Inspur | July 04, 2022
The open engineering consortium MLCommons released the latest MLPerf Training v2.0 results, with Inspur AI servers leading in closed division single-node performance.
MLPerf is the world’s most influential benchmark for AI performance. It is managed by MLCommons, with members from more than 50 global leading AI companies and top academic institutions, including Inspur Information, Google, Facebook, NVIDIA, Intel, Harvard University, Stanford University, and the University of California, Berkeley. MLPerf AI Training benchmarks are held twice a year to track improvements in computing performance and provide authoritative data guidance for users.
The latest MLPerf Training v2.0 attracted 21 global manufacturers and research institutions, including Inspur Information, Google, NVIDIA, Baidu, Intel-Habana, and Graphcore. There were 264 submissions, a 50% increase over the previous round. The eight AI benchmarks cover current mainstream usage AI scenarios, including image classification with ResNet, medical image segmentation with 3D U-Net, light-weight object detection with RetinaNet, heavy-weight object detection with Mask R-CNN, speech recognition with RNN-T, natural language processing with BERT, recommendation with DLRM, and reinforcement learning with MiniGo.
Among the closed division benchmarks for single-node systems, Inspur Information with its high-end AI servers was the top performer in natural language processing with BERT, recommendation with DLRM, and speech recognition with RNN-T. It won the most titles among single-node system submitters. For mainstream high-end AI servers equipped with eight NVIDIA A100 Tensor Core GPUs, Inspur Information AI servers were top ranked in five tasks (BERT, DLRM, RNN-T, ResNet and Mask R-CNN).
Continuing to lead in AI computing performance
Inspur AI servers continue to achieve AI performance breakthroughs through comprehensive software and hardware optimization. Compared to the MLPerf v0.5 results in 2018, Inspur AI servers showed significant performance improvements of up to 789% for typical 8-GPU server models.
The leading performance of Inspur AI servers in MLPerf is a result of its outstanding design innovation and full-stack optimization capabilities for AI. Focusing on the bottleneck of intensive I/O transmission in AI training, the PCIe retimer-free design of Inspur AI servers allows for high-speed interconnection between CPUs and GPUs for reduced communication delays. For high-load, multi-GPU collaborative task scheduling, data transmission between NUMA nodes and GPUs is optimized to ensure that data I/O in training tasks is at the highest performance state. In terms of heat dissipation, Inspur Information takes the lead in deploying eight 500W high-end NVIDIA Tensor Core A100 GPUs in a 4U space, and supports air cooling and liquid cooling. Meanwhile, Inspur AI servers continue to optimize pre-training data processing performance, and adopt combined optimization strategies such as hyperparameter and NCCL parameter, as well as the many enhancements provided by the NVIDIA AI software stack, to maximize AI model training performance.
Greatly improving Transformer training performance
Pre-trained massive models based on the Transformer neural network architecture have led to the development of a new generation of AI algorithms. The BERT model in the MLPerf benchmarks is based on the Transformer architecture. Transformer’s concise and stackable architecture makes the training of massive models with huge parameters possible. This has led to a huge improvement in large model algorithms, but necessitates higher requirements for processing performance, communication interconnection, I/O performance, parallel extensions, topology and heat dissipation for AI systems.
In the BERT benchmark, Inspur AI servers further improved BERT training performance by using methods including optimizing data preprocessing, improving dense parameter communication between NVIDIA GPUs and automatic optimization of hyperparameters, etc. Inspur Information AI servers can complete BERT model training of approximately 330 million parameters in just 15.869 minutes using 2,850,176 pieces of data from the Wikipedia data set, a performance improvement of 309% compared to the top performance of 49.01 minutes in Training v0.7. To this point, Inspur AI servers have won the MLPerf Training BERT benchmark for the third consecutive time.
Inspur Information’s two AI servers with top scores in MLPerf Training v2.0 are NF5488A5 and NF5688M6. The NF5488A5 is one of the first servers in the world to support eight NVIDIA A100 Tensor Core GPUs with NVIDIA NVLink technology and two AMD Milan CPUs in a 4U space. It supports both liquid cooling and air cooling. It has won a total of 40 MLPerf titles. NF5688M6 is a scalable AI server designed for large-scale data center optimization. It supports eight NVIDIA A100 Tensor Core GPUs and two Intel Ice Lake CPUs, up to 13 PCIe Gen4 IO, and has won a total of 25 MLPerf titles.
