tencent cloud

Hyper Computing Cluster

Release Notes
Product Introduction
Overview
Strengths
Scenarios
Instance Specifications
Billing
Billing Overview
Instance Regions
Purchasing Hyper Computing Cluster Instances
Getting Started
User Guide
Managing Hyper Computing Cluster
Installing nvidia-fabricmanager Service on GPU Instance
Installation Instructions of TCCL on GPU Instances
Installing RDMA Millisecond-Level Monitoring Component on GPU Instances
API Document
FAQs
FAQs
RDMA Network Configuration Component RDMA-agent Description
Contact Us
DokumentasiHyper Computing ClusterProduct IntroductionStrengths

Strengths

PDF
Mode fokus
Ukuran font
Terakhir diperbarui: 2024-08-06 14:54:32

Strengths of Hyper Computing Cluster

Super Computing Power

Achieve computing without virtualization loss and maintain server performance based on an elastic high-performance CVM architecture.
Support fully automated production and auto scaling, featuring flexibility and extreme performance.

High-Performance Storage

Support flexible storage capacity expansion with COS (COS) or CFS (CFS) and ensure strong consistency among three replicas.
Offer ultra-fast IO instances for selection in heavy IO load and local caching scenarios when used jointly with local NVMe SSDs.

High-Speed Network Plane

Support VPC and non-blocking RoCEv2 RDMA network for inter-node connection with transmission latency as low as 2 us, widely supporting Ethernet-based applications.
Use intelligent 25G network cards and a VPC network with a PPS of 10 million to help users build cloud-native high-performance computing applications with file storage and other services.

Comparison with Local Supercomputing Clusters

Hyper Computing Cluster has the following advantages over local supercomputing clusters:
Dimension
Feature
Hyper Computing Cluster
Local Supercomputing Cluster
Hardware resource
Computing resource supply
Flexible
Inflexible
Auto scaling of computing resources
Flexible
Inflexible
Hardware iteration and upgrade
Flexible
Inflexible
Storage resource switching
Flexible
Inflexible
Virtualization
Flexibly supported
Partially supported
Hardware resource maintenance
Simple
Complex
Software resource
Operating system type
Flexible
Inflexible
Operating environment configuration
Simple
Complex
Network
VPC
Supported
Not supported
RDMA network
Supported
Supported
Cost
Maintenance cost
Low
High
Purchase cost
Pay-as-you-go
One-time large investment
Availability
Disaster recovery
High
Partially supported
Security
High
Partially supported
Topik sebelumnya: OverviewTopik selanjutnya: Scenarios

Bantuan dan Dukungan

Apakah halaman ini membantu?

masukan