std140

std140 is a memory layout standard used by GLSL to define how data is arranged in uniform blocks and related shader interfaces. It provides a predictable, portable way to transfer data between CPU and GPU across different drivers and platforms, notably in OpenGL and WebGL 2.

The layout specifies alignment and padding rules for types within a block. Scalar types such as float, int, and bool have a base alignment of 4 bytes. A vec2 has an alignment of 8 bytes, while vec3 and vec4 have an alignment of 16 bytes. A matrix is stored as an array of its column vectors, so mat2 consists of two vec2 columns, mat3 of three vec3 columns, and mat4 of four vec4 columns; therefore the base alignment of a matrix corresponds to the alignment of its column type, and its size is the number of columns times the size of a column.

Within a block, each member’s offset must be a multiple of its base alignment. The size of a struct or block is padded so its overall size is a multiple of 16 bytes. Arrays follow the same rules: each element starts at a multiple of the base alignment of the element, and the stride between elements is the element size rounded up to a multiple of 16 bytes.

When writing data from the CPU, it is common to declare uniform blocks with layout(std140) and coordinates such as projection matrices, colors, and scalars in carefully padded slots. For cross-platform correctness, developers often rely on querying the block’s data size with OpenGL functions to ensure the buffer matches the declared std140 layout.

Compared with std430, std140 enforces more padding, providing stronger alignment guarantees at the cost of larger buffers.