Amazon cloud technology launched three new Amazon EC2 instances supported by self-developed chips

The following is the Amazon cloud technology launched three new Amazon EC2 instances supported by self-developed chips recommended by recordtrend.com. And this article belongs to the classification: IaaS, PaaS, SAAS, cloud computing.

The Amazon EC2 c7g instance is supported by the New Amazon graviton3 processor. Compared with the current generation c6g instance supported by Amazon graviton2 processor, the performance of computing intensive workload is improved by up to 25%

Amazon EC2 trn1 instance is supported by Amazon training chip. Amazon EC2 provides the best cost performance and the fastest training speed for most machine learning model training

The Amazon EC2 im4gn / is4gen / i4i instance is equipped with new Amazon nitro SSDs to provide the best storage performance for I / O-Intensive workloads

Beijing – on December 1, 2021, at the Amazon cloud technology re: invent global conference, Amazon cloud technology announced the launch of three new Amazon Elastic Compute Cloud (Amazon EC2) instances supported by self-developed chips to help customers significantly improve the performance, cost and energy efficiency of work load running on Amazon EC2. The new c7g instance is supported by Amazon graviton3 processor. Compared with the current generation c6g instance supported by Amazon graviton2 processor, the performance is improved by 25%. The new trn1 instance supported by Amazon training chip provides the best cost performance and the fastest training speed for training most machine learning models in Amazon EC2. New storage optimized im4gn / is4gen based on self-developed Amazon nitro SSDs (solid state drives)/   The i4i instance provides the best storage performance for I / O intensive workloads running on Amazon EC2. The release of these new Amazon EC2 instances based on Amazon cloud technology self-developed chips will help customers support their key business applications.

David Brown, vice president of Amazon EC2, said: “Our continuous investment in upgrading self-developed chips has enabled customers to gain huge cost performance advantages in some of today’s key workloads. Customers hope that we will continue to break through the boundaries on each new generation of EC2 instances. The continuous innovation of Amazon cloud technology gives customers the opportunity to use these new and game changing instances to run their important workloads and obtain more benefits Good value for money. “

C7g instances are supported by the New Amazon graviton3 processor, which improves the performance by up to 25% compared with the current generation c6g instances supported by Amazon graviton2 processor

Since its launch in 2020, the computing instance based on Amazon graviton2 has been used in production by many customers, such as DirecTV, discovery, Epic Games, formula 1, honeywell.io, intuit, LYFT, mercardolibre, nextroll, Nielsen, SmugMug, snap, Splunk and sprinklr, and has achieved significant performance improvement and cost savings. There are 12 examples based on graviton2, including general-purpose, computing optimized, memory optimized, storage optimized, burst performance and accelerated computing, so that customers can have deep and wide computing options in the cloud, taking into account cost performance and energy efficiency. As customers carry out more computing intensive workloads in the cloud, such as high performance computing (HPC), games and machine learning reasoning, the corresponding requirements for computing, storage, memory and network also grow. Customers need to seek better cost performance and energy efficiency to run these workloads.

C7g instances supported by Amazon graviton3 processor can improve the performance of computing intensive workloads by up to 25% compared with the current generation c6g instances supported by graviton2 processor. Compared with graviton2, Amazon graviton3 processor provides up to twice the floating-point performance for scientific computing, machine learning and media coding workloads, up to twice the speed of encryption workload, and up to three times the performance for machine learning workload. Amazon graviton3 processor is also more energy-efficient. Under the same performance, it can save up to 60% of energy consumption compared with EC2 instances of the same type. C7g instance is the first instance in the cloud to use the latest ddr5 memory. Compared with the instance based on Amazon graviton2, it improves the memory bandwidth by 50%, thus improving the performance of memory intensive applications such as scientific computing. Compared with the Amazon graviton2 based instance, the network bandwidth of the c7g instance is also 20% higher. The c7g instance supports elastic fabric adapter (EFA), which allows applications to communicate directly with the network interface card, provides lower and more consistent latency, and improves the performance of applications requiring large-scale parallel processing (such as HPC and video coding). The c7g instance is now available in preview. For more information on c7g instances, visit aws.amazon.com/ec2/instance-types/c7g.

