Report an Error

NVIDIA H100 CNX

Graphics Processor
GH100
Cores
14592
TMUs
456
ROPs
24
Memory Size
80 GB
Memory Type
HBM2e
Bus Width
5120 bit
GPU Chip
GPU
The H100 CNX is a professional graphics card by NVIDIA, launched on March 21st, 2023. Built on the 5 nm process, and based on the GH100 graphics processor, the card does not support DirectX. Since H100 CNX does not support DirectX 11 or DirectX 12, it might not be able to run all the latest games. The GH100 graphics processor is a large chip with a die area of 814 mm² and 80,000 million transistors. Unlike the fully unlocked H100 SXM5 80 GB, which uses the same GPU but has all 16896 shaders enabled, NVIDIA has disabled some shading units on the H100 CNX to reach the product's target shader count. It features 14592 shading units, 456 texture mapping units, and 24 ROPs. Also included are 456 tensor cores which help improve the speed of machine learning applications. NVIDIA has paired 80 GB HBM2e memory with the H100 CNX, which are connected using a 5120-bit memory interface. The GPU is operating at a frequency of 690 MHz, which can be boosted up to 1845 MHz, memory is running at 1593 MHz.
Being a dual-slot card, the NVIDIA H100 CNX draws power from an 8-pin EPS power connector, with power draw rated at 350 W maximum. This device has no display connectivity, as it is not designed to have monitors connected to it. H100 CNX is connected to the rest of the system using a PCI-Express 5.0 x16 interface. The card measures 267 mm in length, 111 mm in width, and features a dual-slot cooling solution.

Graphics Processor

GPU Name
GH100
Architecture
Hopper
Foundry
TSMC
Process Size
5 nm
Transistors
80,000 million
Density
98.3M / mm²
Die Size
814 mm²

Graphics Card

Release Date
Mar 21st, 2023
Generation
Tesla Hopper
(Hxx)
Predecessor
Tesla Ada
Successor
Tesla Blackwell
Production
Active
Bus Interface
PCIe 5.0 x16

Clock Speeds

Base Clock
690 MHz
Boost Clock
1845 MHz
Memory Clock
1593 MHz
3.2 Gbps effective

Memory

Memory Size
80 GB
Memory Type
HBM2e
Memory Bus
5120 bit
Bandwidth
2.04 TB/s

Render Config

Shading Units
14592
TMUs
456
ROPs
24
SM Count
114
Tensor Cores
456
L1 Cache
256 KB (per SM)
L2 Cache
50 MB

Theoretical Performance

Pixel Rate
44.28 GPixel/s
Texture Rate
841.3 GTexel/s
FP16 (half)
215.4 TFLOPS (4:1)
FP32 (float)
53.84 TFLOPS
FP64 (double)
26.92 TFLOPS (1:2)

Board Design

Slot Width
Dual-slot
Length
267 mm
10.5 inches
Width
111 mm
4.4 inches
TDP
350 W
Suggested PSU
750 W
Outputs
No outputs
Power Connectors
8-pin EPS
Board Number
P1010

Graphics Features

DirectX
N/A
OpenGL
N/A
OpenCL
3.0
Vulkan
N/A
CUDA
9.0
Shader Model
N/A

GH100 GPU Notes

Tensor Cores: 4th Gen
NVENC: No Support
NVDEC: No Support
PureVideo HD: VP11
VDPAU: Feature Set K
May 7th, 2024 07:50 EDT change timezone

New Forum Posts

Popular Reviews

Controversial News Posts