NVIDIA A100 PCIe vs NVIDIA CMP 50HX

GPU Specs Compare

G

NVIDIA A100 PCIe

NVIDIA 2020-06
G

NVIDIA CMP 50HX

NVIDIA 2021-06

Graphics Card

Release Date
Jun 2020
Jun 2021
Generation
Tesla
Mining GPUs
Type
AI GPU
Desktop
Bus Interface
PCIe 4.0 x16
PCIe 1.0 x4

Clock Speeds

Base Clock
765 MHz
1350 MHz
Boost Clock
1410 MHz
1545 MHz
Memory Clock
1215 MHz
1750 MHz

Memory

Memory Size
40GB
10GB
Memory Type
HBM2e
GDDR6
Memory Bus
5120bit
320bit
Bandwidth
1555GB/s
560.0GB/s

Render Config

SM Count
108
56
Shading Units
6912
3584
TMUs
432
192
ROPs
160
80
Tensor Cores
432
448
L1 Cache
192 KB (per SM)
64 KB (per SM)
L2 Cache
40 MB
5 MB
RT Cores
-
56

Theoretical Performance

Pixel Rate
225.6 GPixel/s
123.6 GPixel/s
Texture Rate
609.1 GTexel/s
296.6 GTexel/s
FP16 (half)
77.97 TFLOPS
22.15 TFLOPS
FP32 (float)
19.49 TFLOPS
11.07 TFLOPS
FP64 (double)
9.746 TFLOPS
346.1 GFLOPS

Graphics Processor

GPU Name
GA100
TU102
Architecture
Ampere
Turing
Foundry
TSMC
TSMC
Process Size
7 nm
12 nm
Transistors
54.2 billion
18.6 billion
Die Size
826 mm²
754 mm²
GPU Variant
-
TU102-100-A1

Board Design

TDP
250W
250W
Suggested PSU
600 W
600 W
Outputs
No outputs
No outputs
Power Connectors
8-pin EPS
2x 8-pin

Graphics Features

DirectX
N/A
12 Ultimate (12_2)
OpenGL
N/A
4.6
OpenCL
3.0
3.0
Vulkan
N/A
1.3
CUDA
8.0
7.5
Shader Model
N/A
6.8