Architecture
NVIDIA Ampere 8nm
Most Powerful Variant
GN20-E7
Cores : TMUs : ROPs
6144 : 192 : 96
Typical Memory Size
16384 MB
Typical Memory Bus
256 -bit
GA104 GPU
Overview
Manufacturer
NVIDIA
Release Date
January 12th, 2021
Graphics Processing Unit
Architecture
Ampere
Fabrication Process
8 nm
Die Size
392 mm2
Transistors Count
17.4B
Tensor Cores
192
RT Cores
48
CUDAs
6144
TMUs
192
ROPs
96
Clocks
Boost Clock
1710 MHz
Memory Configuration
Memory Size
16384 MB
Memory Type
GDDR6
Memory Bus Width
256-bit
Memory Bandwidth
448.0 GB/s
Performance
Pixel Fillrate
164.2 GPixels/s
Texture Fillrate
328.3 GTexel/s
SPFP Performance
21 TFLOPS
Performance per mm2
53.6 GFLOPS/mm2
Physical
Interface
PCI-Express 4.0 x16
SLI Support
Unsupported
API Support
DirectX
12.2
Vulkan
1.2
OpenGL
4.6
OpenCL
2.1
| Model | GPU Model | Unified Cores | Boost Clock | Memory Clock | Memory Config. | |
|---|---|---|---|---|---|---|
 | NVIDIA GeForce RTX 3080 Mobile | GA104-775 | ||||
 | NVIDIA GeForce RTX 3070 | GA104-300 | ||||
 | NVIDIA GeForce RTX 3070 Mobile | GA104-770 | ||||
 | NVIDIA GeForce RTX 3060 Ti | GA104-200 |