- ABB
- GE
- NI
- EPRO
- METSO
- weida
- STS
- VMIC
- Hima
- KEBA
- B&R
- FANUC
- YASKAWA
- B&R
- KOLLMORGEN
- Other
- RELIANCE ELECTRIC
- Westinghouse
- ICS TRIPLEX
- Schneider
- MOORE
- YOKOGAWA
- ACQUISITIONLOGIC
- OKUMA
- SELECTRON
- SYNRAD
- PROSOFT
- Motorola
- Honeywell
- Bently
- Allen-Bradley
- Rockwell Ics Triplex
- Woodward
- Other Parts
- Triconex
- Foxboro
- Emerson
NVIDIA H100 Professional Services
NVIDIA H100
1. Product Overview
- The NVIDIA H100 (Codename: GH100) is the first-generation flagship GPU based on the Hopper architecture, unveiled by NVIDIA at GTC 2022 and mass-produced in 2023. It serves as the core cornerstone for the booming development of large AI models.
- The NVIDIA H100 (Codename: GH100) is the first-generation flagship GPU based on the Hopper architecture, unveiled by NVIDIA at GTC 2022 and mass-produced in 2023. It serves as the core cornerstone for the booming development of large AI models.
2. Core Specifications of Single H100 GPU
2.1 Architecture & Process Technology
- Architecture: Hopper (GH100, Single Die)
- Architecture: Hopper (GH100, Single Die)
2.2 Compute Cores
- Streaming Multiprocessors (SMs): 132
- Streaming Multiprocessors (SMs): 132
2.3 Memory System
- VRAM Capacity: 80GB HBM3
- VRAM Capacity: 80GB HBM3
2.4 Single-GPU Dense Computing Performance
- FP8 (Training): 3.94 PFLOPS
- FP8 (Training): 3.94 PFLOPS
2.5 Interconnection & Interfaces
- NVLink 4: 450GB/s unidirectional bandwidth, 900GB/s bidirectional bandwidth
- NVLink 4: 450GB/s unidirectional bandwidth, 900GB/s bidirectional bandwidth
2.6 Power Consumption & Cooling
- TGP: 700W (SXM version) / 600W (PCIe version)
- TGP: 700W (SXM version) / 600W (PCIe version)
3. DGX H100 (8-GPU Server) Specifications
3.1 Hardware Configuration
- GPU Module: 8 × H100 SXM5 GPUs, with a total HBM3 memory of 640GB
- GPU Module: 8 × H100 SXM5 GPUs, with a total HBM3 memory of 640GB
3.2 Overall System Computing Performance
-
- FP8 Training Performance: 31.5 PFLOPS
- FP16 Computing Performance: 15.7 PFLOPS
- FP64 Computing Performance: 494 TFLOPS
____________________________________________________________
NVIDIA B300
NVIDIA B200
NVIDIA H200
NVIDIA H100
- FP8 Training Performance: 31.5 PFLOPS
- FP16 Computing Performance: 15.7 PFLOPS
- FP64 Computing Performance: 494 TFLOPS
NVIDIA B300
NVIDIA B200
NVIDIA H200
NVIDIA H100
------------------------------------------
Many products have not been listed yet. Please contact us for more products
📧 sales5@xrjdcs.com | 📞 +86 18250705533（WhatsApp）丨Sales Manager： Jinny
Main brand
ABB 丨 GE 丨Allen Bradley 丨 Honeywell 丨 Emerson 丨 Bently Nevada 丨 Westinghouse
Triconex 丨Foxboro 丨 ICS Triplex 丨 Hima 丨 Schneider 丨 Yokogawa 丨 Woodward & more
⇒The following are our main activities
Taken the ABB (AC800M/AC800F/AC800PEC/AC900F/excitation control)
★GE VMIVME/ VMIC series CPU mainboard
Relay protection SR469/SR369/SR489/SR750
IS200/IS215/IS220/ IS420/ DS200/DS215
/IC693/IC694/IC695/IC697/IC698
★ Invensys (3008, 3625, 3508E.3721, 3700A 4351B Safety system module)
★ Bentley (3500/42M, 3500/22M, 3500/95.3500/05)
★ Woodward 9907 series 505 governor, 8840 series controller, module
★ Foxboro (FBM230, FBM232, FCP270, H90/H92)
★ Emerson Westinghouse module 1C311/1C312, DELTAV module EPRO sensor
★ Dark Horse (F3330, F3236, F6217, F7126, F8621A)
★ Schneider Modicon 140 series CPU, communication module ELAU servo
★ Prosofte MVI46/56/69, AB, communication module
description2