The Future-Forward Platform Foundation for Agile Digital Services

Discover the Benefits

  1. 1

    Pervasive, breakthrough performance enables greater efficiencies and lower TCO

  2. 2

    Strong, capable platforms for the hybrid-cloud, data-fueled enterprise

  3. 3

    Next-generation platform for cloud-optimized, 5G-ready networks

  4. 4

    Breakthrough HPC and high-performance data analytics innovation

Empowering Transformation in a Digital World
Across an evolving digital world, disruptive and emerging technology trends in business, industry, science, and entertainment increasingly impact the world’s economies. By 2020, the success of half of the world’s Global 2000 companies will depend on their abilities to create digitally enhanced products, services, and experiences,1 and large organizations expect to see an 80 percent increase in their digital revenues, 2 all driven by advancements in technology and usage models they enable.

The Intel® Xeon® Scalable platform provides the foundation for a powerful data center platform that creates an evolutionary leap in agility and scalability.

This global transformation is rapidly scaling the demands for flexible compute, networking, and storage. Future workloads will necessitate infrastructures that can seamlessly scale to support immediate responsiveness and widely diverse performance requirements. The exponential growth of data generation and consumption, the rapid expansion of cloud-scale computing, emerging 5G networks, and the extension of high-performance computing (HPC) and artificial intelligence (AI) into new usages require that today’s data centers and networks urgently evolve—or be left behind in a highly competitive environment. These demands are driving the architecture of modernized, future-ready data centers and networks that can quickly flex and scale.

The Intel® Xeon® Scalable platform provides the foundation for a powerful data center platform that creates an evolutionary leap in agility and scalability. Disruptive by design, this innovative processor sets a new level of platform convergence and capabilities across compute, storage, memory, network, and security. Enterprises and cloud and communications service providers can now drive forward their most ambitious digital initiatives with a feature-rich, highly versatile platform.

View the infographic ›

Enabling Greater Efficiencies and Lower TCO
Across infrastructures, from enterprise to technical computing applications, the Intel Xeon Scalable platform is designed for data center modernization to drive operational efficiencies that lead to improved total cost of ownership (TCO) and higher productivity for users. Systems built on the Intel Xeon Scalable platform are designed to deliver agile services and reduce TCO up to 65 percent3 due to lower software and OS licensing fees, and acquisition, maintenance, and infrastructure costs.

Virtualization in the enterprise has been steadily on the rise for a decade. Most organizations have taken advantage of some form of virtualization, creating greater need for servers to run virtual machines (VMs) in the data center. Intel Xeon Scalable platform enables 4.2x4 more VMs per server compared to earlier generations, allowing IT to increase their consolidation of more services on less hardware.

Pervasive, Breakthrough Performance
From its new Intel® Mesh Architecture and widely expanded resources to its hardware-accelerating and newly integrated technologies, the Intel Xeon Scalable platform enables a new level of consistent, pervasive, and breakthrough performance.

Comparing Intel® Xeon® Scalable platform to both four-year-old and last-generation Intel Xeon processor-based systems, the latest generation Intel CPU delivers better performance and capabilities for enterprise, cloud, communications, and HPC.4 5 6 7 8 9 10 11

Foundational enhancements:

  • Higher per-core performance: Up to 28 cores, delivering high performance and scalability for compute-intensive workloads across compute, storage, and network usages.
  • Greater memory bandwidth/capacity: 50 percent increased memory bandwidth and capacity. Six memory channels versus four memory channels of previous generation for memory-intensive workloads.
  • Expanded I/O: 48 lanes of PCIe* 3.0 bandwidth and throughput for demanding I/O-intensive workloads.
  • Intel® Ultra Path Interconnect (Intel® UPI): Up to three Intel UPI channels increase scalability of the platform to as many as eight sockets, as well as improves inter-CPU bandwidth for I/O-intensive workloads versus previous generation12 (with Intel® Quick Path Interconnect). Intel UPI offers the perfect balance between improved throughput and energy efficiency.
  • Intel® Advanced Vector Extensions 512 (Intel® AVX-512): With double the flops per clock cycle compared to previous generation Intel® AVX2,13 Intel AVX-512 boosts performance, and throughput for the most demanding computational tasks in applications, such as modeling and simulation, data analytics and machine learning, data compression, visualization, and digital content creation.
  • Security without compromise: Near-zero encryption overhead14 enables higher performance on all secure data transactions.

