AMD Radeon Instinct MI50 vs NVIDIA Tesla V100 PCIe

Comparative analysis of AMD Radeon Instinct MI50 and NVIDIA Tesla V100 PCIe videocards for all known characteristics in the following categories: Essentials, Technical info, Video outputs and ports, Compatibility, dimensions and requirements, API support, Memory, Technologies.

AMD Radeon Instinct MI50

NVIDIA Tesla V100 PCIe

Differences

Reasons to consider the AMD Radeon Instinct MI50

Videocard is newer: launch date 1 year(s) 4 month(s) later
Around 27% higher boost clock speed: 1746 MHz vs 1380 MHz
A newer manufacturing process allows for a more powerful, yet cooler running videocard: 7 nm vs 12 nm
Around 14% higher memory clock speed: 2000 MHz vs 1758 MHz

Launch date	18 November 2018 vs 21 June 2017
Boost clock speed	1746 MHz vs 1380 MHz
Manufacturing process technology	7 nm vs 12 nm
Memory clock speed	2000 MHz vs 1758 MHz

Reasons to consider the NVIDIA Tesla V100 PCIe

Around 4% higher core clock speed: 1246 MHz vs 1200 MHz
Around 20% lower typical power consumption: 250 Watt vs 300 Watt

Core clock speed	1246 MHz vs 1200 MHz
Thermal Design Power (TDP)	250 Watt vs 300 Watt

Compare specifications (specs)

	AMD Radeon Instinct MI50	NVIDIA Tesla V100 PCIe
Essentials
Architecture	GCN 5.1	Volta
Code name	Vega 20	GV100
Launch date	18 November 2018	21 June 2017
Place in performance rating	not rated	not rated
Type	Desktop	Desktop
Technical info
Boost clock speed	1746 MHz	1380 MHz
Core clock speed	1200 MHz	1246 MHz
Manufacturing process technology	7 nm	12 nm
Thermal Design Power (TDP)	300 Watt	250 Watt
Transistor count	13,230 million	21,100 million
Floating-point performance		14,131 gflops
Pipelines		5120
Texture fill rate		441.6 GTexel / s
Video outputs and ports
Display Connectors	No outputs	No outputs
Compatibility, dimensions and requirements
Interface	PCIe 4.0 x16	PCIe 3.0 x16
Length	267 mm
Supplementary power connectors	1x 6-pin + 1x 8-pin	2x 8-pin
API support
DirectX	12.0
OpenGL	4.6
Memory
Memory clock speed	2000 MHz	1758 MHz
Maximum RAM amount		16 GB
Memory bandwidth		900.1 GB / s
Memory bus width		4096 Bit
Memory type		HBM2
Technologies
CUDA