This topic describes the recommended specifications and scenarios of E-HPC.
E-HPC supports all specifications of ECS instances. To maximize the performance of E-HPC clusters, we recommend that you select one of the following specifications based on your scenarios. For more information about the specifications of ECS instances, see Instance families and Best practices for instance type selection.
ECS bare metal instances and Super Computing Cluster (SCC) instances
The following instances have no performance loss in virtualization and can be isolated by VPCs. You can have direct access to hardware resources. SCC instances provide a low-latency network that is based on Remote Direct Memory Access (RDMA). The network meets the parallel computing requirements of E-HPC clusters.
SCC instances
Instance type
Instance specification
Computing configuration
Network configurations
Storage configuration
Scenario
High clock speed + low-latency network
ecs.scchfg6.20xlarge
CPU: Uses 3.1 GHz Intel ® Xeon ® Platinum 8269CY (Cascade Lake)
vCPU: 80
Memory: 384 GiB
GPU: none
Ethernet: 30 Gbit/s
RDMA: 50 Gbit/s
Ultra disk
Enhanced SSD
Standard SSD
This instance type is suitable for scenarios that require high clock speed and parallel computing on multiple clusters. These scenarios include manufacturing simulation, life science, and weather forecast.
ecs.scch5s.16xlarge
CPU: Uses 3.1 GHz Intel ® Xeon ® Platinum 8269CY (Cascade Lake)
vCPU: 64
Memory: 192 GiB
GPU: none
Ethernet: 30 Gbit/s
RDMA: 50 Gbit/s
Ultra disk
Enhanced SSD
Standard SSD
ecs.scch5.16xlarge
CPU: Uses 3.1 GHz Intel ® Xeon ® Platinum 8269CY (Cascade Lake)
vCPU:64
Memory: 192 GiB
GPU: none
Ethernet: 10 Gbit/s
RDMA: 50 Gbit/s
Ultra disk
Standard SSD
This instance type is suitable for scenarios that require high clock speed and parallel computing on multiple clusters. These scenarios include manufacturing simulation, life science, and weather forecast. Enhanced SSDs (ESSDs) are not supported. If you require high-speed storage in local disks, we recommend that you use scchfg6 or scch5s.
General-purpose + low-latency network
ecs.sccg5.24xlarge
CPU: Uses 2.5 GHz Intel ® Xeon ® Gold 6149 (Skylake)
vCPU: 96
Memory: 384 GiB
GPU: none
Ethernet: 10 Gbit/s
RDMA: 50 Gbit/s
Ultra disk
Standard SSD
This instance type is suitable for scenarios that require parallel computing on multiple clusters, such as manufacturing simulations. ESSD is not supported. For high-speed local hard disk storage, we recommend that you use scchfg6 or scch5s.
GPU type + low-latency network
ecs.sccgn6.24xlarge
CPU: Uses 2.5 GHz Intel ® Xeon ® Gold 6149 (Skylake)
vCPU: 96
Memory: 384 GiB
GPU: 8 * Nvidia Tesla V100
Ethernet: 30 Gbit/s
RDMA: 50 Gbit/s
Ultra disk
Enhanced SSD
Standard SSD
This instance type is suitable for scenarios that require parallel computing on one or more clusters. These scenarios include machine learning and large-scale molecular dynamics.
ECS bare metal instance
Instance type
Instance specification
Computing configuration
Network configurations
Storage configuration
Scenario
High clock speed
ecs.ebmhfc6.20xlarge
CPU: Uses 3.1 GHz Intel ® Xeon ® Platinum 8269CY (Cascade Lake)
vCPU:80
Memory: 384 GiB
GPU: none
Ethernet: 30 Gbit/s
Ultra disk
Standard SSD
This instance type is suitable for scenarios that require high-performance computing on one cluster.
High clock speed + large memory
ecs.ebmhfr6.20xlarge
CPU: Uses 3.1 GHz Intel ® Xeon ® Platinum 8269CY (Cascade Lake)
vCPU: 80
Memory: 768 GiB
GPU: none
Ethernet: 30 Gbit/s
Ultra disk
Standard SSD
This instance type is suitable for scenarios that require high-performance computing on one cluster.
General-purpose
ecs.ebmg6.26xlarge
CPU: Uses 2.5 GHz Intel ® Xeon ® Gold 6149 (Skylake)
vCPU:104
Memory: 384 GiB
GPU: none
Ethernet: 30 Gbit/s
Ultra disk
Enhanced SSD
Standard SSD
This instance type is suitable for computing scenarios in life science. These scenarios include bioinformatics, dynamic simulation, and drug development.
General-purpose + large memory
ecs.ebmr6.26xlarge
2.5
vCPU: 104
Memory: 768 GiB
GPU: none
Ethernet: 30 Gbit/s
Ultra disk
Enhanced SSD
Standard SSD
This instance type is suitable for scenarios such as electronic design automation and life science.
GPU Type
ecs.ebmgn6v.24xlarge
CPU: Uses 2.5 GHz Intel ® Xeon ® Gold 6149 (Skylake)
vCPU: 96
Memory: 384 GiB
GPU: 8 * Nvidia Tesla V100
Ethernet: 30 Gbit/s
RDMA: 50 Gbit/s
Ultra disk
Enhanced SSD
Standard SSD
This instance type is suitable for scenarios that require GPU computing, such as machine learning.
Enterprise-class computing instance families based on the x86 architecture and enterprise-class heterogeneous computing instance families
The following instances have performance loss in virtualization. You can select an instance type that applies to your scenario.
Instance type | Instance specification | Computing configuration | Network configurations | Storage configuration | Scenario |
General-purpose | ecs.g6.8xlarge |
| Ethernet: 10 Gbit/s |
| This instance type is suitable for scenarios that require general-purpose, large-scale concurrent processing capabilities on multiple clusters, such as animation rendering and life science. |
High clock speed | ecs.hfg6.16xlarge |
| Ethernet: 25 Gbit/s |
| This instance type is suitable for scenarios that require high clock speed, such as computing tasks in life science. |
Memory optimized instance | ecs.re6.26xlarge |
| Ethernet: 16 Gbit/s |
| This instance type is suitable for scenarios that involve a single cluster with large memory. These scenarios include computing tasks in bioinformatics. |
GPU Type | ecs.gn6i-c24g1.6xlarge |
| Ethernet: 7.5 Gbit/s |
| Supports GPU computing. |
ecs.gn6i-c24g1.12xlarge |
| Ethernet: 15 Gbit/s |
| Supports GPU computing. |