HPC, AI, and cloud computing: When the intelligent era trident is perfectly integrated in Amazon Cloud Technology

In the Greek mythology, the gods of the sea god and the god of harvest Hatong held the trident to escort the ship's protection and bring clear spring to the farmers. Since then, Tridents have often been used to compare three things to form a close combination to form a joint force. For example, many teams in the football world have classic front line tridents.

In the wave of enterprise cloud and industry intelligence, there are also trigaites in cutting -edge technical trends: business and organization clouds are becoming the digital development option of most enterprises; computing power has gradually become the strategic resource of the enterprise, and the cloud The almost infinite cluster computing power allows more and more industry and scene innovation to rely on the cloud high -performance calculation to complete; AI is changing the production method of thousands of industries, becoming a pioneer force, machine learning and deep learning for scientific research and industry exploration, machine learning and industry exploration It also brings the outbreak of artificial intelligence's special computing power.

In the impression of many people, HPC high -performance computing, AI, and cloud services are still in the stage of independence and parallel development. Especially in the cloud, high -performance calculations seem to be too avant -garde.

However, from the logical perspective of industrial efficiency, the high -performance calculation of high -quality and efficient AI training and deployment in the cloud is actually the general trend of intensive industrial intensive and social low -cost innovation. Only by integrating these three technologies can we cast digital tridents needed in the intelligent era.

Amazon Cloud Technology has some answers on how to melt this era trident.

Towards the cloud: high -performance calculation industrial trends and challenges

Is the relationship between cloud computing and high -performance computing, can it really be incompatible? The answer may not be the case.

According to the data of the Hyperion Research market, 18.8%of HPCs will run in the cloud by the end of 2022, and this data will be 12.3%in 2021. Although most HPC tasks still depend on supercomputing centers and local hardware, they are Getting high -performance calculations in the cloud can be said to be the general trend of industrial development. Getting high -performance calculations in the cloud, customers will worry about encountering some challenges at this stage. For example: management challenges, large -scale computing clusters are difficult to create and manage, can there be a fast deployment method and efficient and convenient management methods; energy efficiency challenges, or the cost -effective consideration of high -performance computing on the cloud, how to play HPC in the cloud in the cloud The maximum energy efficiency is a topic that many users are worried about; security challenges, a large number of HPC processing tasks and data are inseparable. Data will inevitably have data security concerns. How to deliver the data security in the cloud to the user's assured environment.

However, from the perspective of the development trend of the high -performance computing industry, these problems can be solved one by one in practice. From the basic computing logic, the cloud acquisition of high -performance computing is more economical, and users can flexibly obtain heterogeneous computing resources to truly realize the adaptation of computing and tasks. From the perspective of a single node performance, the computing resource performance of the cloud is better; in the computing cluster scenario, the cloud can allow users to obtain linear growth computing performance and avoid waste waste.

Therefore, it is not impossible to achieve high -performance calculations in the cloud. Instead, because of the huge amount of computing power of the cloud, the computing performance of the node is constantly enhanced, and the convenient and efficient computing power management means, as well as the safety guarantee of the cloud -native system and the data of data , Make high -performance calculations in many industries run in the cloud.

In the exploration of how to achieve the cloud acquisition of reliable HPC, Amazon Cloud Technology has achieved industry -leading exploration.

Technology integration and industrial balance: Amazon Cloud Technology's high -performance computing exploration

At the current stage, Amazon Cloud Technology can already provide a highly customized HPC computing platform to bring users a variety of heterogeneous computing resources and customized computing examples. It is particularly noteworthy that Amazon Cloud Technology, known for its rich software ecology, also provides a large number of software that can be available and low -cost in the HPC field to help users solve problems such as management and dispatching.

Overall, the HPC exploration of Amazon Cloud Technology shows two core differences: the technical experience of chips, clouds, storage, software, AI and other fields, as well as the industry needs and user pain points. Software and hardware ecology.

Amazon Cloud Technology provides customers with mature HPC -related service guarantees in the computing, network, storage, and application software ecological adaptation of high -performance computing customers.

In the computing power layer, Amazon Cloud Technology provides diversified heterogeneous computing support including CPU, GPU, ARM, and customized elastic computing examples to meet users' computing resources in HPC high -income tasks such as AI.

In the storage layer, clustering computing power needs will lead to high and high -storage access, which makes the storage performance very critical. Amazon Cloud Technology provides storage and support for high -performance computing scenes, and can achieve multi -level file storage strategies in the cloud to help users achieve elastic storage and use according to calculation needs, thereby reducing the storage costs of cloud HPCs, increasing data calling data calls , Management efficiency.

