NVIDIA GH200 super server
Product description
NVIDIA GH200 super server is a high-performance AI superserver powered by NVIDIA GH200 Tensor Core Gpus and NVIDIA NVLink technology that delivers powerful computing performance for AI workloads.
Product parameter
GPU: NVIDIA GH200 Tensor Core GPU
CUDA core: 12,480
Memory: 120GB HBM3
Video memory bandwidth: 2.2TB/s
Power consumption: 600W
NVLink: Four NVLink connections
Product specification
Dimensions: 762 x 533 x 180mm
Weight: 12kg
Port: PCIe Gen5 x16
Product application
The NVIDIA GH200 super server can be used for a variety of AI workloads, including:
Image classification
Natural language processing
Machine learning
Deep learning
Usage method
The NVIDIA GH200 super server can be used to build high-performance AI workstations or servers. It can be used to run a variety of AI-accelerated applications, including:
Artificial intelligence training and reasoning
Machine vision
Natural language processing
Data analysis
Brand introduction
NVIDIA is the world’s leading graphics processor and artificial intelligence company, providing high-performance computing solutions for gaming, professional applications and data centers.
Product advantage
The NVIDIA GH200 super server offers the following benefits:
Powerful computing performance
Efficient AI acceleration
Multiple application support
High reliability
manufacturer
The NVIDIA GH200 super server is manufactured by NVIDIA Corporation.
Detailed description
The NVIDIA GH200 super server features NVIDIA GH200 Tensor Core Gpus, which feature the new Hopper Tensor Core and RT Core to deliver powerful computing performance for AI workloads.
The NVIDIA GH200 super server is also equipped with 120GB of HBM3 video memory, providing up to 2.2TB/s video bandwidth to meet the needs of multiple AI workloads.
The NVIDIA GH200 super server uses four NVLink connections to connect multiple GH200 cards together to form a more powerful computing cluster. NVLink connections provide up to 200GB/s of bandwidth and can be used for multi-card parallel computing to improve the performance of AI workloads.
Overall, NVIDIA GH200 super server is a powerful, feature-rich AI super server for a wide variety of AI workloads.
Force Computers SYS68K CPU-40B/4-01 VME MVME
IRONICS VMEbus IV-3201-1MB VME CONTROL BOARD
MOTOROLA SMM 167-34B CPU PROCESSOR MODULE MVME VME
1 FUEL FILTER 3576059, VME, VOLVO
PHONOGRAPH NEEDLE RCA RP-229A VLP VME VMP VMT VPE VPT
VME-MMCPU Vintage 68000 CPU Card 450-01172-01 Rev A
MOTOROLA MVME162-522A VME BOARD 64-W4125C01A, REV AAA
Zilog CCU 09-0201-02 Z80 Z80A CPU Card 1983 Vintage VME
Vintage Motorola 3 Port Serial VMEmodule VME Bus
Vero VME 48680 Disk Module
Mizar 8300-0-01 VME 4-CH Serial I/O Board
Micro Memory Inc MM-6326 VME Board
MOTOROLA MVME (01-W3584B) VME MODULE 705-1 BOARD;)
Micro Memory MM-6490R/1G Dual-Port VME VME64 RACEway
Motorola MVME 204-2F VME Memory Module MVME-204-2F
Battery for CANON E ES H J L UC UCS VME EX1 EQ305 new
Sun3 Sun4 VME SMD Controller Xylogics 7053 501-1249
Force SYS68K CPU-33B/4 VME Board
battery for CANON E, ES, H, J, L, UC, UCS, VME Series
IRONICS VMEbus IV-3201A-1MB VME CONTROL
Motorola VME Card Memory 1985 1MB + 1MB Mezz VMEmodule
FSI Polaris Whedco VME-3550 Servo Module VMEbus
Idacom HP ITO Card VME 68000 Main CPU 3 5063-3508 68440
Vintage Rockwell Intl CMC VME Bus Ethernet Card 320057
GE FANUC VME-6450 USPP VME6450
Vintage Formation VME Bus Token Ring Network Card
Vintage Interphase VME Bus Fiber Optic Network Card
GE FANUC VME-6450 USPP VME6450
VME SBC CPU Card 6809 Motorola 1052-170-02D Biometrie
MODCOMP IEEE488 National Instruments NI GPIB-1014 VME
Motorola VME Card MVME131 SBC 68020 CPu M68KVMMB-851
Force SPARC/CPU-10/128-1251 SPARC/CPU-10/128-1251H1 VME
Ultratech Stepper VME Combo PCB Board 03-20-00870
TANBAC VME SBC card TB3240 68040 CPU MC68EC040RC33
188 80188 SBC VME XRE 1990 Intel Z8530 CPU Card
Battery for CANON E, ES, H, J, L, UC, UCS, VME Series
