I am trying to set up a Shader Storage Buffer Object to create an array of 16-bit data and an array of 8-bit data. I want to pass the 16-bit data to the shader, and then perform some calculations within the shader to stretch the 16-bit data to fit into an 8-bit format and read it out.

However, I’ve found that every time I read the data within the shader, it’s in 32-bit format, which means that in each workgroup, I can only process two 16-bit pieces of data and four 8-bit pieces of data. But this does not allow for a one-to-one mapping because I want to put the 16-bit data at the same index into the 8-bit data.

What should I do to make the one-to-one mapping work properly?

genBuffer：

    // some gpu init 
    uint16_t data[10] = { 1, 2, 3, 4, 5, 6, 7, 8, 9, 511}; 

    glGenBuffers(1, &data_buffer);
    glBindBuffer(GL_SHADER_STORAGE_BUFFER, data_buffer);
    glBufferData(GL_SHADER_STORAGE_BUFFER, sizeof(uint16_t) * 10, (void*)data, GL_STREAM_COPY);
    glBindBufferBase(GL_SHADER_STORAGE_BUFFER, 1, data_buffer);

    glGenBuffers(1, &output_buffer);
    glBindBuffer(GL_SHADER_STORAGE_BUFFER, output_buffer);
    glBindBufferBase(GL_SHADER_STORAGE_BUFFER, 2, output_buffer);

    glBufferData(GL_SHADER_STORAGE_BUFFER, sizeof(uint8_t) * 10, nullptr, GL_DYNAMIC_READ);

    ... // gpu use, dispatch, wait

    // print output data
     GLuint* ptr = (GLuint*)glMapBufferRange(GL_SHADER_STORAGE_BUFFER, 0, sizeof(uint8_t) * 10, GL_MAP_READ_BIT);
    uint8_t* tData = (uint8_t*)ptr;
    for (int i = 0; i < 10; i++) {
        std::cout << " " << static_cast<unsigned int>(*(tData + i));
    }

shader glsl:

#version 320 es
layout (local_size_x = 1, local_size_y = 1, local_size_z = 1) in;
layout(std430) buffer;
layout(binding = 2) writeonly buffer Output {
    uint elements[];
} output_data;
layout(binding = 1) readonly buffer Input0 {
    uint elements[];
} input_data0;
void main()
{
    uint ident = gl_GlobalInvocationID.x;
    // some processs
    //output_data.elements[ident] = input_data0.elements[ident];
    // test
    output_data.elements[ident] = 4u;
}

i try to run this code, then print

4 0 0 0 4 0 0 0 4 0

What I hope for the output is:

4 4 4 4 4 4 4 4 4 4