AMD Zhongshan Subor Z+ GPU vs NVIDIA H100 PCIe

GPU · Specs Compare

GPU AMD Zhongshan Subor Z+ GPU AMD · 2018-08

GPU NVIDIA H100 PCIe NVIDIA · 2022-03

Graphics Card

Release Date

Aug 2018

Mar 2022

Generation

Console GPU

Tesla Hopper

Type

Game console

AI GPU

Bus Interface

—

PCIe 5.0 x16

Clock Speeds

Memory Clock

1200 MHz

1593 MHz

Base Clock

—

1095 MHz

Boost Clock

—

1755 MHz

Memory

Memory Size

8GB

80GB

Memory Type

GDDR5

HBM2e

Memory Bus

256bit

5120bit

Bandwidth

153.6GB/s

2039GB/s

Render Config

Compute Units

—

Shading Units

1536

14592

TMUs

456

ROPs

SM Count

—

114

Tensor Cores

—

456

L1 Cache

—

256 KB (per SM)

L2 Cache

—

50 MB

Theoretical Performance

Pixel Rate

41.60 GPixel/s

42.12 GPixel/s

Texture Rate

124.8 GTexel/s

800.3 GTexel/s

FP16 (half)

7.987 TFLOPS

204.9 TFLOPS

FP32 (float)

3.994 TFLOPS

51.22 TFLOPS

FP64 (double)

249.6 GFLOPS

25.61 TFLOPS

Graphics Processor

GPU Name

Fenghuang

GH100

GPU Variant

Zhongshan Subor

—

Architecture

GCN 5.0

Hopper

Foundry

TSMC

Process Size

14 nm

4 nm

Transistors

Unknown

80 billion

Die Size

397 mm²

814 mm²

Board Design

TDP

100W

350W

Outputs

No outputs

Suggested PSU

—

750 W

Power Connectors

—

1x 16-pin

Graphics Features

DirectX

12 (12_1)

N/A

OpenGL

4.6

N/A

OpenCL

2.1

3.0

Vulkan

1.2

N/A

Shader Model

6.2

N/A

CUDA

—

9.0

▲ Specs Compare Data sourced from official specifications and benchmark tests.