Skip to main content
Contact Us
Advantech

Advantech SKY-TESL-H100-80P NVIDIA H100 PCIe 80GB HBM2e Data Centre GPU Card with Up to Seven MIG Instances

The SKY-TESL-H100-80P is an NVIDIA H100 PCIe data centre GPU card on the Hopper architecture, with 80 GB of HBM2e at 2 TB/s, 48 TFLOPS FP32, 24 TFLOPS FP64 and up to seven MIG instances, in a passive 350 W dual-slot card. It suits generative AI and HPC in air-cooled rack servers with a tighter power budget.

  • Up to Seven MIG Instances: Second-generation Multi-Instance GPU partitions the card into up to seven isolated instances, so several users or services share one GPU.
  • 80 GB HBM2e at 2 TB/s: High-bandwidth HBM2e memory with ECC keeps generative AI models and HPC datasets resident on the card.
  • Hopper Compute with Fast FP64: 14,592 CUDA cores and 456 Tensor Cores, rated at 48 TFLOPS FP32 and 24 TFLOPS FP64 for double-precision HPC work.
  • 350 W Power Draw: Up to 350 W through a 16-pin PCIe connector, which suits air-cooled racks with a tighter power budget per GPU.
  • NVLink and PCIe 5.0: A PCIe 5.0 x16 interface, plus three NVLink bridges that link two cards at 600 GB/s using optional dual or three-slot bridges.
Form Factor: PCI ExpressPCI Express
USA Support Team: +1 407 214 9446 Call
Select section
  • Overview
  • Specifications
  • Downloads

SKY-TESL-H100-80P Overview

Not every AI or HPC rack can feed a 600 W card in every slot. The SKY-TESL-H100-80P is the NVIDIA H100 PCIe from Advantech, a compute-optimised data centre GPU on the Hopper architecture with 80 GB of HBM2e and a maximum draw of 350 W. It is a passive, dual-slot card for lower-power, air-cooled enterprise rack designs, and suits generative AI and high-performance computing (HPC) workloads.

The GPU has 14,592 CUDA cores and 456 fourth-generation Tensor Cores, rated at 48 TFLOPS FP32 and 24 TFLOPS FP64, with fast FP64 for double-precision codes. Its 80 GB of HBM2e with ECC runs at 2 TB/s. Second-generation Multi-Instance GPU can divide the SKY-TESL-H100-80P into up to seven instances with partitioned hardware and guaranteed quality of service, so one card can serve several teams or inference services. Seven JPEG decoders and seven video decoders handle image and video ingest.

The card connects over PCI Express 5.0 x16 and takes power through a 16-pin PCIe connector. It is headless and full height, with a passive heatsink that relies on the server’s airflow. Two cards can be linked with three NVLink bridges at 600 GB/s, using the optional SKY-QUAD-NVL-2A for dual-slot spacing or SKY-QUAD-NVL-3A for three-slot spacing. NVIDIA’s NGC catalog supplies software, libraries and optimized AI models.

We supply the SKY-TESL-H100-80P tested, and fit, configure and burn-in test it in the GPU servers we build, with airflow, power and any NVLink bridges checked before delivery, and we can help size the card count to the power and cooling available. The SKY-TESL-H100N-94P is the H100 NVL with 94 GB of HBM3 at 400 W, and the SKY-TESL-H200N-141 moves to 141 GB of HBM3e for larger models.

SKY-TESL-H100-80P Specifications

Overview
Brand & ModelAdvantech SKY-TESL-H100-80P
Performance
GPU / AI AcceleratorDiscrete GPU (PCIe)
Supported GPUsH100 PCIe
AI Performance3026
GPU FamilyNVIDIA Data Center
GPU ArchitectureNVIDIA Hopper
PCIe Gen SupportPCIe 5.0
Mechanical Properties
Dimensions267mm x 112mm (10.50" x 4.40")

Official Advantech USA Distributor

Things Embedded is an official Advantech USA distributor, and we supply the SKY-TESL-H100-80P tested, and fit, configure and burn-in test it in the GPU computers and servers we build. Within the Advantech range, the SKY-TESL-H100N-94P is the H100 NVL with 94 GB of HBM3 and 60 TFLOPS FP32 at 400 W.

Across our range, the SKY-TESL-H100-80P sits with our data centre GPUs, NVIDIA GPU range and edge AI servers.

Shop Advantech

Consult. Design. Integrate.

Things Embedded is a leading USA supplier of industrial computing and edge AI systems. We combine technical expertise with a collaborative approach to deliver optimal solutions for your application. Whether it’s an off-the-shelf platform or a custom design, we stay at the forefront of the latest technologies to ensure your success at the edge. Our team works closely with you from concept through deployment, providing guidance, engineering support, and integration expertise every step of the way.

Connect with our solutions experts today to review your requirements and find the ideal solution fit for your application.

Services
Send Inquiry