Innovative Integrations
First-time platform integrations deliver improvements in performance and latency across the infrastructure:

  • Integrated Intel® Omni-Path Architecture (Intel® OPA) Host Fabric Interface: End-to-end high-bandwidth, low-latency fabric that optimizes performance and eases deployment of HPC clusters by eliminating the need for a discrete host fabric interface card. Integrated in the CPU package.
  • Integrated Intel® QuickAssist Technology (Intel® QAT): Chipset-based hardware acceleration for growing compression and cryptographic workloads for greater efficiency while delivering enhanced data transport and protection across server, storage, and network infrastructure.
  • Integrated Intel® Ethernet with scalable iWARP* RDMA*: Provides up to four 10 Gbps high-speed Ethernet ports for high data throughput and low-latency workloads. Ideal for software-defined storage solutions, NVM Express* over Fabric solutions, and virtual machine migrations. Integrated in the chipset.

Integrated Intel® Omni-Path Architecture (Intel® OPA) Host Fabric Interface is an end-to-end high-bandwidth, low-latency fabric that optimizes performance and eases deployment of HPC clusters.

Industry-Leading Storage Support
Storage innovations can drive significant improvements in efficiency and performance of data-hungry workloads.

  • Support for Intel® Optane™ SSDs and Intel® 3D NAND Solid State Drives: Delivers industry-leading combination of high throughput, low latency, high QoS, and ultra-high endurance to break through data access bottlenecks.
  • Deploy next generation storage with confidence with Intel® Volume Management Device (Intel® VMD): Enables hot swapping of NVMe SSDs from the PCIe bus without shutting down the system, while standardized LED management helps provide quicker identification of SSD status. This commonality brings enterprise reliability, availability, and serviceability (RAS) features to NVMe SSDs, enabling deployment of next-generation storage with confidence.
  • Intel® Intelligent Storage Acceleration Library (Intel® ISA-L): Optimizes storage operations, such as encryption, for increased storage performance.
    • Up to 3.1x faster cryptographic hashing of SHA algorithms using Intel AVX-512.15
    • Up to 1.2x algorithm speedup for AES-128-GCM.15
    • Up to 2x faster Reed Solomon Erasure Code performance using Intel AVX-512.15

Complementary Offerings for Even Greater Performance, Scalability
Intel offers a broad hardware and software portfolio that complements this new processor.

  • The Intel® Xeon Phi™ Processor is an ideal foundation for highly parallel applications, such as machine learning training, simulation, and visualization.
  • Intel® FPGAs offer efficient acceleration16 along with the flexibility that comes with using programmable hardware for low-latency applications, such as virtual switching, network services, data analytics, and AI.
  • A range of software tools and libraries for general and highly parallel computing help developers optimize applications for Intel® Architecture.

Enhanced Platform Trust
Data and platform reliability and protection are key concerns for enterprises dealing with increasing concerns and scrutiny regarding data security and privacy. Intel Xeon Scalable platform helps build highly trusted infrastructures with platform data protection, resiliency, and uptime.

Increased data protection and reliability across every workload:

  • Enhanced Intel® Run Sure Technology: New enhancements deliver advanced Reliability, Availability, and Serviceability (RAS) and server uptime for a company’s most critical workloads. Hardware-assisted capabilities, including enhanced MCA and recovery and adaptive multi-device error correction, diagnose, and recover from previously fatal errors. And, they help ensure data integrity within the memory subsystem.
  • Intel® Key Protection Technology (Intel® KPT) with Integrated Intel QAT and Intel® Platform Trust Technology (Intel® PTT): Deliver hardware-enhanced platform security by providing efficient key and data protection at rest, in-use, and in-flight.
  • Intel® Trusted Execution Technology (Intel® TXT) with One-Touch Activation: Enhanced platform security, while providing simplified and scalable deployment for Intel TXT.

As more data-rich workloads flow through the data center, this comprehensive suite of hardware-enhanced features brings better data- and platform-level protection mechanisms for trusted services in enterprise and cloud environments.

Intel® Xeon® Processor Provides Consistent Performance for Diverse Workloads

 

What Matters?

Intel® Xeon® Scalable Platform Advantages

Enterprise and Cloud

Minimize complexity with compatible virtualization infrastructure

Meet stringent customer SLAs

Deploy quickly. Intel VMs coexist w/ other Intel® technology-based servers.

Fast response times.

HPC

Maximize vector floating performance and efficiency

High performance with fewer servers.

Storage

Ensure deterministic storage response

Deterministic performance. Cores, cache, memory, I/O in a single die.

Communications

Deliver diverse services efficiently

Incredible efficiency and hardware acceleration from a platform with features required for application, control, packet, and signal processing.

Dynamic and Highly Efficient Service Delivery
The convergence of enhanced compute, memory, network, and storage performance, combined with software ecosystem optimizations, make Intel Xeon Scalable  platform the ideal platform for fully virtualized, software-defined data centers that dynamically self-provision resources—on-premise, through the network, and in the public cloud—based on workload needs.

