Report an Error

AMD Barts

Default
Barts
Die Shot
Shane Phillips
Die Shot
Block Diagram
Block Diagram
AMD's Barts GPU uses the TeraScale 2 architecture and is made using a 40 nm production process at TSMC. With a die size of 255 mm² and a transistor count of 1,700 million it is a medium-sized chip. Barts supports DirectX 11.2 (Feature Level 11_0). It features 1120 shading units, 56 texture mapping units and 32 ROPs.
Further reading: TeraScale 2 Instruction Set Architecture

Graphics Processor

GPU Name
Barts
Architecture
TeraScale 2
Foundry
TSMC
Process Size
40 nm
Transistors
1,700 million
Density
6.7M / mm²
Die Size
255 mm²
Released
Oct 21st, 2010

Graphics Features

DirectX
11.2 (11_0)
OpenGL
4.4
OpenCL
1.2
Vulkan
N/A
Shader Model
5.0
Compute
GFX4

Render Config

Shading Units
1120
TMUs
56
ROPs
32
Compute Units
14
Z-Stencil
40
L1 Cache
16 KB per CU
L2 Cache
512 KB
Max. TDP
300 W

All TeraScale 2 GPUs

Graphics cards using the AMD Barts GPU

Name Chip Memory Shaders TMUs ROPs GPU Clock Memory Clock
AMD Radeon HD 6850 Barts PRO 1024 MB 960 48 32 775 MHz 1000 MHz
AMD Radeon HD 6870 Barts XT 1024 MB 1120 56 32 900 MHz 1050 MHz
AMD Radeon HD 6790 Barts LE 1024 MB 800 40 16 840 MHz 1050 MHz
AMD Radeon HD 6870 X2 Barts XT 1024 MB 1120 56 32 900 MHz 1050 MHz
AMD Radeon HD 6850 X2 Barts PRO 2 GB 960 48 32 800 MHz 1050 MHz
AMD Radeon HD 7720 OEM Barts LE 1024 MB 800 40 32 840 MHz 1050 MHz

Barts GPU Notes

Architecture Codename: Northern Islands
Codename: Victoria
Graphics/Compute: GFX4
Display Core Engine: 5.0
Unified Video Decoder: 3.1