Little Endian, Big Endian difference

 

 Problem

Occasionally, when running a program written in C and C++, communication failure occurs between computer A and computer B.

It's a common problem when computer A's CPU manufacturing is Intel and computer B's manufacturer is AMD

It is likely due to the difference between Little Endian and Big Endian, which is one of the CPU's data storage methods.

The Endian method represents the order in which a computer stores data in bytes, which plays an important role in communication and data exchange. Intel CPUs adopt the Little Endian method and store it in memory from the lowest bytes, while AMD CPUs use the Big Endian method to store it from the highest bytes. This difference in Endian can lead to problems due to inconsistent order of data during data communication between the two CPUs. Accordingly, during programming work, Endian conversion of data must be performed in consideration of communication between the two CPUs.

 

 

 Endian Convert

Endian conversion of data plays an important role in data communication between Intel CPUs and AMD CPUs. This is a process of aligning the bytes of data so that the two CPUs can understand each other, and Endian conversion is mainly required for data exchange tasks such as network communication or file input/output.


Intel CPUs and AMD CPUs have differences in how data is stored. Intel CPUs store data using the Little Endian method. In this method, the lowest byte (LSB) is stored at the lowest address of a memory, and the highest byte (MSB) is stored at the highest address. For example, a 32-bit integer of 0x12345678 is equal to 0x12345678

It is stored in memory as follows: 0x78 0x56 0x34 0x12.

Meanwhile, the AMD CPU stores data using the Big Endian method, in which the highest byte is stored at the lowest address, and the lowest byte is stored at the highest address. In the above example, a 32-bit integer of 0x12345678 is used to store data

It is stored in memory as follows: 0x12 0x34 0x56 0x78.

Therefore, when transmitting data generated by the Intel CPU to the AMD CPU, an Endian conversion operation of the data is required. Through this conversion operation, communication between the two CPUs may be smoothly performed by changing the byte order of the data.

Endian conversion of the data is carried out in the following process:

1. Check the byte order of the data
2. If the data was generated by the Intel CPU, it would be stored in the Little Endian method, and if it was generated by the AMD CPU, it would be stored in the Big Endian method.
3. Change the byte order of the data.
To do this, you can rearrange bytes in reverse order or use a different type of conversion algorithm.
5. The converted data is transmitted to the destination. In this case, the destination refers to a CPU or a device that will use data.

Endian transforms can be implemented simply using a library or function provided by the programming language. For example, C or C++ provides a variety of Endian transform functions. I will introduce it later when I have time.
In addition, in network programming, Endian conversion should be considered even in socket communication.

When exchanging data in the network, the Endian of the data should be converted in consideration of compatibility with the other party's system.

Endian conversion of data can also affect performance. It is important to process it efficiently, especially when converting a large amount of data. Therefore, the performance can be improved by optimizing the Endian conversion algorithm.

Finally, Endian conversion of data is important to ensure interoperability between various systems. Endian conversion should always be considered when exchanging data between systems using different Endian methods. Through this, the accuracy and reliability of data communication can be maintained, and program compatibility in various environments can be guaranteed.

 

 

 Last

If a program that I worked with C or C++ has a communication problem, I hope that the solution will be solved well.
And, For developers who are exposed to low-level programming, it seems beneficial to unify the CPU of a company or company into a CPU of one manufacturer. If you experience an error due to this problem, you may be wasting a lot of time and resources.