Report an Error

AMD Vega 20

Default
Vega 20
Vega 20 XT
Vega 20 XT
Block Diagram
Block Diagram
CU Diagram
CU Diagram
AMD's Vega 20 GPU uses the GCN 5.1 architecture and is made using a 7 nm production process at TSMC. With a die size of 331 mm² and a transistor count of 13,230 million it is a large chip. Vega 20 supports DirectX 12 (Feature Level 12_1). For GPU compute applications, OpenCL version 2.0 can be used. It features 4096 shading units, 256 texture mapping units and 64 ROPs.
Further reading: Vega Architecture Whitepaper , Vega Instruction Set Architecture

Graphics Processor

GPU Name
Vega 20
Architecture
GCN 5.1
Foundry
TSMC
Process Size
7 nm
Transistors
13,230 million
Density
40.0M / mm²
Die Size
331 mm²
Released
Nov 18th, 2018

Graphics Features

DirectX
12 (12_1)
OpenGL
4.6
OpenCL
2.0
Vulkan
1.2
Shader Model
6.4
Compute
GFX9
ROCm
Yes

Render Config

Shading Units
4096
TMUs
256
ROPs
64
Compute Units
64
ACEs
4
HWSs
2
GEs
4
L1 Cache
16 KB per CU
L2 Cache
4096 KB
Max. TDP
475 W

All GCN 5.1 GPUs

AMD GPU Architecture History

Graphics cards using the AMD Vega 20 GPU

Name Chip Memory Shaders TMUs ROPs Base Clock Boost Clock Memory Clock
AMD Radeon Instinct MI60 Vega 20 GL 32 GB 4096 256 64 1200 MHz 1800 MHz 1000 MHz
AMD Radeon Pro Vega II Duo Vega 20 XT 32 GB 4096 256 64 1400 MHz 1720 MHz 1000 MHz
AMD Radeon Instinct MI50 Vega 20 GLXT 16 GB 3840 240 64 1200 MHz 1746 MHz 1000 MHz
AMD Radeon VII Vega 20 XT 16 GB 3840 240 64 1400 MHz 1750 MHz 1000 MHz
AMD Radeon Pro Vega II Vega 20 XT 32 GB 4096 256 64 1574 MHz 1720 MHz 806 MHz
AMD Radeon Pro VII Vega 20 GL XT 16 GB 3840 240 64 1400 MHz 1700 MHz 1000 MHz

Vega 20 GPU Notes

Architecture Codename: Vega
Codename: Moonshot
Old Codename: Aruba
CLRX Version: GCN 1.4.1
Graphics/Compute: GFX9 (gfx906) / (gfx907)
Display Core Engine: 12.1
Unified Video Decoder: 7.2
Video Compression Engine: 4.1