Servers & Systems: The Right Compute
ComputeExperts

32 reasons why HPE ProLiant servers with AMD EPYC processors define innovation

HPE ProLiant servers and HPE Apollo systems—powered by AMD EPYC processors—hold 32 world records on a range of benchmarks. Learn about these record-setting solutions and what they can do for you.   

HPE_ProLiant_AMD_EPYC_World_Record.jpg

HPE ProLiant servers and HPE Apollo systems, powered by AMD EPYC processors, hold 32 world records—achieving leadership positions in virtualization, energy efficiency, database analytic workloads, and Java applications.

At HPE compute matters—it’s the foundation for the modern datacenter and the digital engine at the heart of every business. That’s why at HPE our systems are driven by a relentless pace of innovation. Innovation that delivers the real-world performance, efficiency, and investment protection our customers need.

Delivering density in virtualized environments

COVID-19 has forced an unprecedented change in the way we work, challenging businesses to find ways to enable remote work. This massive shift is pushing companies to support more complex types of virtualization deployments.

In order to test how a platform will perform in these types virtualized environments, the HPE Performance Benchmarking team targeted the VMmark3.1.1 benchmark—which allows accurate and reliable benchmarking for virtual data center performance and consumption.

The HPE ProLiant DL385 Gen10 Plus v2 server has taken the title in VMmark 3.1.1, providing 39% better performance than the previous 4-node 2P record holder.1 In fact, performance of the HPE ProLiant DL385 Gen10 Plus v2 server edged out a competitor’s 16 socket result with an 8 socket result by almost 2%! HPE ProLiant servers offer industry-leading density—giving businesses the power to consolidate their compute footprint by packing more VM’s into fewer servers.

Efficiency means saving green in more ways than one

With each generation of HPC technology, HPE strives to dramatically improve the performance of next-generation systems while reducing energy consumption. We are rethinking the way large-scale systems are designed and built in terms of more efficient and cost-effective power and cooling techniques, density, energy-aware software, and ML AIOps to optimize energy effectiveness in order to do exponentially more with less.

To get an understanding of the efficiency of a server in the real world, the HPE Performance Benchmarking team ran SPECpower_ssj® 2008 benchmarks on the HPE Apollo 2000 Gen10 Plus system. Results from this benchmark provide metrics on power management and energy efficiency; the resulting score gives a comparable view of a server’s or system’s energy efficiency.

The HPE Apollo 2000 Gen10 Plus system with HPE ProLiant XL225n Gen10 Plus server captures an astounding 18 world records and up to 54% higher energy efficiency than other 4-node systems.2 As a result, HPE truly has the edge on power when it comes to HPC. This is not just a clever slogan, but fact-based and benchmarked capabilities we can hang our hat on. The HPE Apollo systems are defining a new era of supercomputers intelligently designed to continue to deliver maximum performance without sacrificing sustainability, lowering operating costs, and always balancing environmental footprint.

A need for speed within decision support workloads

Finding a way to harness the surging amount of data to make meaningful insights can seem daunting. A solution needs to be able to examine large volumes of data and execute highly complex queries, then provide answers to critical business questions.

TPC Benchmark® H (TPC-H) provides a way to test systems to these extremely complex workloads—showcasing query execution speed, as well as cost of the system at different database sizes. HPE currently holds 9 records in this space in two different scale factors—2 of which are at a database size of 3 TB and 7 records at the 10 TB scale factor. 

The HPE ProLiant DL345 Gen10 Plus server, a brand new server line targeted for storage-optimized solutions for database workloads, achieved a new world record non-clustered result for decision support database workloads3, and the HPE ProLiant DL385 Gen10 Plus v2 server achieved two number one non-clustered results providing 14.4% better performance compared to a 4-socket server and 15.5% cost reduction.4 Grabbing numerous records on multiple HPE ProLiant platforms, plus holding records that are still standing from previous generation platforms, really showcases the strength of collaboration between HPE and AMD in database workloads.

Optimizing large scale Java-based application performance

From turning on your favorite on demand video service to ordering a shirt from an online retailer, Java-based applications are everywhere. The way they are deployed are almost as numerous as the number of applications developed on them—containers, hypervisors, virtual machines and cloud computing all being popular methods to get large scale Java-based applications to the end customer.

HPE offers two new, record setting, platforms that help deliver stellar performance on Java-based applications—the HPE ProLiant DL385 Gen10 Plus v2 and the HPE ProLiant DL345 Gen10 Plus servers. These new servers are capable of handling the large amounts of data that today’s businesses experience thanks to the blistering performance attributes on the 3rd Gen AMD EPYC processor.