The trn1 instance supported by Amazon training chip provides the best cost performance and the fastest training speed for training most machine learning models in Amazon EC2

More and more customers are building, training and deploying machine learning models to support applications that can reshape their business and customer experience. In order to improve the accuracy, these machine learning models must be built on more and more training data, resulting in higher and higher training cost. This may limit the number of machine learning models that customers can deploy. Amazon cloud technology provides a wide range of computing service choices for machine learning, including EC2 p4d instances using NVIDIA A100 tensor core GPU and Habana labs   EC2 DL1 instance of Gaudi accelerator. However, even with today’s fastest acceleration examples, training a continuously growing machine learning model is still very expensive and time-consuming.

Trn1 instance supported by Amazon training chip provides the best cost performance and the fastest training speed for in-depth learning model training in Amazon EC2. Compared with p4d instance, the cost of training in-depth learning model through trn1 instance is reduced by up to 40%. Trn1 instance provides 800gbps EFA network bandwidth (twice as high as the latest GPU based EC2 instance) and integrates with Amazon FSX for lustre high-performance storage, so that customers can start trn1 instance with EC2 ultraclusters function. With EC2 ultraclusters, developers can extend machine learning training to more than 10000 training accelerators interconnected with Pb networks, allowing customers to access supercomputing performance on demand. Even for the largest and most complex models, the training time can be reduced from a few months to a few days. The trn1 instance is now available in preview. For more information on trn1 instances, please visit aws.amazon.com/ec2/instance-types/trn1.

Im4gn / is4gen with new Amazon nitro SSDs/   I4i instances provide optimal storage performance for I / O-Intensive workloads

Today, customers use I3 / i3en storage optimized instances for applications that need direct access to local storage datasets, such as horizontally expanded transactional and relational databases (such as MYSOL and PostgreSQL), NoSQL databases (such as Cassandra, mongodb, redis, etc.), big data (such as Hadoop) and data analysis workloads (such as spark, hive, presto, etc.). I3 / i3en instances provide instance storage supported by non-volatile memory standard (nvme) SSD at low cost, and are optimized for low latency, high I / O performance and throughput. Customers like the fast transaction processing capability provided by the I3 / i3en instance, but with the continuous upgrading of their workload – processing more complex transactions on larger data sets, they need to obtain higher computing performance and faster data access speed without increasing cost.

The im4gn / is4gen / i4i instance is designed to maximize the storage performance of I / O-Intensive workloads through the architecture. Through self-developed Amazon nitro SSDs, im4gn / is4gen / i4i instances provide up to 30 TB of nvme storage. Compared with the previous generation I3 instances, I / O latency is reduced by 60% and latency variability is reduced by 75%, thus maximizing application performance. Amazon nitro SSDs are tightly integrated with Amazon nitro systems by optimizing storage stacks, virtualization hypervisors, and hardware. Compared with using commercial SSDs, Amazon cloud technology manages the hardware and firmware of Amazon nitro SSDs at the same time, making SSD update delivery faster, so that customers can benefit from improved functions. The im4gn instance (now available) uses Amazon graviton2 processor, which improves the price / performance by up to 40% and reduces the storage cost per TB by up to 44% compared with the I3 instance. The is4gen instance (now available) also uses Amazon   Graviton2 processor, compared with i3en instance, reduces storage cost per TB by up to 15% and improves computing performance by up to 48%. To start using im4gn / is4gn instances, visit aws.amazon.com/ec2/instance-types/i4g. The i4i instance (to be available soon) uses Intel’s third generation scalable processor (ice lake), which improves the computing performance by up to 55% compared with the current generation of I3 instances. To learn more about im4gn / is4gen/   For information on i4i instances, please visit aws.amazon.com/ec2/instance-types/i4i.