Powerful tools and technologies for an agile data center:

  • Intel® VT-x new features:
    • Mode based execution control (MBE) virtualization: Provides an extra layer of protection from malware attacks in a virtualized environment by enabling hypervisors to more reliably verify and enforce the integrity of kernel level code.
    • Timestamp counter scaling (TSC) virtualization: Provides workload optimization in hybrid cloud environments by allowing virtual machines to move across CPUs operating at different base frequencies.
  • Intel® Node Manager 4.0: Helps IT intelligently manage and optimize power, cooling, and compute resources in the data center, maximizing efficiency, while reducing the chances of costly overheats.

Strong, Capable Platforms for the Hybrid-Cloud, Data-Fueled Enterprise
Enterprises are keen to extract value from the exploding data streams being presented to them for rapid insights that can shape their business initiatives. Traditional and emerging applications in the enterprise, including predictive analytics, machine learning, and HPC, require new levels of powerful compute capabilities and massive tiered data storage volumes. The modernized data center is being architected using a converged and holistic approach that can flexibly deliver new services and improve TCO across infrastructure assets today, while providing the most seamless and scalable on-ramp to a self-governing, hybrid data center.

Yet, organizations running their foundational business workloads, such as OLTP and web infrastructure, seek to reduce TCO with higher performing infrastructures.

The Intel Xeon Scalable platform delivers next-generation enterprise capabilities to businesses through a future-ready platform that can serve the hybrid-cloud, data-fueled era, plus it helps improve day-to-day operations with up to 58 percent more requests/second on runtime workloads.17 This versatile platform brings disruptive levels of compute performance, coupled with memory and I/O advances, to compute-hungry and latency-sensitive applications. Combined with innovative Intel® SSD Data Center Family to manage large data volumes across storage, caching, and memory, platforms built on the Intel Xeon Scalable platform are ready to handle the intense demands of the data and cloud era. 

With a scalable portfolio of packaging options to suit diverse workload requirements, the Intel Xeon Scalable platform is a performance workhorse designed for deploying highly efficient, virtualized infrastructures for compute, storage, and networking.

Highlights for enterprise innovation:

  • Up to 28 high-performance cores, six memory channels, and 48 lanes of PCIe 3.0
  • Up to three Intel UPI channels
  • Support for Intel Optane SSDs and Intel® 3D NAND SSDs
  • Enhanced Intel Run Sure Technology

Combined with the Data Plane Development Kit (DPDK) and Intel QAT, this new processor can boost network performance, enabling service providers to accommodate greater traffic to grow services—and revenue—and helping them to prepare for 5G.18 19 20

Next-Generation Platform for Cloud-Optimized, 5G-Ready Networks, and Next-Generation Virtual Networks
The coming era of 5G will enable entirely new ecosystems and classes of consumer and enterprise services along with media applications on wireless and wireline networks. These data-rich, innovative use cases, driven by the new Internet of Things (IoT), visual computing, and analytics, represent significant future opportunities for communications service providers (CommSPs) to grow revenue. 

The transition from purpose-built, fixed function infrastructure to a new generation of open networks is the essential first step to prepare for a 5G-enabled world. Software-defined networking with Network Functions Virtualization (NFV) is enabling new service opportunities and operations efficiencies for both communication services providers and enterprises alike. Using flexible, optimized, industry-standard servers and virtualized, orchestrated network functions will allow future-ready infrastructures to be able to deliver innovative services with efficiency and ease. 

Such distributed communications networks can support extreme levels of scalability, agility, programmability, and security across an ever-growing volume and variety of networking workloads—from the network core to the edge. 

The Intel Xeon Scalable platform is the basis for next-generation platforms to build virtualized, cloud-optimized, 5G-ready networks. It offers an architecture that scales and adapts with ease to handle the demands of emerging applications and the convergence of key workloads, such as applications and services, control plane processing, high-performance packet processing, and signal processing. This new processor provides a foundation for agile networks that can operate with cloud economics, be highly automated and responsive, and support rapid and more secure delivery of new and enhanced services enabled by 5G.

Highlights for CommSP innovation:

  • Hardware-based acceleration of encryption and compression using integrated Intel® QAT
  • Integrated Intel Ethernet
  • Intel FPGAs maximize versatility in communications infrastructure
  • Intel KPT with Integrated Intel QAT and Intel PTT
  • Intel TXT with One-Touch Activation

Additional Resources Optimized for CommSPs
The open source Data Plane Development Kit (DPDK) enables optimized communications operations on Intel Architecture. DPDK has demonstrated ability to scale performance as processor core count and performance increase; workloads, such as Vector Packet Processing (VPP) IPSec, benefit from this enhanced performance. Additionally, these libraries provide pre-optimized mechanisms to allow new processor capabilities (such as Intel AVX-512 and memory and I/O enhancements) to be able to utilize the new functionality for improved packet processing performance with less  direct development effort.

