CMP 50HX

GPU Code: TU102

10 GB

GDDR6

1545 MHz

Boost Clock

250 W

TDP

12 nm

Process
Share on Social Media
Share CMP 50HX specifications with others

Technical Specifications

Core Specifications
Architecture: Turing
Process Size: 12 nm
Base Clock: 1350 MHz
Boost Clock: 1545 MHz
Shader Units: 3584
SMs: 56
Tensor Cores: 448
RT Cores: 56
TMUs: 192
ROPs: 80
Memory & Performance
Memory Size: 10 GB
Memory Type: GDDR6
Memory Bandwidth: 560.0 GB/s
Bus Interface: PCIe 1.0 x4
TDP: 250 W
Suggested PSU: 600 W
Cache & Die Information
L1 Cache: 64 KB
L2 Cache: 5.0 MB
Die Size: 754 mm²
Transistors: 18.6 billion
Transistor Density: 24.7 million/mm²

API Support

DirectX
12.2
OpenGL
4.6
Vulkan
1.3
Cuda Compute
7.5 sm_75

Physical Specifications

Board Length: 267 mm
Board Width: 116 mm
Power Connectors: 2x 8-pin

Quick Facts

Release Date: 2021-06-24
Manufacturer: NVIDIA
Generation: Mining GPUs

Performance

Pixel Fill Rate 123.6 GPixel/s
Texture Fill Rate 296.6 GTexel/s
FP32 Performance 11.1 TFLOPS

CUDA Toolkit Compatibility

Cuda Compute
7.5 sm_75
Architecture

Turing

Supported CUDA Toolkits
10.0 10.1 10.2 11.0 11.1 11.2 11.3 11.4 11.5 11.6 11.7 11.8 12.0 12.1 12.2 12.3 12.4 12.5 12.6 12.7 12.8

Turing architecture with 1st gen RT cores and hardware integer operations