The benchmark used to test these environments mimic the use case of a worldwide supermarket chain with an IT infrastructure that handles point-of-sale requests, online purchases, and data-mining operations. SPECjbb2015 also has other real world applications to workloads such as banking and financial services, manufacturing, and healthcare and life sciences.

The HPE ProLiant DL385 Gen10 Plus v2 server achieved two-processor leadership with better performance by up to 17.1%5 and the HPE ProLiant DL345 Gen10 Plus server with one-processor performance by up to 16.8%.6 These results cement the performance of HPE ProLiant servers, powered by AMD EPYC processors, as record setting solutions within wide-reaching and fast moving virtualized environments handling large amounts of data.

Record-setting performance as-a-service

Customers can easily adopt these latest HPE solutions using HPE GreenLake, an elastic, pay-per-use, as-a-service platform that can run on-premises, at the edge, or in a colocation facility. HPE GreenLake combines the simplicity and agility of the cloud with the governance, compliance, and visibility that comes with hybrid cloud.

Realize innovation and accelerate next

HPE ProLiant servers and HPE Apollo systems with 3rd Gen AMD EPYC™ processors provide foundational compute platforms that deliver unmatched performance, security, automation, and remote management capabilities to support a range of critical workloads that are essential to digital transformation. All these record-setting solutions are ready to tackle demanding workloads and unlock new customer experiences.

Learn more about how HPE and AMD can meet your specific requirements. 


Meet Tyler Hofstede, HPE Compute Product Marketing Manager 

Tyler_Hofstede1.jpgTyler is the Compute Product Marketing Manager for HPE ProLiant rack servers. He brings technical and creative marketing ideas to life and has made a career out of translating technical jargon into relatable content. Connect with Tyler on LinkedIn

 

 


Compute Experts
Hewlett Packard Enterprise

twitter.com/hpe_compute
linkedin.com/showcase/hpe-servers-and-systems/
hpe.com/servers


[1] VMmark® is a product of VMware, Inc. The competitive benchmark claims are based on being the best 4-node, 2P 4-node, and AMD result on the VMmark 3.1.1 benchmark, with a score of 33.58 @ 36 tiles. Results published as of March 15, 2021.  VMmark disclosures are available at vmware.com/products/vmmark/results3x.html.

[2] SPEC and SPECpower_ssj are registered trademarks of the Standard Performance Evaluation Corporation (SPEC); all rights reserved, see spec.org as of March 15, 2021; 4-, 3-, 2-node configurations and 4-, 3-, 2-node 2-processor configurations for the overall, Linux, and Windows categories.

[3] TPC Benchmark™ H (TPC-H) performance as of March 15, 2021. See tpc.org for more information. Claim based on having the #1 performance for a non-clustered system on the TPC-H @ 3000GB scale factor. Configuration: 1 HPE ProLiant DL345 Gen10 Plus server used 1 AMD EPYC 7763 2.45 GHz processor; 1 socket/64 cores/128 threads; Red Hat Enterprise Linux 8.3; Microsoft® SQL Server 2019 Enterprise Edition. TPC-H results show the HPE ProLiant DL345 Gen10 Plus with a result of 1,346,932.7 QphH @ 3000GB and $0.40 USD/QphH @ 3000GB with a system availability of April 19, 2021; see tpc.org/3352 for details.

[4] TPC-H performance as of March 15, 2021. See tpc.org for more information. Claim based on having the #1 performance and price-performance for a non-clustered system on the TPC-H @ 10000GB scale factor. Configuration: 1 HPE ProLiant DL385 Gen10 Plus v2 server used 2 AMD EPYC 7763 2.45 GHz processors; 2 socket/128 cores/256 threads; Red Hat Enterprise Linux 8.3; Microsoft® SQL Server 2019 Enterprise Edition. TPC-H results show the HPE ProLiant DL385 Gen10 Plus v2 with a result of 1,883,497.4 QphH @ 10000GB and $0.56 USD/QphH @ 10000GB with a system availability of April 19, 2021; see tpc.org/3351 for details.

 [5, 6] SPEC and SPECjbb are registered trademarks of the Standard Performance Evaluation Corporation (SPEC); all rights reserved, see spec.org as of March 15, 2021.

About the Author

ComputeExperts

Our team of Hewlett Packard Enterprise server experts helps you to dive deep into relevant infrastructure topics.