NVIDIA next-gen H100 Hopper GPU for High Performance Computing pictured

NVIDIA H100, company’s first 4nm GPU

Tomorrow NVIDIA CEO Jensen Huang will announce the new H100 series of products for data-centers. 

The render that leaks just hours ahead of the announcement confirms some key details on NVIDIA’s Hopper GH100 GPU. Such as the fact that GH100 is a large monolithic GPU with six Graphics Processing Clusters. In fact, we can also count a total of 144 Streaming Multiprocessors. As reported earlier today, one of the GPCs is different, because it also carries the render engine and supports graphics.

Furthermore, the chip supports up to 6 stacks of High-Bandwidth-Memory, unfortunately since the full specs are not yet available, we are not sure if all 6 stacks are used.

NVIDIA H100, Source: VideoCardz

NVIDIA will be rolling out a number of products based on GH100 GPU, such an SXM based H100 card for DGX mainboard, a DGX H100 station and even a DGX H100 SuperPod. More importantly, NVIDIA is also announcing PCIe-based H100 model at the same time.

NVIDIA H100 Product Family, Source: VideoCardz

All products will be presented tomorrow by NVIDIA CEO Jensen Huang during a special GTC 2022 opening keynote.

RUMORED NVIDIA Data-Center GPUs Specifications
VideoCardz.com	NVIDIA H100	NVIDIA A100	NVIDIA Tesla V100	NVIDIA Tesla P100
Picture
GPU	GH100	GA100	GV100	GP100
Transistors	80B	54.2B	21.1B	15.3B
Die Size	814 mm²	828 mm²	815 mm²	610 mm²
Architecture	Hopper	Ampere	Volta	Pascal
Fabrication Node	TSMC N4	TSMC N7	12nm FFN	16nm FinFET+
GPU Clusters	132	108	80	56
CUDA Cores	16896/14592*	6912	5120	3584
L2 Cache	50MB	40MB	6MB	4MB
Tensor Cores	528/456*	432	320	–
Memory Bus	5120-bit	5120-bit	4096-bit	4096-bit
Memory Size	80 GB HBM3/HBM2e*	40/80GB HBM2e	16/32 HBM2	16GB HBM2
TDP	700W/350W*	250W/300W/400W	250W/300W/450W	250W/300W
Interface	SXM5/*PCIe Gen5	SXM4/PCIe Gen4	SXM2/PCIe Gen3	SXM/PCIe Gen3
Launch Year	2022	2020	2017	2016