Industry Trends
HPC Development Trends
To meet the processing demands of high performance computing applications, enterprises are faced with the decision to build local HPC clusters or leverage the scalability, performance, and price of the cloud. In-cloud HPC enables nearly unlimited computing capacity that scales with demand, reduces unnecessary investment and costs, and allows customers to focus on their own business development.
展开详情Business Challenges
-
Asset-Heavy, Time-intensive Builds
High construction costs and long rollouts that still often fail to meet requirements for service flexibility.
-
Heavy Maintenance Workloads
Heavy maintenance workloads consume manpower and time.
-
Rapid Hardware Iterations
New technology emerges and evolves so quickly that enterprises can barely keep up with computing requirements.
-
High OPEX
High operational costs including rent, electricity, water, and system management.
Typical Scenarios
-
Industrial Design
-
Biomedicine
-
Energy Exploration
-
Graphic Rendering and Heterogeneous Computing
Industrial Design
Industrial and manufacturing enterprises typically use design simulation software, including CAD and CAE, to improve efficiency. Huawei HPC Cloud provides powerful computing services for a wide range of simulation software, helping enterprises shorten service rollouts and improve product quality.
High-Performance ECS
Provides H1 (high computing I, which uses a 10GE network), H2 (ultra-high computing II, which uses a 100 Gbit/s InfiniBand network), and M2 (memory-optimized type) ECSs. H2 ECSs use local NVMe PCIe SSDs, offering up to 3.2 TB capacity, 300,000 IOPS, and 3,000 Mbit/s bandwidth storage capacity.
Excellent-Performance BMS
Provides excellent computing performance using dedicated physical servers, avoiding loss due to virtualization and offering high performance. It allows you to automatically provision BMSs on the management console, fully meeting scalability and flexibility requirements for HPC services.
High-Performance Lustre Parallel File System
Uses optimized and dedicated ECSs and EVS disks to provide up to 4.5 Gbit/s throughput for each OSS and up to 2 Tbit/s throughput in total. Each file supports up to 32 PB of capacity. Lustre file system can fully meet the file system requirements of HPC.
Distributed and Scalable Block Storage Service
Allows EVS disks to be attached to each ECS on demand and offers storage space for HPC computing and storage nodes. Each disk provides up to 32 TB of capacity, 20,000 IOPS, and 350 Mbit/s throughput.
Biomedicine
Biomedicine relies heavily on high performance computing for molecular modeling, data management, medicine molecular design, and next-generation gene sequencing. Huawei HPC Cloud provides biopharmaceutical companies with industry-leading computing and storage services.
Scalable and High-Performance ECS
Provides various ECSs and BMSs to meet different HPC application requirements for computing and memory capacity. ECS types include H1 (high computing I, which uses a 10GE network), H2 (ultra-high computing II, which uses a 100 Gbit/s InfiniBand network), P1 (Artificial Intelligence), and M1 (memory-optimized type).
Distributed and Scalable Block Storage Service
Allows EVS disks to be attached to each ECS on demand and offers storage space for HPC computing and storage nodes. Each disk provides up to 32 TB of capacity, 20,000 IOPS, and 350 Mbit/s throughput.
Excellent VPC Network
Builds a secure, isolated virtual network environment on HUAWEI CLOUD for HPC users. HPC computing networks connect to each other with 10GE Ethernet or Direct Connect, providing the high bandwidth for the users on your network.
Massive OBS
Provides a huge amount of storage space for HPC applications. You can flexibly upload and download HPC data.
Related Services
Energy Exploration
Energy exploration is a computing-intensive industry. Huawei HPC Cloud provides powerful computing services for data collecting, seismic data processing, and oil reservoir modeling to deliver unprecedented levels of accuracy and productivity.
Flexible High-Performance ECS
ECSs are used to deploy HPC head and work nodes. D1 ECSs provide optimized data processing capabilities, and attach enforced SAS and SSD disks as local storage (supporting up to 24 local disks). D1 ECSs provide HPC applications with high I/O speed and rapid data processing capabilities to meet requirements for data-intensive computing applications. P1 ECSs provide superb dual-precision computing power of up to 36.8 TeraFLOPS and support local NVMe SSDs. P1 ECSs can visualize and analyze PB-level well location data in milliseconds. H2 ECS computing clusters support RDMA and a wide range of mainstream MPI libraries.
Excellent-Performance BMS
Provides excellent computing performance using dedicated physical servers, avoiding loss due to virtualization and offering high performance. It allows you to automatically provision BMSs on the management console, fully meeting scalability and flexibility requirements for HPC services.
High-Performance Lustre Parallel File System
Uses optimized and dedicated ECSs and EVS disks to provide up to 4.5 Gbit/s throughput for each OSS and up to 2 Tbit/s throughput in total. Each file supports up to 32 PB of capacity. Lustre file system can fully meet the file system requirements of HPC.
Distributed and Scalable Block Storage Service
Allows EVS disks to be attached to each ECS on demand and offers storage space for HPC computing and storage nodes. Each disk provides up to 32 TB of capacity, 20,000 IOPS, and 350 Mbit/s throughput.