Intel offers programs, such as Intel® Network Builders University, ideal for network evolution in the 5G era. With solution guidance and training from these programs, CommSPs can drive their network transformation initiatives forward with increased confidence.

Across 13 common HPC workloads, Intel® Xeon® Scalable platform delivers improved performance.

Breakthrough HPC and High-Performance Data Analytics Innovation
Today’s scientific discoveries are fueled by innovative algorithms, new sources and volumes of data, and advances in compute and storage. Benefitting from exponentially expanding volumes and variety of data, HPC clusters are also the engine for running evolving High-Performance Data Analytics (HPDA) workloads, leading to incredible discoveries and insight for business and human understanding. Machine learning, deep learning, and AI converge the capabilities of massive compute with the flood of data to drive next-generation applications, such as autonomous systems and self-driving vehicles.

Intel Xeon Scalable platform offers a common platform for AI with high throughput for both inference and training—up to 18x higher inference21 and up to 19x higher training22 throughput, compared to four-year-old systems.

HPC is no longer just the domain of large scientific institutions. Enterprises are increasingly consuming a massive number of HPC compute cycles; some of the world’s largest HPC clusters are in private oil and gas companies. Research in personalized medicine applies HPC for highly focused treatment plans. New HPC installations are engaging innovative, converged architectures for non-traditional usages that combine simulation, AI, visualization, and analytics in a single supercomputer.

HPC platforms—from the smallest clusters to largest supercomputers—demand a balance across compute, memory, storage, and network. The Intel Xeon Scalable platform was designed to deliver and enable such balance with massive scalability—to tens of thousands of cores. From its improved core count and mesh architecture to newly integrated technologies and support for Intel Optane memory and storage devices, the Intel Xeon Scalable platform enables the ultimate goals of HPC—to maximize performance across compute, memory, storage, and network without inducing bottlenecks at any intersection of resources.

The integration of Intel Omni-Path Architecture, an end-to-end high-performance fabric, into the Intel Xeon Scalable platform delivers both increased performance and scaling to distributed, parallel computing clusters. Near linear scaling up to 32 nodes enables building large HPC solutions that are not inhibited by the interconnect. And, the Intel Xeon Scalable platform, compared to previous generation Intel Xeon processors, delivers 1.63x better overall performance across 13 common HPC workloads used throughout the sciences and financial services.23 The Intel Xeon Scalable platform and Intel Omni-Path Architecture can enable new discoveries and faster solutions for highly parallel workloads in many data centers.

Highlights for HPC innovation:

  • Intel Ultra Path Interconnect
  • Intel Advanced Vector Extensions 512
  • Integrated Intel Omni-Path Architecture Host Fabric Interface
  • Support for Intel Optane SSDs and Intel® 3D NAND Solid State Drives

Additional technologies for HPC, HPDA, and AI:

  • The bootable Intel® Xeon Phi™ processor 7200 Series, with its Many Integrated Core (MIC) architecture, is an ideal foundation for highly parallel workloads, such as machine learning training, simulation, and visualization.
  • A range of high-productivity software tools, optimized libraries, foundational building blocks, and flexible frameworks for general and highly parallel computing help simplify workflows and assist developers to create codes that maximize the capabilities of IA for HPC and AI.
  • Optimizations for popular deep learning frameworks for IA, including Neon,* Caffe,* Theano,* Torch,* and TensorFlow* offer increased value and performance for data scientists.
  • Intel® Parallel Studio XE 2017 includes performance libraries, such as Intel® Math Kernel Library for Deep Neural Networks (Intel® MKL-DNN) to accelerate deep learning frameworks on IA, and Intel® Data Analytics Acceleration Library (Intel® DAAL) to speed big data analytics.

Resources Optimized for HPC
To continue to advance discovery through HPC into the Exascale era, the Intel® Modern Code Developer Program offers developers and data scientists easily accessible online and face-to-face code modernization technical sessions on techniques, such as vectorization, memory and data layout, multi-threading, and multi-node programming.

Product and Performance Information

3

Up to 65% lower 4-year TCO estimate example based on equivalent rack performance using VMware ESXi* virtualized consolidation workload comparing 20 installed 2-socket servers with Intel® Xeon® processor E5-2690 (formerly “Sandy Bridge-EP”) running VMware ESXi* 6.0 GA using Guest OS RHEL* 6.4 compared at a total cost of $919,362 to 5 new Intel® Xeon® Platinum 8180 (Skylake) running VMware ESXi* 6.0 U3 GA using Guest OS RHEL* 6 64 bit at a total cost of $320,879 including basic acquisition. Server pricing assumptions based on current OEM retail published pricing for Broadwell based servers – subject to change based on actual pricing of systems offered.

