ASUS Sets Most World Records for 1P and 2P Server Performance on SPEC.org

2020/11/05

Leading server systems provider scores 683 records for top performance across multiple architectures and platforms in intensive computing workloads

TAIPEI, Taiwan, October 29, 2020 ASUS, a leading server systems, server motherboards, and workstations provider, today announced that ASUS servers continue to dominate computing performance with a total of 683 new world-record benchmark results confirmed by the Standard Performance Evaluation Corporation (SPEC, www.spec.org). ASUS holds the most amount of records on the SPEC CPU® 2017 benchmark in single-socket (1P) and dual-socket (2P). These world records are set by servers running across Intel® and AMD® platforms, including the 1st and 2nd Gen Intel Xeon® Scalable processors and AMD EPYC™ 7002 processors and workloads ranging from general business infrastructure, software-defined deployment, data analytics, AI, and HPC (High Performance Computing). These results are correct as of September 29, 2020.



ASUS offers a comprehensive range of AMD EPYC product portfolio

ASUS servers with AMD EPYC 7002 processors currently hold an incredible 213 world records. This total consists of 167 No. 1 results for single-socket performance, plus 46 No. 1 results for dual-socket performance across all SPEC CPU 2017 benchmarks, including integer rate, floating-point rate, integer speed and floating-point. ASUS offer a wide range of AMD EPYC server categories including GPU-optimized, high-density and general-purposed in 1U and 2U chassis, supporting up to 64 cores, 128 threads, 8-channel DDR4 3200 and 128 PCIe® 4.0 lanes, all powered by AMD EPYC 7002 processors. Each server category is designed to fit the different workloads of AI inference, virtualization, VDI, analytics and software-defined deployments.



GPU-optimized server: ESC4000A-E10 supports up to four double-slot GPUs

ASUS ESC4000A-E10 is a 2U GPU server powered by single-socket AMD EPYC processor designed for AI, HPC and virtual desktop infrastructure (VDI) applications in data center or enterprise environments requiring powerful CPU cores, multiple GPU support and faster transmission speeds.


ESC4000A-E10 delivers GPU-optimized performance with support for up to four high-performance double-slot or eight single-slot GPUs, including the latest NVIDIA® A100 PCIe GPUs built on the NVIDIA Ampere architecture, Tesla T4 and Quadro® GPUs. This performance also provides benefits for virtualization by consolidating GPU resources into a shared pool, enabling users to utilize resources in more efficient ways.



High-density server: RS620SA-E10-RS12 is world’s first EPYC 2U6N high-density server

RS620SA-E10-RS12 contains six nodes in 2U chassis and each node is powered by a AMD EPYC 7002 processor. This server supports a TDP of up to 280 W and up to 64 cores per node, and PCIe 4.0 bandwidth including one PCIe 4.0 x16 slot and one OCP 3.0 slot per node. RS620SA-E10-RS12 also has support for 12 hot-swappable storage bays, enabling extreme flexibility. This includes up to six SATA/SAS hard drives and six NVMe™ SSDs, or all 12 bays populated with NVMe SSD storage. With the dual advantages of high core-counts and high transfer speeds, RS620SA-E10-RS12 offers superior overall system flexibility.



General-purpose rack servers: RS720A and RS500A support diverse sockets and form factors

ASUS also offers a range of general-purpose rack servers to fit diverse business needs and environments, with a variety of storage options, power supplies, networking and expansion options. ASUS RS720A-E9 and RS700A-E10 are dual-socket servers in 2U and 1U chassis offering great balance on performance, efficiency and manageability. ASUS RS500A-E10 and RS500A-E9 are single-socket servers in compact 1U chassis, and position essential features for growing organization at an affordable price point.



ASUS Performance Boost tuning technology

ASUS servers feature exclusive Performance Boost tuning technology to achieve the best server performance and agility by tuning servers to match workload requirements, allowing users to gain greater control of data centers. This improves workload throughput by maximizing processor frequency and boost power, making it ideal for time-sensitive applications such as financial services or data center operations. Performance Boost consists of three key technologies. Core Optimizer maximizes the processor frequency in multi-core operations, avoiding frequency shifting for reduced latency. Engine Boost delivers automatic power acceleration with an innovative voltage design to increase server overall performance, delivering up to 24.5%-greater Tflops performance . Workload Presets is a preconfigured BIOS-server profile based on workloads and benchmarks for improved performance and efficiency, which reaches 5.8% performance improvement at peak frequency.



About SPEC

The Standard Performance Evaluation Corporation (SPEC) is a nonprofit corporation formed to establish, maintain and endorse standardized benchmarks and tools to evaluate performance and energy efficiency for the newest generation of computing systems. SPEC develops benchmark suites and also reviews and publishes submitted results from its member organizations and other benchmark licensees.



The SPEC CPU 2017 benchmark package contains SPEC's next-generation, industry-standardized, CPU intensive suites for measuring and comparing compute-intensive performance, stressing a system's processor, memory subsystem and compiler. SPEC designed these suites to provide a comparative measure of compute-intensive performance across the widest practical range of hardware using workloads developed from real user applications.