UINT16 - 16-Bit Unsigned Integer | Integer Format Guide

Bit Layout

Value

V

16 bits

A 16-bit unsigned integer uses all 16 bits to represent magnitude, giving a range of 0 to 65,535 (2¹⁶ - 1). It is the unsigned short type in C/C++ and u16 in Rust.

Overview

UINT16 is ideal for compact storage of non-negative values that fit within 0–65,535. It uses half the memory of UINT32 while covering a range sufficient for many common applications.

Common Uses

Graphics index buffers: UINT16 is the standard type for mesh vertex indices in OpenGL and Vulkan, supporting up to 65,535 vertices per draw call at half the memory cost of UINT32.
Unicode: UTF-16 code units are UINT16 values, covering the Basic Multilingual Plane (most common characters) in a single unit.
Network protocols: TCP/UDP port numbers (0–65,535) are stored as UINT16.
Sensor data: 16-bit ADC readings, telemetry values, and fixed-width file format fields.

See also: INT16 For the signed variant using two's complement, commonly used for 16-bit PCM audio samples.

Range & Properties

Property	Value
Total Bits	16
Minimum Value	0
Maximum Value	65,535
Representable Values	65,536

Key Bit Patterns

Value	Bit Pattern	Signed	Unsigned
Zero	0000000000000000	0	0
One	0000000000000001	1	1
Max Unsigned	1111111111111111	-1	65535
Max Signed	0111111111111111	32767	32767
All Ones (-1)	1111111111111111	-1	65535
Min Signed (-32768)	1000000000000000	-32768	32768

Interactive Bit Visualizer

Click any bit to flip it. All 16 bits contribute directly to the unsigned magnitude.

Decimal:

Hex:

Format Comparison

Property	UINT16	UINT32	UINT8	UINT4
Total Bits	16	32	8	4
Signed Min	-32,768	-2,147,483,648	-128	-8
Signed Max	32,767	2,147,483,647	127	7
Unsigned Max	65,535	4,294,967,295	255	15
Step (ULP)	1	1	1	1
Values	65,536	4,294,967,296	256	16

Where UINT16 Is Used

GPU packed operations: The PTX ISA 9.2 introduces .u16x2 packed types (two 16-bit unsigned integers in a 32-bit register) with add/sub/min/max and .sat qualifier support.
Graphics index buffers: UINT16 is the standard vertex index type in OpenGL and Vulkan, supporting up to 65,535 vertices per draw call at half the memory cost of UINT32 indices.
ML frameworks: PyTorch provides torch.uint16. The ONNX specification defines UINT16 = 4 as a tensor data type.
Unicode and networking: UTF-16 code units and TCP/UDP port numbers are both UINT16 values, making this the standard type for text processing and network protocol implementations.