4

Up to 4.2x more VMs based on server virtualization consolidation workload: Based on Intel® internal estimates 1-Node, 2 x Intel® Xeon® processor E5-2690 on Romley-EP with 256 GB Total Memory on VMware ESXi* 6.0 GA using Guest OS RHEL* 6.4, glassfish3.1.2.2, postgresql9.2. Data Source: Request Number: 1718, Benchmark: server virtualization consolidation, Score: 377.6 @ 21 VMs vs. 1-Node, 2 x Intel® Xeon® Platinum 8180 processor on Wolf Pass SKX with 768 GB Total Memory on VMware ESXi6.0 U3 GA using Guest OS RHEL* 6 64 bit. Data Source: Request Number: 2563, Benchmark: server virtualization consolidation, Score: 1580 @ 90 VMs. Higher is better.

5

Up to 5x claim based on OLTP Warehouse workload: 1-Node, 4 x Intel® Xeon® processor E7-4870 on Emerald Ridge with 512 GB Total Memory on Oracle Linux* 6.4 using Oracle 12c* running 800 warehouses. Data Source: Request Number: 56, Benchmark: HammerDB, Score: 2.46322e+006 Higher is better vs. 1-Node, 4 x Intel® Xeon® Platinum 8180 processor on Lightning Ridge SKX with 768 GB Total Memory on Red Hat Enterprise Linux* 7.3 using Oracle 12.2.0.1 (including database and grid) with 800 warehouses. Data Source: Request Number: 2542, Benchmark: HammerDB, Score: 1.2423e+007 Higher is better.

6

Up to 1.52x claim based on HammerDB*: 1-Node, 4 x Intel® Xeon® processor E7-8890 v4 on Brickland (Patsburg) with 1536 GB Total Memory on Oracle Linux* 7.1 using Oracle* 12.1.0.2.0 (including database and grid) with 800 warehouses. Data Source: Request Number: 2239, Benchmark: HammerDB, Score: 8.17145e+006 vs. 1-Node, 4 x Intel® Xeon® Platinum 8180 processor on Lightning Ridge SKX with 768 GB Total Memory on Red Hat Enterprise Linux* 7.3 using Oracle 12.2.0.1 (including database and grid) with 800 warehouses. Data Source: Request Number: 2542, Benchmark: HammerDB, Score: 1.2423e+007 Higher is better.

7

Up to 1.5x more VMs based on virtualization consolidation workload : 1-Node, 2 x Intel® Xeon® processor E5-2699 v4 on Grantley-EP (Wellsburg) with 512 GB Total Memory on VMware ESXi* 6.0 Update 1 using Guest VM’s utilize RHEL 6 64bit OS. Data Source: Request Number: 1637, Benchmark: server virtualization workload, Score: 1034 @ 58 vs. 1-Node, 2 x Intel® Xeon® Platinum 8180 processor on Wolf Pass SKX with 768 GB Total Memory on VMware ESXi6.0 U3 GA using Guest VM’s utilize RHEL 6 64bit OS. Data Source: Request Number: 2563, Benchmark: server virtualization workload, Score: 1580 @ 90 VMs Higher is better.

8

Up to 2.7x claim based on DPDK L3 Packet Forwarding: Intel® Xeon® E5-2650 processors 2.00GHz, 8 GT/s QPI, 20MB L3 cache, Patsburg Chipset (C0 stepping), Grizzly Pass Platform (R2216GZBPP), DDR3 1333MHz, 8 x dual rank 4GB (total 32GB), 4 memory channels per socket Configuration, 1 DIMM per channel, 6 x Intel® 82599 dual-port PCI-Express* Gen2 x8 10 Gb Ethernet NIC, 1 x Intel® 82599 dual-port Gen2 x8 I/O expansion module10 Gb Ethernet NIC, BIOS version S5500.86B.01.00.0048, Operating system: Fedora Core* 15, Kernel version: 2.6.38.4, IxNetwork* 6.0.400.22, DPDK version: FD5_1 Score: 102Gbits/s packet forwarding at 256B using cores vs. Gold 6152: Estimates based on Intel internal testing on Intel® Xeon® 6152 2.1 GHz, 2x Intel®, FM10420(RRC) Gen Dual Port 100GbE Ethernet controller (100Gbit/card) 2x Intel® XXV710 PCI Express Gen Dual Port 25GbE Ethernet controller (2x25G/card), DPDK 17.02. Score: 281 Gbits/s packet forwarding at 256B packet using cores, IO and memory on a single socket.

