Covering Scientific & Technical AI | Wednesday, December 25, 2024

Inspur NF8480M6 Sets New Record in SAP SD 2-Tier Benchmark for 4-Socket Servers from the Cooper Lake Platform 

SAN JOSE, Calif., Jan. 4, 2023 -- Inspur Information, a leading IT infrastructure solutions provider, announced that its NF8480M6 has ranked first in the SAP SD 2-tier performance benchmark for 4-socket servers with the Copper Lake platform and hit a world record of 359,780 SAPS.

SAP SD 2-Tier Benchmark: A Critical Benchmark for Operational Performance of Real Production Systems

The SAP Sales & Distribution 2-Tier Benchmark (SAP SD 2-Tier Benchmark) is a benchmark based on sales and distribution modules in a two-tier architecture. It is the most representative server hardware performance test among SAP benchmark tests. In this test, applications and databases are installed on the same single physical server. Frequent, repetitive standardized sales and distribution business operations are simulated upon request by highly concurrent users, in order to verify the maximum concurrent load pressure that the hardware platform can bear and further quantify and measure the capability of a single-server system to execute applications and databases concurrently.

The SAP Application Performance Standard (SAPS) is a unit of measure developed by SAP to measure the performance of hardware servers. SAPS can be understood as the standard performance value of an SAP SD application module. 100 SAPS is defined as 2,000 fully business processed order line items per hour, each of which refers to a complete business process of an order line item that consists of 18 specific steps, including new order generation, shipping order generation, order display, change of shipping content, goods entry, listing of orders and invoice generation. Also, 100 SAPS is equivalent to processing 6,000 dialog steps (screen changes) and 2,000 postings per hour in the SD benchmark or 2,400 SAP transactions per hour.

Considering the actual production needs of enterprise users in the design, the SAP SD 2-Tier Benchmark Test has become a key measure of operational performance for the production systems of real enterprises, and is considered a more versatile and valuable basis for selecting a server platform by numerous enterprise users. In the test, a higher benchmark test score (i.e., a higher SAPS value) means that the server system can support more users and provide higher productivity for enterprises.

Inspur NF8480M6 Hits a New Record in SAP SD 2-Tier Benchmark for 4-Socket Servers

The SAP SD 2-Tier Benchmark simulates the maximum number of concurrent users the system can process within an acceptable response time (<1s) in real scenarios. As one of the most CPU-consuming benchmark tests in the industry, it is highly valued by many server hardware manufacturers and continuously showcases new performance records.

The benchmark test is a process of gradually increasing the load until the system reaches its bottleneck or a performance point it cannot process, and then determining the concurrent performance of the system via a comprehensive analysis of transaction execution indicators and resource monitoring indicators. The goal is to test and determine the peak performance of the system based on a large number of requests from concurrent users, response times, CPU load and memory usage, etc., as the load gradually increases, thus identifying the maximum service level that the system can offer.

The concurrent performance test mainly aims to fulfill three tasks: Representative and critical business operations are selected on the basis of real business to design test cases for evaluating the current performance of the system. For extending the functionality of an application or deploying a new application, the load test will help determine whether the system can still process the desired user load, and customers can reference the SAPS result for the model selection of application hardware. By simulating and repeatedly executing the operations by thousands of users under well-defined business scenarios in the standard test, it can help confirm the performance bottlenecks of the whole system, thereby discovering the weaknesses of the system and implementing optimization and adjustment in time.

The Inspur Information NF8480M6 participated in the SAP SD 2-Tier Benchmark Test with SUSE 15 SP2 as its operating system and the application version based on SAP ERP 6.0 SP5. In the end, NF8480M6 achieved 359,780 SAPS, which is equivalent to processing 7,195,670 business order line items per hour, at the high concurrent requests of 62,900 users. The average session response time was just 0.49 seconds. This test result is a new global record for a 4-socket server from the Cooper Lake platform, demonstrating that NF8480M6 is able to provide excellent performance for large-scale and critical business applications.

Inspur Information Multi-Socket Server Provides an Excellent System Platform for Critical Business Applications

This groundbreaking performance record of a 4-socket server in the SAP SD 2-Tier Benchmark Test is attributed to the design philosophy of extremeness, excellence, security and openness for Inspur NF8480M6 to optimize the server's performance, reliability and ease of use.

In terms of performance, the NF8480M6 processor has been upgraded with 6 UPI channels, and the data communication and interconnection bandwidth has been doubled to over 20.8 GT/s. Inspur Information has designed a unique topology that can achieve a better balance between increasing the throughput and reducing the overall energy consumption, through which the performance of high-concurrent applications in a non-NUMA architecture can be improved by 10%, providing users with powerful concurrent computing and processing capabilities.

In terms of I/O throughput, the NF8480M6 motherboard supports up to 18 TB of memory and has the memory rate upgraded to 3,200 MT/s, which can significantly reduce the system waiting time caused by frequent I/Os. In terms of data storage, the NF8480M6 supports 24 hot-swap NVMe SSDs with an IOPS 10 times higher than that of SATA SSDs, thereby breaking through the data access bottleneck and satisfying intense I/O requirements.

In terms of security, the NF8480M6 achieves redundancy of power supplies, fans, and firmware chips and supports memory fault isolation, hot-swapping and intelligent management at three-level upgrades of hardware, firmware, and system. This significantly improves system reliability, reduces the risk of unplanned system shutdowns, increases the mean time between failures (MTBF) of the whole unit by more than 200,000 hours and guarantees trouble-free, stable operation of the system for more than 8 years. Moreover, the NF8480M6 is also strongly open; it supports hot-swap OCP 3.0 cards, shortening the NIC replacement time from 20 minutes to 1 minute and improving the O&M efficiency.

Since 2016, Inspur Information has established a global technology partnership (GTP) with SAP, and has worked closely with the SAP PAC team to successfully provide multiple solutions for customers. Right now, Inspur Information has various configuration models of mainstream 2-socket, 4-socket and 8-socket servers certified by SAP HANA Appliance, covering two major system architectures—Scale-up and Scale-out, which can provide a safe, reliable, flexible and scalable computing platform of excellent performance for SAP application loads and therefore help enterprises gain an insight into data value more efficiently and accelerate digital transformation.

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.


Source: Inspur

AIwire