0% found this document useful (0 votes)
31 views7 pages

G550 Heterogeneous

Uploaded by

Điện Tử NT
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
0% found this document useful (0 votes)
31 views7 pages

G550 Heterogeneous

Uploaded by

Điện Tử NT
Copyright
© © All Rights Reserved
We take content rights seriously. If you suspect this is your content, claim it here.
Available Formats
Download as PDF, TXT or read online on Scribd
You are on page 1/ 7

FCIA Training Material

GPU Server Technical Introduction


G5500 V5
Fusion X, Digital Infinity

Copyright©2021 xFusion Digital Technologies Co., Ltd.


All Rights Reserved.

The information in this document may contain predictive


statements including, without limitation, statements regarding
the future financial and operating results, future product
portfolio, new technology, etc. There are a number of factors that
could cause actual results and developments to differ materially
from those expressed or implied in the predictive statements.
Therefore, such information is provided for reference purpose
only and constitutes neither an offer nor an acceptance. xFusion
may change the information at any time without notice.
G5500 Data Center Heterogeneous Server Overview

Rich and flexible hardware configurations: 1 chassis, 2 node types, and 3 models

GP608 node • 8 x V100/P100/P40/T4 PCIe

G560 V5 full-width dual-socket


G5500 full-width model compute node

GP308 node • 4 x V100 or 8*V100(150 W)


/T4 PCIe

G530 half-width dual-socket GP316 node • 16 x T4 PCIe


G5500 half-width model compute node

2
G5500 Product Specifications
G5500
Node G560 V5 G530 V5
Application High-density AI inference deployment;
G560 V5 front view Large-scale AI training
Scenario HPC
Full-width heterogeneous
Type Half-width heterogeneous compute node
compute node
GPU module
2 half-width GPU boards
1 full-width GPU board
GPU Each board: 16 T4 or 4 P100/V100 (PCIe)
8 P40/P100/V100 (PCIe)
or 8 V100 (single PCIe slot)
CPU module
1 full-width 2S CPU compute
2 half-width 2S CPU compute modules
module
Server Each node: 2 Xeon® Skylake/Cascade
2 Xeon® Skylake/Cascade CPUs,
G530 V5 front view CPUs, 24 DIMMs
24 DIMMs
G5500

Each node: 2 x 2.5-inch SAS/SATA/NVMe


GPU module 2 x 2.5-inch SAS/SATA + 6 x 2.5-
drives
inch NVMe
Drive Or 4 M.2 SATA SSDs
PCIe model: scalable to 8 x 3.5-
Scalable to 4 x 3.5-inch SAS/SATA drives
CPU module inch SAS/SATA drives
(GP308)
RAID RAID 0, 1, 10, 5, 50, 6, or 60
G5500 rear view
IO 4 PCIe x16 LP + 2 x 10GE LOM Each node: 2 PCIe x16 LP + 2 x 10GE LOM

Power 4 hot-swappable 2200 W AC or 240 V HVDC PSUs, with support for the N+N
Supply Units working mode
Fan Modules 6 hot-swappable fan modules, with support for N+1 redundancy
Temperature 5°C to 35°C (41°F to 95°F)
Dimensions
175 mm x 447 mm x 790 mm (6.89 in. x 17.60 in. x 31.10 in.)
3 (H x W x D)
G5500 Front View — Various Local Storage Resources

1
1

2
2
3
3

1 8 x 3.5-inch SAS/SATA HDDs* 1 4 x 3.5-inch SAS/SATA HDDs*

2 6 x 2.5-inch NVMe SSDs or SAS/SATA HDDs 2 2 x 2.5-inch NVMe SSDs or SAS/SATA HDDs (OS
installation)

3 2 x 2.5-inch SAS/SATA HDDs (OS installation) 3 1 VGA + 2 USB ports

*Note: The G530 V5+GP316 model does not support 4 x


3.5-inch drives.

4
G5500 Rear View — Unified Management and Shared
Power and Fan Resources
1 1

2 2
4

3 3

1 Management modules 2 I/O modules 3 PSUs 4 Fan modules

⚫ 2 management modules ⚫ 2 I/O modules ⚫ 4 Platinum PSUs, N+N ⚫ 6 hot-swappable fan modules,
⚫ Each management module ⚫ Each I/O module supports 2 redundancy N+1 redundancy
supports 2 x 10GE optical PCIe x16 slots ⚫ Maximum output power of ⚫ Online maintenance without
ports ⚫ Supports half-height half- each PSU: 2200 W opening the chassis cover
⚫ GE management network length PCIe cards ⚫ Support for AC and 240 V
ports and serial ports HVDC

5
FusionServer G5500 Key Messages

Superior Heterogeneous Computing Performance


• Supports 8 FHFL or 32 HHHL AI training &
inference heterogeneous accelerator cards
• Supports GPUDirect RDMA/P2P

One-Click Switching of Heterogeneous Topologies


• AI and HPC topologies switched at one click
• Support diverse applications and reduce hardware
investment

Fully Modular Design


• Independent CPU module and heterogeneous
modules, supports long-term evolution
• Power supply, hard drive and fan use modular
Artificial Cloud GPU design, simplifying O&M
Intelligence HPC
Acceleration

6
Thank you.

Fusion X, Digital Infinity

Copyright©2021 xFusion Digital Technologies Co., Ltd.


All Rights Reserved.

The information in this document may contain predictive


statements including, without limitation, statements regarding
the future financial and operating results, future product
portfolio, new technology, etc. There are a number of factors that
could cause actual results and developments to differ materially
from those expressed or implied in the predictive statements.
Therefore, such information is provided for reference purpose
only and constitutes neither an offer nor an acceptance. xFusion
may change the information at any time without notice.

You might also like

pFad - Phonifier reborn

Pfad - The Proxy pFad of © 2024 Garber Painting. All rights reserved.

Note: This service is not intended for secure transactions such as banking, social media, email, or purchasing. Use at your own risk. We assume no liability whatsoever for broken pages.


Alternative Proxies:

Alternative Proxy

pFad Proxy

pFad v3 Proxy

pFad v4 Proxy