9

Up to 1.7x claim based on DPDK L3 Packet Forwarding: E5-2658 v4: 5 x Intel® XL710-QDA2, DPDK 16.04. Benchmark: DPDK l3fwd sample application Score: 158 Gbits/s packet forwarding at 256B packet using cores. Gold 6152: Estimates based on Intel internal testing on Intel® Xeon® 6152 2.1 GHz, 2x Intel®, FM10420(RRC) Gen Dual Port 100GbE Ethernet controller (100Gbit/card) 2x Intel® XXV710 PCI Express* Gen Dual Port 25GbE Ethernet controller (2x25G/card), DPDK 17.02. Score: 281 Gbits/s packet forwarding at 256B packet using cores, IO and memory on a single socket.

10

Up to 8.2x claim based on Intel® Distribution for LINPACK* Benchmark: 1-Node, 2 x Intel® Xeon® processor E5-2690 on Intel® Server Board S2600CP2 with 32 GB Total Memory on Red Hat Enterprise Linux* 6.0 (Santiago) kernel version 2.6.32-504.el6.x86_64 using Intel® Distribution for LINPACK Benchmark using 56000 problem size. Score: 366.0 GFLOPS/s vs. 1-Node, 2 x Intel® Xeon® Platinum 8180 processor on Purley-EP (Lewisburg) with 192 GB Total Memory on Ubuntu 17.04 using MKL 2017 Update 2. Data Source: Request Number: 2535, Benchmark: Intel® Distribution for LINPACK Benchmark, Score: 3007.8 GFLOPS/s Higher is better.

11

Up to 2.2x claim based on Intel® Distribution for LINPACK* Benchmark: 1-Node, 2 x Intel® Xeon® processor E5-2699 v4 on Grantley-EP (Wellsburg) with 64 GB Total Memory on Red Hat Enterprise Linux* 7.0 kernel 3.10.0-123 using MP_LINPACK 11.3.1 (Composer XE 2016 U1). Data Source: Request Number: 1636, Benchmark: Intel® Optimized MP LINPACK*, Score: 1446.4 GFLOPS/s vs. 1-Node, 2 x Intel® Xeon® Platinum 8180 processor on Purley-EP (Lewisburg) with 192 GB Total Memory on Ubuntu* 17.04 using MKL 2017 Update 2. Data Source: Request Number: 2535, Benchmark: Intel® Distribution for LINPACK Benchmark, Score: 3007.8 GFLOPS/s Higher is better.

12

More inter-CPU bandwidth. 10.4 GT/s on Intel® Xeon® Scalable processors vs. 9.6 GT/s on Intel® Xeon® processor E5-2600 v4 product family.

13

As measured by Intel® Xeon® processor Scalable family with Intel® AVX-512 compared to an Intel® Xeon® E5 v4 with Intel® AVX2.

14

BigBench*, Near Zero encryption overhead: BigBench query Runtime/second. Testing done by Intel.  BASELINE: Platform 8168, NODES 1 Mgmt + 6 Workers, Make Intel Corporation, Model S2600WFD, Form Factor 2U, Processor Intel® Xeon® Platinum 8168, Base Clock 2.70 GHz, Cores per socket 24, Hyper-Threading Enabled, NUMA mode Enabled, RAM 384 GB DDR4, RAM Type 12x 32 GB DDR4, OS Drive Intel® SSD DC S3710 Series (800 GB, 2.5 in SATA 6Gb/s, 20nm, MLC), Data Drives 8x – Seagate Enterprise 2.5 HDD ST2000NX0403 2 TB, Intel® SSD DC P3520 Series (2.0TB), Temp Drive DC 3520 2 TB, NIC Intel  X722 10 GBE – Dual Port, Hadoop Cloudera* 5.11, Benchmark TPCx-BB 1.2, Operating System CentOS* Linux release 7.3.1611 (Core); HDFS encryp-tion turned OFF. vs. NEW: Platform 8168, NODES 1 Mgmt + 6 Workers, Make Intel Corporation, Model S2600WFD, Form Factor 2U, Processor Intel® Xeon® Platinum 8168, Base Clock 2.70 GHz, Cores per socket 24, Hyper-Threading Enabled, NUMA mode Enabled, RAM 384 GB DDR4, RAM Type 12x 32 GB DDR4, OS Drive Intel® SSD DC S3710 Series (800 GB, 2.5 in SATA 6 Gb/s, 20nm, MLC), Data Drives 8x – Seagate Enterprise  2.5 HDD ST2000NX0403 2 TB, Intel® SSD DC P3520 Series (2.0 TB), Temp Drive DC 3520 2 TB, NIC Intel X722 10 GBE – Dual Port, Hadoop Cloudera* 5.11, Benchmark TPCx-BB 1.2, Operating System CentOS* Linux release 7.3.1611 (Core); HDFS encryption turned ON.

