GeekRadar / The comparison desk
COMPARE.
GPU
NVIDIA

NVIDIA Switch GPU 20nm

Graphics processor

Video memory
4 GB*
Memory bus
64 bit*
Rated memory bandwidth
25.60 GB/s*

No matched benchmark results are available for this entry yet.

Processing & platform

Architecture
Maxwell 2.0*
Process node
20 nm*
GPU cores
256*
Reported GPU frequency
384 MHz*
Reported chip die area
118 mm²*
Chip/package TDP
15 W*
Transistors
2,000 million*
Graphics processor
GM20B*
Texture mapping units
16*
Raster output units
16*
Chip foundry
TSMC*
Declared GPU TDP
15 W*
DirectX / feature level
12 (12_1)*
OpenGL
4.6*
OpenCL
1.2*
Vulkan
1.4*
CUDA compute capability
5.3*
Shader model
6.0*

Memory & storage

Memory options
4*
Video memory
4 GB*
Video memory type
DDR4*
Memory bus
64 bit*
Rated memory bandwidth
25.60 GB/s*

Power specifications

Thermal design power
15 W*

System configuration

Detailed specifications

Specifications 1 · 13 fields
Foundry
TSMC*
Vendor
NVIDIA*
FP16 GFLOPS
786.4* GFLOPS
FP32 GFLOPS
393.2* GFLOPS
FP64 GFLOPS
12.29* GFLOPS
Product
NVIDIA Switch GPU 20nm*
Type
GPU*
Release Date
2017-03-17*
Process Size (nm)
20* nm
TDP (W)
15* W
Die Size (mm^2)
118* mm²
Transistors (million)
2,000*
Freq (MHz)
384* MHz
Specifications 2 · 24 fields
proc transistors
2,000 million*
proc die size
118* mm²
cores
256*
base clock
384*
boost clock
768*
memory clock
1,600*
ram
4*
name
NVIDIA Switch GPU 20nm*
ram type
DDR4*
bus width
64*
ram bandwidth
25.6* GB/s
fp16 performance
0.7864*
fp32 performance
0.3932*
fp64 performance
0.0123*
tdp
15* W
api cuda
5.3*
is retail board
No*
gpi value
0.52*
vendor
nvd*
manufacturer
NVIDIA*
arch name
Maxwell 2.0*
card release date
2017-03-17*
proc foundry
TSMC*
proc process size
20* nm
Specifications 3 · 31 fields
CUDA compute capability
5.3*
DirectX / feature level
12 (12_1)*
OpenCL
1.2*
OpenGL
4.6*
Shader model
6.0*
Vulkan
1.4*
Architecture
Maxwell 2.0*
Memory bus
64* bit
Reported shader processors
256*
Reported theoretical FP16 peak
0.7864* TFLOPS
Scope and conditions
Arithmetic density
Not stated*
Reported theoretical FP32 peak
0.3932* TFLOPS
Scope and conditions
Arithmetic density
Not stated*
Reported theoretical FP64 peak
0.0123* TFLOPS
Scope and conditions
Arithmetic density
Not stated*
Reported display outputs
1x USB Type-C*
Theoretical pixel fill rate
12.29 GPixel/s*
Reported chip die area
118 mm²*
Chip foundry
TSMC*
Process node declaration
20 nm*
Transistors
2,000 million*
Graphics processor
GM20B*
Reported video memory
4* GB
Rated memory bandwidth
25.60 GB/s*
Video memory type
DDR4*
Raster output units
16*
Reported shading units
256*
2:4 tensor sparsity
No*
Declared GPU TDP
15 W*
Theoretical texture fill rate
12.29 GTexel/s*
Texture mapping units
16*
GPU cores
256*
Video memory
4* GB
Video memory type
DDR4*

* Provisional data