SAP Hana is the world’s leading in memory database and the foundation of SAP business technology platform. Irfan Khan, President of SAP Hana database and analysis, said: “Over the past decade, SAP Hana has helped customers manage their most critical transactions and analysis workloads. Amazon cloud technology’s continuous investment and innovation in arm based Amazon graviton processor and its cooperation with Amazon cloud technology have brought many possibilities, bringing significant operational efficiency and benefits to our enterprise customers and SAP cloud analysis and data management solutions supported by SAP Hana cloud Performance improvement. “

Twitter carries what is happening and people are discussing. Nick tornow, director of twitter platform, said: “Twitter is working on a multi-year project to use Amazon EC2 instance based on Amazon graviton processor to provide twitter timeline. To further improve efficiency, we tested a new c7g instance based on Amazon graviton 3 processor. In a series of benchmarks that can represent the performance of twitter workload, we found c7g based on Amazon graviton 3 processor Instance can improve performance by 20% – 80% and reduce tail latency by up to 35% compared with c6g instance based on Amazon graviton2 processor. We are very happy and look forward to better price / performance with Amazon graviton3 processor based instance. “

Formula one (F1) racing began in 1950. It is the most famous racing race in the world and the most popular annual sports event in the world. “The Amazon graviton2 processor based c6g instance has provided the best price / performance for some of our CFD (Computational Fluid Dynamics) workloads. Now, we find that in the same simulation, the graviton3 based c7g instance is 40% faster than the graviton2 based c6g instance.” said Pat Symonds, chief technology officer of formula one Management: “We are pleased that EFA will be standard in this instance type. Based on the excellent performance of graviton3 in terms of cost performance, we expect it to be the best choice to run all CFD workloads.”

Founded in 1991, Epic Games is the creator of fortnite, unreal, gears of war, shadow complex and infinity blade games. Epic’s unreal engine technology brings high fidelity and interactive experience to PC, host, mobile phone, AR, VR and web. Mark imbriaco, senior engineering director of Epic Games, said: “when we look to the future and build a more engaging immersive experience for players, we are happy to use EC2 instances based on Amazon gravition3 processor. Our tests show that they are even suitable for the most demanding and delay sensitive workloads, while providing excellent cost performance and improving the performance of the game Fortress night and any game experience based on a virtual engine. “

Honeycomb has developed a visualization platform that allows engineering teams to visualize, analyze and improve the quality and performance of cloud applications. “We are pleased to be able to test our high throughput workload based on a preview example of Amazon graviton3.”   Liz Fong Jones, chief developer evangelist of honeycomb.io, said: “under the same workload, we run 30% fewer c7g instances than c6g, and the latency is reduced by 30%. We look forward to putting the c7g instance of Amazon graviton3 into our production environment as soon as it is officially available.”

Meta AI focuses on connecting people with their concerns, providing meaningful and safe experiences, promoting machine learning and guiding open research. Lin Qiao, senior director of engineering at meta AI pytorch, said: “Pytorch developers are constantly innovating to improve the accuracy of in-depth learning models and find better ways to solve problems. At the same time, the scale of these models is growing exponentially, which makes it more difficult and costly to train them. Our pytorch team has been working with Amazon neuron team to set a high standard for ease of use and performance to ensure the success of pytorch Provide native support for accelerators such as training in. This includes studying collective computing primitives and setting an appropriate foundation for extended performance and distributed training. We look forward to working with Amazon cloud technology to integrate training more with pytorch native products such as Eagle mode and dynamic shapes. “

Anthropologic has built reliable, interpretable and controllable artificial intelligence systems, which will have the opportunity to create value for business and public interests. Tom Brown, co-founder of anthropologic, said: “Our research interests span many fields, including natural language, artificial feedback, scaling law, reinforcement learning, code generation and interpretability. A key to our success is to use modern infrastructure to enable us to start clusters of super large and high-performance deep learning accelerators. We look forward to using trn1 instances supported by Amazon training chips because they have The unprecedented ability to scale to tens of thousands of nodes and higher network bandwidth will enable us to iterate faster while controlling costs. “