About Inspur Information
Inspur Information is a leading provider of data center infrastructure, cloud computing, and AI solutions. It is the world’s 2nd largest server manufacturer. Through engineering and innovation, Inspur Information delivers cutting-edge computing hardware design and extensive product offerings to address important technology sectors such as open computing, cloud data center, AI, and deep learning. Performance-optimized and purpose-built, our world-class solutions empower customers to tackle specific workloads and real-world challenges.
IDC | July 01, 2022
According to the International Data Corporation (IDC) Worldwide Quarterly Enterprise Infrastructure Tracker: Buyer and Cloud Deployment, spending on compute and storage infrastructure products for cloud deployments, including dedicated and shared environments, increased 17.2% year over year in the first quarter of 2022 (1Q22) to $18.3 billion. This growth continues a series of strong year-over-year increases in spending on infrastructure products by both service providers and enterprises despite tight supply of some system components and disruptions in transportation networks. Investments in non-cloud infrastructure increased 9.8% year over year in 1Q22 to $14.8 billion, continuing a streak of growth for this segment into its fifth quarter.
Spending on shared cloud infrastructure reached $12.5 billion in the quarter, increasing 15.7% compared to a year ago. IDC expects to see continuously strong demand for shared cloud infrastructure with spending expected to surpass non-cloud infrastructure spending in 2022 for the first time. Spending on dedicated cloud infrastructure increased 20.5% year over year in 1Q22 to $5.9 billion. Of the total dedicated cloud infrastructure, 47.8% was deployed on customer premises.
For the full year 2022, IDC is forecasting cloud infrastructure spending to grow 22% compared to 2021 to $90.2 billion – the highest annual growth rate since 2018 – while non-cloud infrastructure is expected to grow 1.8% to $60.7 billion. The increased forecast for both segments is partially driven by inflationary pressure and expectations of higher systems prices during 2022 as well as improvements in the supply chain in the second half of the year. Shared cloud infrastructure is expected to grow by 24.3% year over year to $63.9 billion for the full year. Spending on dedicated cloud infrastructure is expected to grow 16.8% to $26.3 billion for the full year.
As part of the Tracker, IDC follows various categories of service providers and how much compute and storage infrastructure these service providers purchase, including both cloud and non-cloud infrastructure. The service provider category includes cloud service providers, digital service providers, communications service providers, and managed service providers. In 1Q22, service providers as a group spent $18.3 billion on compute and storage infrastructure, up 14.5% from 1Q21. This spending accounted for 55.3% of total compute and storage infrastructure spending. Spending by non-service providers increased 12.9% year over year, the highest growth in fourteen quarters. IDC expects compute and storage spending by service providers to reach $89.1 billion in 2022, growing 18.7% year over year.
At the regional level, year-over-year spending on cloud infrastructure in 1Q22 increased in most regions. Once again Asia/Pacific (excluding Japan and China) (APeJC) grew the most at 50.1% year over year. Japan, Middle East and Africa, China, and the United States all saw double-digit growth in spending. Western Europe grew 6.4% and growth in Canada slowed to 1.2%. Central & Eastern Europe, affected by the war between Russia and Ukraine, declined 10.3%, while Latin America declined 11.3%. For 2022, cloud infrastructure spending for most regions is expected to grow, with four regions, APeJC, China, the U.S., and Western Europe, expecting to post annual growth in the 20-25% range. Impact of the war will continue to hurt spending in Central and Eastern Europe, which is now expected to decline 54.6% in 2022.
Long term, IDC expects spending on compute and storage cloud infrastructure to have a compound annual growth rate (CAGR) of 14.5% over the 2021-2026 forecast period, reaching $145.2 billion in 2026 and accounting for 69.7% of total compute and storage infrastructure spend. Shared cloud infrastructure will account for 72.6% of the total cloud amount, growing at a 15.4% CAGR. Spending on dedicated cloud infrastructure will grow at a CAGR of 12.1%. Spending on non-cloud infrastructure will grow at 1.2% CAGR, reaching $63.1 billion in 2026. Spending by service providers on compute and storage infrastructure is expected to grow at a 13.4% CAGR, reaching $140.8 billion in 2026.
A graphic illustrating IDC's 2021-2026 forecast for worldwide enterprise infrastructure spending by deployment type (Shared Cloud, Dedicated Cloud, and Non-Cloud) is available by viewing this press release on IDC.com.
IDC's Worldwide Quarterly Enterprise Infrastructure Tracker: Buyer and Cloud Deployment is designed to provide clients with a better understanding of what portion of the compute and storage hardware markets are being deployed in cloud environments. The Tracker breaks out each vendors' revenue into shared and dedicated cloud environments for historical data and provides a five-year forecast. This Tracker is part of the Worldwide Quarterly Enterprise Infrastructure Tracker, which provides a holistic total addressable market view of the four key enabling infrastructure technologies for the datacenter (servers, external enterprise storage systems, and purpose-built appliances: HCI and PBBA).