Related Services
Graphic Rendering and Heterogeneous Computing
High performance computing nodes need to use GPU heterogeneous computing to accelerate computing for various HPC applications including AI, deep learning, neural networking, and 3D rendering. Huawei HPC Cloud greatly improves application program execution efficiency.
GPU-accelerated Cloud Server
Huawei GPU-accelerated cloud server is developed using NVIDIA Tesla P100 GPU accelerators. A single GPU uses 3,584 stream processors, with 9.3 TeraFLOPS of single-precision computing power and 4.7 TeraFLOPS of double-precision computing power. Huawei GPU-accelerated cloud server uses 16 GB HBM2 GPU memory and 732 Gbit/s bandwidth, GPU Direct and NvLink technologies, and PCI bus to increase data transmission efficiency between GPUs by 5X.
FPGA-accelerated Cloud Server
Provides tools and environment for you to develop and use FPGAs. With the ECSs, you can easily develop FPGA accelerators, deploy FPGA-based services, and provide simple-to-use, cost-effective, agile, and secure FPGA cloud services.
Excellent-Performance BMS
Provides excellent computing performance using dedicated physical servers. Each BMS uses up to eight Tesla P100 GPU accelerators to support parallel GPU acceleration and attaches SSD disks to provide storage acceleration, providing GPU acceleration. You can automatically provision BMSs on the management console to meet the flexibility and scalability requirements for such services as AI and deep learning.
Agile Graphic Workstation
Provides desktop cloud services. Each desktop cloud can attach a Tesla M60 GPU accelerator, providing you with excellent graphic rendering performance and flexible, simple-to-use desktop cloud services.
Related Services
Solution Architectures
Description
Provides high-performance, reliable, secure, and simple-to-use simulation services, fully satisfying the requirements for parts manufacturing, shortening service rollouts, and boosting enterprise revenue.
Highlights
- Quick deployment: one-click deployment and on-demand auto scaling capabilities.
- Excellent performance: 100 Gbit/s InfiniBand computing network, 3.2 TB enterprise-level SSD disks (the largest local cache disks in cloud services).
- Rich instances: Supports deployment of ECSs and BMSs to suit the specific requirements for your application mash-up.
- All-around protection: 7-layer security protection, Layer 3 Anti-DDoS, VPC, and IAM.
Applicable Products & Services
Solution Advantages
Excellent Performance
Industry-leading performance is achieved using a 100 Gbit/s InfiniBand network and 3.2 TB enterprise-level SSDs (local high-speed cache disks in cloud services).
All-Round Security Protection
End-to-end security solutions protect user data from leakage and ensure a reliable high-performance computing platform.
Strong Scalability
Full OpenStack architecture provides openness and heterogeneous capabilities while avoiding vendor lock-in.
Extensive Ecosystem
Integration with advanced HPC applications of other enterprises including Dassault, Altair, and ESI Group and comprehensive evaluation of system performance ensure that HPC Cloud services run reliably.
Recommended Configurations
Industrial Design
Industrial Design
Biomedicine
Energy Exploration
Heterogeneous Computing
Industrial Design
Pre- and post-simulation scenarios
-
ServiceSpecificationsQuantity
-
ECS
G1 | 4 vCPUs 8 GB | 80 GB high I/O system disk
2
-
EVS
100 GB high I/O data disk
N/A
-
VPC
10 Mbit/s bandwidth
N/A
-
AS
Free
N/A
-
IMS
Free
N/A
-
Cloud Eye
Free
N/A
Industrial Design
Calculation scenarios
-
ServiceSpecificationsQuantity
-
ECS
H2 | 16 vCPUs 128 GB | 3.2 TB SSD | 100 Gbit/s InfiniBand
10
-
SFS
4 TB capacity
N/A
-
VPC
10 Mbit/s bandwidth
N/A
-
AS
Free
N/A
-
IMS
Free
N/A
-
Cloud Eye
Free
N/A
Biomedicine
Gene conversion
-
ServiceSpecificationsQuantity
-
ECS
H2 | 16 vCPUs 128 GB | 3.2 TB SSD | 100 Gbit/s InfiniBand
10
-
SFS
4 TB capacity
N/A
-
VPC
10 Mbit/s bandwidth
N/A
-
AS
Free
N/A
-
IMS
Free
N/A
-
Cloud Eye
Free
N/A
Energy Exploration
-
ServiceSpecificationsQuantity
-
ECS
H2 | 16 vCPUs 256 GB | 3.2 TB SSD | 100 Gbit/s InfiniBand
10
-
SFS
4 TB capacity
N/A
-
VPC
10 Mbit/s bandwidth
N/A
-
AS
Free
N/A
-
IMS
Free
N/A
-
Cloud Eye
Free
N/A
Heterogeneous Computing
-
ServiceSpecificationsQuantity
-
ECS
P1 | 8 x P100 | 100 Gbit/s InfiniBand
2
-
SFS
4 TB capacity
N/A
-
VPC
10 Mbit/s bandwidth
N/A
-
AS
Free
N/A
-
IMS
Free
N/A
-
Cloud Eye
Free
N/A