15

3.1x, 1.2x, 2x ISA-L configuration. Intel® Xeon® processor Scalable family: Intel® Xeon® Platinum 8180 processor, 28C, 2.5 GHz, H0, Neon City CRB, 12x16 GB DDR4 2666 MT/s ECC RDIMM, BIOS PLYCRB1.86B.0128.R08.1703242666.

Intel® Xeon® E5-2600v4 series processor, E5-2650 v4, 12C, 2.2 GHz, Aztec City CRB, 4x8 GB DDR4 2400 MT/s ECC RDIMM, BIOS GRRFCRB1.86B.0276.R02.1606020546

Operating System: Redhat Enterprise Linux* 7.3, Kernel 4.2.3, ISA-L 2.18, BIOS Configuration, P-States: Disabled, Turbo: Disabled, Speed Step: Disabled, C-States: Disabled, ENERGY_PERF_ BIAS_CFG: PERF.

16

As measured by Intel® Xeon® processor Scalable family with Intel® FPGA optimized workload and Intel® Xeon® processor Scalable family without FPGA optimized workload.

17

Up to 1.58x claim based on Ghost-NodeJS workload: 1-Node, 2 x Intel® Xeon® processor E5-2699 v4 on Wildcat Pass with 384 GB Total Memory on Ubuntu* 16.04 LTS using Node.js version 6.9.2, MySQL* Maria DB version 15.1 Distrib 10.0.30. Data Source: Request Number: 2687, Benchmark: Ghost-NodeJS, Score: 2308 Higher is better vs. 1-Node, 2 x Intel® Xeon® Platinum 8180 processor on Wolf Pass SKX with 384 GB Total Memory on Ubuntu 16.10 using Node.js version 6.9.2, MySQL Maria DB version 15.1 Distrib 10.0.30. Data Source: Request Number: 2687, Bench-mark: Ghost-NodeJS, Score: 3647 Higher is better.

18

Up to 1.77x claim based on DPDK L3 Forwarding: E5-2658 v4: 5 x Intel® XL710-QDA2, DPDK 16.04. Benchmark: DPDK l3fwd sample application. Vs. Gold 6152: Estimates based on Intel internal testing on Intel® Xeon® 6152 2.1 GHz, 2x Intel®, FM10420(RRC) Gen Dual Port 100GbE Ethernet controller (100Gbit/card) 2x Intel® XXV710 PCI Express* Gen Dual Port 25GbE Ethernet controller (2x25G/card), DPDK 17.02. Score: 281 Gbits/s packet forwarding at 256B packet using cores, IO and memory on a single socket.

19

Up to 2.5x claim based on DPDK IPSec Seg-gw benchmark: Intel® Xeon® E5-2658 v4, Intel® PCH C612, DDR4-2400 Intel® 895XCC based Quick Assist Accelerator Adapter PCIe* Gen3 x8 links, DPDK 16.11 IPSec-secgw,1420 B packet). Intel® DPDK 16.11 IPsec-secgw sample application. Cores, IO, packet buffer memory, and processing cores are on a single socket. 6 cores used on one Socket, Crypto algorithm: AES-128-CBC-HMAC-SHA1 vs. Intel® Xeon® 6152 2.1 GHz, 3x Intel® Corporation, Ethernet Controller X710 (4x10 Gbe ports per card), Lewisburg B0 Quick Assist Accel-erator with PCIe Gen3 x24 links, Intel® DPDK 17.02 IPsec-secgw,  Intel® QAT1.7.Upstream.L.1.0.0-15, 6 cores used. Cores, IO, packet buffer memory, and processing cores are on a single socket. 6 cores used on one Socket, Crypto algorithm: AES-128-CBC-HMAC-SHA1.

20

Up to 2.4x claim based on TLS Web Proxy using NGINX®: Intel® Xeon® E5-2658 v4, DDR4-2133, Intel® PCH C612, Intel® 895XCC based QuickAssist Accelerator Adapter PCIe* Gen3 x8 links, OpenSSL-Async (0.4.9-009) + NGINX-1.6.2 (0.1.0-008), QAT1.6.L.2.6.0-60. Cores, IO, packet buffer memory, and processing cores are on a single socket. 6 cores used on one Socket 12Cores are used, Crypto algorithm: AES-128-CBC-HMAC-SHA1 vs. Intel Xeon 6152 2.10 GHz, DDR4-2400 3x Intel® Corporation Ethernet Controller X710 (4 x10 Gbe ports per card), 1x Intel® Corpora-tion Ethernet Controller X710 (2 x10 Gbe ports per card), PCIe x16 to 2 x8 PCIe bifurcation plugin card, Lewisburg-L B1 QuickAssist Accelerator with PCIe Gen3 x24 links, Intel®OpenSSL-1.0.1u + NGINX-1.9.6, Intel® QAT1.7.Upstream.L.1.0.0-15. Cores, IO, packet buffer memory, and processing cores are on a single socket. 6 cores used on one Socket, 20Core are used. Crypto algorithm: AES-128-CBC-HMAC-SHA1.