IDC defines cloud services more formally through a checklist of key attributes that an offering must manifest to end users of the service.
Shared cloud services are shared among unrelated enterprises and consumers; open to a largely unrestricted universe of potential users; and designed for a market, not a single enterprise. The shared cloud market includes a variety of services designed to extend or, in some cases, replace IT infrastructure deployed in corporate datacenters; these services in total are called public cloud services. The shared cloud market also includes digital services such as media/content distribution, sharing and search, social media, and e-commerce.
Dedicated cloud services are shared within a single enterprise or an extended enterprise with restrictions on access and level of resource dedication and defined/controlled by the enterprise (and beyond the control available in public cloud offerings); can be onsite or offsite; and can be managed by a third-party or in-house staff. In dedicated cloud that is managed by in-house staff, "vendors (cloud service providers)" are equivalent to the IT departments/shared service departments within enterprises/groups. In this utilization model, where standardized services are jointly used within the enterprise/group, business departments, offices, and employees are the "service users."
About IDC Trackers
IDC Tracker products provide accurate and timely market size, vendor share, and forecasts for hundreds of technology markets from more than 100 countries around the globe. Using proprietary tools and research processes, IDC's Trackers are updated on a semiannual, quarterly, and monthly basis. Tracker results are delivered to clients in user-friendly Excel deliverables and on-line query tools.
International Data Corporation (IDC) is the premier global provider of market intelligence, advisory services, and events for the information technology, telecommunications, and consumer technology markets. With more than 1,300 analysts worldwide, IDC offers global, regional, and local expertise on technology, IT benchmarking and sourcing, and industry opportunities and trends in over 110 countries. IDC's analysis and insight helps IT professionals, business executives, and the investment community to make fact-based technology decisions and to achieve their key business objectives. Founded in 1964, IDC is a wholly owned subsidiary of International Data Group (IDG), the world's leading tech media, data, and marketing services company.
Vodafone | June 22, 2022
In order to modernize the operator's European IT infrastructure and hasten its shift to the cloud, Oracle and Vodafone, the largest pan-European and African technological communications firm, established a strategic alliance.
A major portion of Vodafone's systems will be updated and moved to OCI Dedicated Region, a fully managed cloud region that integrates all of Oracle's public cloud services into Vodafone's own network and data centers, as part of the multi-year arrangement. As a result, in addition to updating its tens of thousands of Oracle databases, Vodafone will be able to support and extend its mission-critical OSS and BSS services, such as CRM and order management, on a dedicated cloud platform.
Vodafone will be able to develop new cloud-based applications more quickly thanks to the deployment, and by utilizing its geographic reach, it will be able to simultaneously launch them in a number of markets.
In Vodafone's primary data centers, which oversee its European IT and network operations, Oracle will implement OCI Dedicated Region. As a result, Vodafone will be able to more easily meet the latency and performance requirements of these applications by deploying public cloud services inside of its own network and data centers. In addition, this will allow the operator to modernize, manage, and automate its critical systems using new technologies like autonomous services.
"As Vodafone focuses on growth, data is key to how we evolve our business, build new capabilities and innovate to meet the needs of our customers. Our collaboration with Oracle supports our vision of becoming a technology communications company. The agreement enables Oracle to bring its entire portfolio of cloud services directly into Vodafone data centers. This includes the same architecture, software, services and control plane used in OCI public cloud. The flexibility offered by OCI enables us to build a robust, secure, and extensible cloud platform in our own data centers, while also providing the operational agility and scalability required to support the growth and diversification of our business."
Scott Petty, Chief Digital & IT Officer, Vodafone
The collaboration backs Vodafone's multi-year commitment to streamline and update the IT infrastructure that underpins its mission-critical systems into a common, on-premises, open-standard platform ready to support and scale next-generation digital services. Additionally, it will assist Vodafone in achieving its Tech 2025 objectives, which include speeding up the time it takes to market new services, delivering standout client experiences through always-on services, and cutting operating costs through automation.
"Now more than ever, telecom companies need to quickly adopt new technologies to deliver new innovative products at speed while continuing to meet evolving regulatory requirements. Our partnership with Vodafone is based on achieving this balance, providing a cloud platform that enables Vodafone to modernize and consolidate its existing infrastructure while also building a foundation for a digital future. We are looking forward to partnering with one of the telecom sector's digital trailblazers as we help shape the next generation of communication services and business models," said Jonathan Tikochinsky, executive vice president, global strategic clients, Oracle.