD). Qualcomm Snapdragon improves graphics performance of smartphone games.
Qualcomm was the first to introduce a 28nm processor in 2012 [4]. In 2013, Qualcomm
Snapdragon released the Snapdragon 800 processor which beats all the processors in the field of
Smartphones. Snapdragon 800 opens apps and web browsers very fast and it has outstanding
graphics, good battery life, ultra HD video and multiple HD audio [4].
3.3 NVidias Tegra Processors
Tegra is also a SOC architecture developed by NVidia. It includes an ARM processor,
graphic processing unit and memory controllers. NVidia Tegra 4 processor is a quad core SOC
with improved performance. NVidia is technically faster, but when power efficiency is compared
Qualcomm snapdragon wins. Different processor performance can be compared using the
software called GeekBench3, which calculates the benchmark utilizing a baseline score of 2500
(which is the score of Intel Core i5) [4]. Score that is higher than the benchmark score implies
better processor performance. Figures 1 shows Android operating system processors,
benchmarks which reflect the best processors in 2015.

Figure 1: Android Operating System (OS) processor benchmarks [7].

4. Smartphone Memory:
The enormous growth of smartphone users led to an increase in the demand for app
developers which in turn increases the demand for memory space in smartphones. Smartphones
are provided with two types of memory: Internal flash storage and an External SD card. Flash memory is
of electronically Erasable Programmable Read-Only Memory (EEPROM). EEPROM erases data as a
group of blocks so that it is suitable for applications where astronomically immense amounts of data
require frequent update. Flash memory uses two distinct logical technologies NOR and NAND. NOR is
known for providing high speed reading, writing and access. The main purpose of NOR flash is to store
the mobile Operating System (OS). NAND flash reads and writes consecutively at high speed which
handles a single block of data called pages. NAND flash reads quicker than it writes and transfers the
whole page of data. Many reviews do not consider internal NAND storage speed as a key factor for
specifications. To compare the NAND performance of most popular Smartphones, the AndroBench tool is
used, which tests the sequential and random speed [5]. AndroBench test focuses on large amount of data

storage. For example, to transfer a movie stored in a sequential block of memory. The faster the results
are, the faster is the ability to write or read a file. Figure 2 shows the results.

Figure 2: Reading and writing speeds of different Smartphones [5].

Android and Apple iOS use the SQLite database to store the structured data as a primary
means of storage. SQLite is a light weight database which occupies less memory space.
Applications which have querying, retrieving and searching functions use one or more SQLite
databases. Apple iOS core data is a data model which is built on top of SQLite providing access
to applications, such as save, restore, undo and redo [3]. When the focus shifts to external
storage, a large memory size always lead to a decrease in speed. Smartphones come with two
different and standard memory cards called mini and micro-SD. They may vary in storage
capacities from 16MB to 2TB. Table 1 lists SD cards performances on desktop and smartphones
with respect to reading and writing speed.

Table 1: SD card cost and performance [3].

5. Additional features of Smartphones:

Implementation of additional feature, such as accelerometer or electronic compass,
introduces challenges in predicting the walking speed and orientation of smartphone users. At
first, the map is downloaded based on the user location to track the user. Using the assistance
GPS reading, the phone calculates the location of a user frequently. Implementation of the
accelerometer and electronic compass together form CompAcc. It is first implemented on Nokia
N95 and works well for a small range [2].
6. Conclusion:
Different vendors are working for better smartphones which contain a powerful processor
chip. Though all processors are ARM-based, they design with a different name by various
cellular companies. Though enough progress is achieved in Smartphone development, to achieve
an optimum level of satisfaction some factors like low power consumption, huge memory usage
by applications, etc., have to be improved.

BIBLIOGRAPHY:
[1] Bi, M., Duan, R., Gniady C., Exploring memory energy optimizations in smartphones,
Proceeding IGCC '11 Proceedings of the 2011 International Green Computing Conference and
Workshops, IEEE Computer Society Washington, DC, USA , pages 1-8.
[2] Constandache, I., Choudhury, R. R., Rhee, I., CompAcc: Using Mobile Phone Compasses
and Accelerometers for Localization, in IEEE INFOCOM, San Diego, CA, USA, march 2010.
[3] Kim, H., Agrawal, N., and Ungureanu, C. Revisiting storage for smartphones. In Proceedings
of USENIX Conference on File and Storage Technologies (FAST) 2012, USENIX Association.
[4] Singh, M, P., Jain, M, K., Evolution of Processor Architecture in Mobile Phones, Proceedings
International Journal of Computer Applications (0975 8887), Volume 90 No 4, March 2014.
[5] Victor, H., Phone Arena, PhoneNews, Android storage speed comparison: which phone has
the fastest IO performance, 04 Feb2015, http://www.phonearena.com/news/Android-IO-speedcomparison-which-phone-has-the-fastest-internal-storage_id65588.

[6]G.Geoffre.,GeekBenchBrowser,AndroidBenchmarks,http://browser.Primatelabs.com/android
-benchmarks.

[7] K, Russel., Computerworld, FlashMemory, 7Jun,2010, http://www.computerworld.com/

article/ 2550624/data-center/flash-memory.html