In the cloud network, Amazon Cloud Technology can provide customers with the continuous low latency required for super computing applications, high -bandwidth network environment, users can use the up to 100Gbps bandwidth through Amazon Cloud Technology, support the EFA (EFASTIC FABRIC of Elastic Fabric Adapter) network card has launched a low latency and reduced network jitter SRD (Scalable Reliable DataGram) protocol to accelerate communication between nodes. On the software layer, Amazon Cloud Technology faces the needs of migration and dispatch, including visualization and other HPC scenarios, providing rich and low -cost software tools. For example, the use of Amazon Cloud Technology ParallelCluster can be said to achieve the rapid construction of the HPC computing environment and simplify the deployment and management of the HPC cluster. Amazon Cloud Technology Step Functions is a low -code, visual workflow service that can help developers build distributed applications, automate IT and business processes, and build data and machine learning channels, thereby reducing comprehensive development costs. This is very important for high -performance computing tasks in AI and other fields. Rich, professional and low -threshold software ecology allows Amazon Cloud Technology to help high -performance computing users save huge software custom development costs and achieve industrial -grade high -performance computing applications.

Based on the high -performance computing exploration of Amazon Cloud Technology's diversification, the surging computing power of clustering in the cloud has become possible. The direct impact of such a possibility is to lay the foundation for the large -scale AI application tide.

Intelligent morning Xi: The wave of calculation brought by AI navigation

As pre -training large models and AI scientific computing begin to become the mainstream of the industry, the computing power required for AI training and deployment has begun to increase, especially the dependence of AI tasks on high -performance calculations has gradually been enlarged. Perhaps it can be said that the morning light of industrial intelligence is gradually blooming, and it must be based on the solid computing power base of HPC.

AI tasks with complex structures and large data volume of new drug research and development, research and research, and geological exploration have begun to increase, and a series of new demands have been put on HPC. For example, the requirements for computing clusterization are continuously improved, the ability requirements for heterogeneous computing are more stringent, data throughput and throughput efficiency requirements are continuously strengthened. In this era of "AI navigation", if enterprises and scientific research institutions still use the direction of building a hardware calculation pool to achieve HPC, then it is obvious that the industrial efficiency is very low, and the comprehensive cost is greatly waste. It takes a long time to deploy. For high -performance computing tasks with high timeliness requirements, it is obviously unable to meet its needs.

Faced with the computing power requirements brought by machine learning, deep learning, and other AI tasks, Amazon Cloud Technology not only provides computing resources equipped with enterprise GPUs in the cloud, but also independently develops corresponding to the characteristics of machine learning and deep learning. The chip is delivered to customers through the form of cloud services. At the current stage, Amazon Cloud Technology can provide customers with a large -scale computing cluster equipped with 4,000 NVIDIA A100 GPUs, providing 400 GBPS non -blocking network infrastructure, as well as high throughput capabilities achieved through FSX for Lustre, Low delay storage. The computing cluster of such scale is actually difficult to achieve in the physical supercomputing center. In the era of AI, high -performance computing power for machine learning and deep learning from the cloud is obviously the most reasonable solution.

Facing HPC demand for high speeds in the intelligent era. Amazon Cloud Technology has finally integrated AI, HPC, and cloud computing through the accumulation of industrial knowledge and service experience. This trident will continue to evolve to help users go out of the sea in the wave of intelligence and harvest value in digital fields.

At the Global ISC2022 conference in early June this year, Amazon Cloud Technology launched a series of cloud services for high -performance computing, with computing instances specifically for HPC workload HPC6A. After optimization, it can operate the workload of intensive and high -performance computing, such as calculating fluid dynamics, hiding modeling, weather simulation, and finite element analysis. Compared to the comparable Amazon EC2 based on the X86 -based computing and optimized example, the cost -effectiveness provided by the HPC6A instance is up to 65%higher. Using the HPC6A instance, you can greatly reduce the cost of HPC workload, while using the elasticity and scalability of AWS. On the GPU instance, the new instance Amazon EC2 P4DE launches a preview version. This instance can provide the excellent performance required for machine learning (ML) training and high -performance calculation (HPC) applications, such as object detection, semantic segmentation, natural language, natural language Treatment, earthquake analysis, and calculation of fluid dynamics. The ARM -based chip Graviton series that Amazon Cloud Technology has always been committed to development has also released the latest product Graviton3 of the third -generation Graviton processor series this year. Compared with the AWS Graviton2 processor, their computing performance has increased by 25%, the floating -point performance is doubled, and the encrypted workload performance has been doubled. Want to understand the continuous evolution of Amazon Cloud Technology in the field of high -performance computing; if you want to clarify how high -performance calculations are learned with machines, and truly quantum computing is a combination of cutting -edge technology; if you want to insight into time in advance At 13:30 on August 24th, the "Amazon Cloud Technology HPC + Cloud Business Accelerated Innovation Forum" was held at the third -floor conference hall of the Westin Big Hotel in Jinmao.

This event will bring together technical experts from Amazon Cloud Technology and various industries to jointly sort out the development trajectory of computing and intelligence, and reveal the innovation opportunities in the "HPC +" era.

On August 24th, we saw it.

- END -