From OpenGL Wiki
Jump to: navigation, search
Core in version 4.5
Core since version 4.3
Core ARB extension ARB_vertex_attrib_binding

glVertexAttribFormat: specify the organization of vertex arrays

Function Definition

 void glVertexAttribFormat(GLuint attribindex​, GLint size​, GLenum type​, GLboolean normalized​, GLuint relativeoffset​);
 void glVertexAttribIFormat(GLuint attribindex​, GLint size​, GLenum type​, GLuint relativeoffset​);
 void glVertexAttribLFormat(GLuint attribindex​, GLint size​, GLenum type​, GLuint relativeoffset​);
The generic vertex attribute array being described.
The number of values per vertex that are stored in the array.
The type of the data stored in the array.
GL_TRUE if the parameter represents a normalized integer (type​ must be an integer type). GL_FALSE otherwise.
The offset, measured in basic machine units of the first element relative to the start of the vertex buffer binding this attribute fetches from.


glVertexAttribFormat, glVertexAttribIFormat and glVertexAttribLFormat specify the organization of data in vertex arrays. attribindex​ specifies the index of the generic vertex attribute array whose data layout is being described, and must be less than the value of GL_MAX_VERTEX_ATTRIBS.

size​ determines the number of components per vertex are allocated to the specifed attribute and must be 1, 2, 3 or 4. type​ indicates the type of the data. If type​ is one of GL_BYTE, GL_SHORT, GL_INT, GL_FIXED, GL_FLOAT, GL_HALF_FLOAT, and GL_DOUBLE indicate types GLbyte, GLshort, GLint, GLfixed, GLfloat, GLhalf, and GLdouble, respectively; the values GL_UNSIGNED_BYTE, GL_UNSIGNED_SHORT, and GL_UNSIGNED_INT indicate types GLubyte, GLushort, and GLuint, respectively; the values GL_INT_2_10_10_10_REV and GL_UNSIGNED_INT_2_10_10_10_REV indicating respectively four signed or unsigned elements packed into a single GLuint; and the value GL_UNSIGNED_INT_10F_11F_11F_REV indicating three floating point values packed into a single GLuint.

glVertexAttribLFormat is used to specify layout for data associated with a generic attribute variable declared as 64-bit double precision components. For glVertexAttribLFormat, type​ must be GL_DOUBLE. In contrast to glVertexAttribFormat, which will cause data declared as GL_DOUBLE to be converted to 32-bit representation, glVertexAttribLFormat causes such data to be left in its natural, 64-bit representation.

For glVertexAttribFormat, if normalized​ is GL_TRUE, then integer data is normalized to the range [-1, 1] or [0, 1] if it is signed or unsigned, respectively. If noramlized​ is GL_FALSE then integer data is directly converted to floating point.

relativeoffset​ is the offset, measured in basic machine units of the first element relative to the start of the vertex buffer binding this attribute fetches from.

glVertexAttribFormat should be used to describe vertex attribute layout for floating-point vertex attributes, glVertexAttribIFormat should be used to describe vertex attribute layout for integer vertex attribute and glVertexAttribLFormat should be used to describe the layout for 64-bit vertex attributes. Data for an array specified by glVertexAttribIFormat will always be left as integer values; such data are referred to as pure integers.


GL_UNSIGNED_INT_10F_11F_11F_REV is accepted for type​ only if the GL version is 4.4 or higher.


GL_INVAILD_VALUE is generated if attribindex​ is greater than or equal to the value of GL_MAX_VERTEX_ATTRIBS.

GL_INVALID_VALUE is generated if size​ is not one of the accepted values.

GL_INVALID_VALUE is generated if relativeoffset​ is greater than the value of GL_MAX_VERTEX_ATTRIB_RELATIVE_OFFSET.

GL_INVALID_ENUM is generated if type​ is not one of the accepted tokens.

GL_INVALID_OPERATION is generated if no vertex array object is bound.

Associated Gets


glGetVertexAttrib with argument GL_VERTEX_ATTRIB_RELATIVE_OFFSET.

See Also

glBindVertexArray, glEnableVertexAttribArray, glVertexAttribBinding, glVertexAttribPointer, glVertexBindingDivisor


Copyright © 2012 Khronos Group. This material may be distributed subject to the terms and conditions set forth in the Open Publication License, v 1.0, 8 June 1999.