Splunk is a leading provider of data platforms designed to investigate, monitor, analyze and process data of any size. Brad Murphy, vice president of Splunk cloud platform and infrastructure, said: “We run a C / C + + based workload to index and search event data. Our workload is limited by CPU and benefits from high-capacity and low latency SSD storage. When evaluating the new im4gn / is4gen instances supported by Amazon graviton2, we observed that the search run time is reduced by 50% compared with the I3 / i3en instances currently in use. Im4gn and is4gen instances are running me They are a great choice for storage intensive workloads, with significantly better price / performance and lower TCO. “

Sprinklr helps large enterprises around the world improve customer satisfaction through more than 30 digital channels – using the advanced and complex artificial intelligence engine built for enterprises to create insight driven strategies and create a better customer experience. “We benchmarked the Java based search workload on the Amazon EC2 im4gn / is4gen instance supported by the Amazon graviton2 processor. Compared with the larger i3en instance, the smaller is4gen instance can provide similar performance, which means there is an opportunity to significantly reduce TCO,” said Abhay Bansal, vice president of engineering at sprint “When we migrated the workload from the I3 instance to the im4gn instance, we also found that the query latency was significantly reduced by up to 50% and the price / performance was improved by 40%. It was easy to migrate to the Amazon graviton2 based instance and it took two weeks to complete the benchmark. We are very satisfied with our experience and look forward to running these workloads in the production environment based on the im4gn and is4gen instances.”

Redis enterprise provides mission critical applications and services to more than 8000 organizations around the world by helping the software team create a high-performance data layer for the real-time world. Yiftach shoulman, redis co-founder and chief technology officer, said: “We are very pleased to see that the New Amazon EC2 i4i instance of low latency Amazon nitro SSDs provides faster transaction speed than the previous generation instance. We expect that the stronger storage performance and faster network and processor speed of i4i instance will bring significant improvement and more attractive TCO to our customers who use redis on flash based on i4i instance.” More reading: Amazon cloud technology accelerates the implementation of database services in China. Amazon cloud technology Gu fan: Modern Applications face “uncertainty” NASDAQ joins hands with Amazon cloud technology to reshape the coupling of big data companies and cloud computing giants in the capital market: Shence data and Amazon cloud technology mutually support ISV cloud innovation and usher in new opportunities. Amazon cloud technology enables partners to reshape innovation ability. AWS China distributes financial management services to enable customers to use cloud services more cost effectively. AWS announces the volume of Amazon bracket The sub computing service was officially launched. AWS released the apple MacOS instance for Amazon EC2. Gartner’s latest report suggested that enterprise customers learn from Amazon’s digital model. Amazon cloud service (AWS) announced that Amazon route 53 would launch Amazon cloud service (AWS) in China Two new file storage services with self-developed processor Amazon graviton2 are officially launched in China. Amazon cloud technology in China is rich in ecological diversity. Amazon EC2 x2gd with self-developed processor Amazon graviton2 is fully available. Netease games choose AWS global infrastructure to accelerate the development of overseas markets. Shouqi car hailing and Amazon cloud service (AWS) Release the first customized intelligent voice solution in the travel industry

If you want to get the full report, you can contact us by leaving us the comment. If you think the information here might be helpful to others, please actively share it. If you want others to see your attitude towards this report, please actively comment and discuss it. Please stay tuned to us, we will keep updating as much as possible to record future development trends.

RecordTrend.com is a website that focuses on future technologies, markets and user trends. We are responsible for collecting the latest research data, authority data, industry research and analysis reports. We are committed to becoming a data and report sharing platform for professionals and decision makers. We look forward to working with you to record the development trends of today’s economy, technology, industrial chain and business model.Welcome to follow, comment and bookmark us, and hope to share the future with you, and look forward to your success with our help.