21

Inference: Platform: 2S Intel® Xeon® Platinum 8180 CPU @ 2.50 GHz (28 cores), HT disabled, turbo disabled, scaling governor set to “performance” via intel_pstate driver, 384GB DDR4-2666 ECC RAM. CentOS* Linux release 7.3.1611 (Core), Linux kernel 3.10.0-514.10.2.el7.x86_64. SSD: Intel® SSD DC S3700 Series (800GB, 2.5in SATA 6Gb/s, 25nm, MLC). Performance measured with: Environment variables: KMP_AFFINITY=’granularity=fine, compact‘, OMP_NUM_THREADS=56, CPU Freq set with cpupower frequency-set -d 2.5G -u 3.8G -g performance.
Deep Learning Frameworks: Caffe: (http://github.com/intel/caffe/), revision f96b759f71b2281835f690af267158b82b150b5c. Inference measured with “caffe time --forward_only” command, training measured with “caffe time” command. For “ConvNet” topologies, dummy dataset was used. For other topologies, data was stored on local storage and cached in memory before training. Topology specs from https://github.com/intel/caffe/tree/master/models/intel_optimized_models (GoogleNet, AlexNet, and ResNet-50), https://github.com/intel/caffe/tree/master/models/default_vgg_19 (VGG-19), and https://github.com/soumith/convnet-benchmarks/tree/master/caffe/imagenet_winners (ConvNet benchmarks; files were updated to use newer Caffe prototxt format but are functionally equivalent). Intel C++ compiler ver. 17.0.2 20170213, Intel® Math Kernel Library (Intel® MKL) small libraries version 2018.0.20170425. Caffe run with “numactl l”.

22

Training: Platform: 2S Intel® Xeon® CPU E5-2697 v2 @ 2.70GHz (12 cores), HT enabled, turbo enabled, scaling governor set to “performance” via intel_pstate driver, 256GB DDR3-1600 ECC RAM. CentOS Linux release 7.3.1611 (Core), Linux kernel 3.10.0-514.21.1.el7.x86_64. SSD: Intel® SSD 520 Series 240GB, 2.5in SATA 6Gb/s, 25nm, MLC.

Performance measured with: Environment variables: KMP_AFFINITY=’granularity=fine, compact,1,0‘, OMP_NUM_THREADS=24, CPU Freq set with cpupower frequency-set -d 2.7G -u 3.5G –g performance.

Deep Learning Frameworks: Caffe: (http://github.com/intel/caffe/), revision b0ef3236528a2c7d2988f249d347d5fdae831236. Inference measured with “caffe time --forward_only” command, training measured with “caffe time” command. For “ConvNet” topologies, dummy dataset was used. For other topologies, data was stored on local storage and cached in memory before training. Topology specs from https://github.com/intel/caffe/tree/master/models/intel_optimized_models (GoogleNet, AlexNet, and ResNet-50), https://github.com/intel/caffe/tree/master/models/default_vgg_19 (VGG-19), and https://github.com/soumith/convnet-benchmarks/tree/master/caffe/imagenet_winners (ConvNet benchmarks; files were updated to use newer Caffe prototxt format but are functionally equivalent). GCC 4.8.5, Intel® Math Kernel Library (Intel® MKL) small libraries version 2017.0.2.20170110.

23

Up to 1.63x Gains based on Geomean of Weather Research Forecasting - Conus 12Km, HOMME, LSTCLS-DYNA Explicit, INTES PERMAS V16, MILC, GROMACS water 1.5M_pme, VASPSi256, NAMDstmv, LAMMPS, Amber GB Nucleosome, Binomial option pricing, Black-Scholes, Monte Carlo European options. Results have been estimated based on internal Intel analysis and are provided for informational purposes only. Any difference in system hardware or software design or configuration may affect actual performance. Software and workloads used in performance tests may have been optimized for performance only on Intel® microprocessors. Performance tests, such as SYSmark and MobileMark, are measured using specific computer systems, components, software, operations and functions. Any change to any of those factors may cause the results to vary. You should consult other information and performance tests to assist you in fully evaluating your contemplated purchases, including the performance of that product when combined with other products. For more information go to http://www.intel.com/performance/datacenter.