Disabling integer overflow capping for uint8 datatypes

Question

JMB 2024-6-20

0
链接

此问题的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/2130621-disabling-integer-overflow-capping-for-uint8-datatypes

编辑： Voss 2024-6-20

Hello!

I am currently in the middle of optimizing some high-performance code, and have run into a weird situation where I need integer overflow to occur on some uint8 variables. Unfortunately, it seems like MATLAB features datatype capping, meaning that when I perform the operation

uint8(100) * uint8(30)

I get a value of 255, where I would expect something like 184.

The same occurs when I perform addition:

uint8(100) + uint8(200)

Here, I also expect to get 44, but instead get 255.

Is there a way to turn this capping off, so that I can have the overflow occur? I understand that this could potentially be a "dangerous" operation to perform, but I really would benefit from having a way to test some code with integer overflow in MATLAB before I do any form of compilation or other code optimization work.

0 个评论
显示 -2更早的评论隐藏 -2更早的评论

请先登录，再进行评论。

请先登录，再回答此问题。

Answer 1

Voss 2024-6-20

0
链接

此回答的直接链接

https://ww2.mathworks.cn/matlabcentral/answers/2130621-disabling-integer-overflow-capping-for-uint8-datatypes#answer_1475111

编辑：Voss 2024-6-20

在 MATLAB Online 中打开

I don't think you can disable integer overflow capping, but you could of course perform the operations on corresponding floating-point variables and then take the results mod 256.

x = uint8(100);
y = uint8(30);
z = uint8(200);
% original integer arithmetic
x * y
ans = uint8
    255
x + z
ans = uint8
    255
% with capping "disabled"
uint8( mod( double(x) * double(y) , 256) )
ans = uint8
    184
uint8( mod( double(x) + double(z) , 256) )
ans = uint8
    44