Hyper-Threading Technology on the Intel®XeonTMProcessor Family for ServersOffering increased server performance through on-processor thread-level parallelismIntel inCommunicationsWhite PaperHyper-Threading Technology on the Intel®XeonTMProcessor Family for ServersWhite PaperHyper-Threading Technology on the Intel®XeonTMProcessor Family for ServersContentsIntroduction 3Overview of Hyper-Threading Technology 4Resource Utilization 5Microarchitectural Details 6How Server Platforms Use Hyper-Threading Technology 8How Software Takes Advantage of Hyper-Threading Technology 9Business Benefits of Hyper-Threading Technology 10Summary 103White PaperHyper-Threading Technology on the Intel®XeonTMProcessor Family for ServersIntroductionHyper-Threading Technology is a groundbreaking innovationfrom Intel that enables multithreaded server software appli-cations to execute threads in parallel within each processorin a server platform. The Intel®XeonTMprocessor family usesHyper-Threading Technology, along with the Intel®NetBurstTMmicroarchitecture, to increase compute power and through-put for today’s Internet, e-Business, and enterprise serverapplications. This level of threading technology has neverbeen seen before in a general-purpose microprocessor.Hyper-Threading Technology helps increase transaction rates,reduces end-user response times, and enhances businessproductivity providing a competitive edge to e-Businessesand the enterprise. The Intel Xeon processor family for serversrepresents a significant leap forward in processor designand performance by being the first Intel®-based processorto support thread-level parallelism on a single processor.Intel’s commitment to providing businesses with world-class performance is long-standing. Each new processorintroduction and microarchitecture shift has resulted in significant performance advances. The Intel Xeon processorfamily for servers has changed the landscape of processordesign and performance by implementing support for multiplesoftware threads on each processor core. Hyper-ThreadingTechnology provides thread-level-parallelism (TLP) on eachprocessor resulting in increased utilization of processorexecution resources.With processor and application parallelism becoming moreprevalent, today’s server platforms are increasingly turning tothreading as a way of increasing overall system performance.Server applications have been threaded (split into multiplestreams of instructions) to take advantage of multiple processors. Multithreaded applications are designed to be separated into distinct execution streams, or threads.These applications make the best use of multiple processorsbecause they are designed with their own execution andperformance in mind rather than relying on operating systemsor compilers. For example, most Java Virtual Machine*implementations are multithreaded, as many of the administrative tasks are separate threads. Multi-processing-aware operating systems can schedule these threads forprocessing in parallel, across multiple processors within theserver system. These same applications can run unmodifiedon the Intel Xeon processor family for servers and takeadvantage of thread-level-parallelism on each processor in the system. Hyper-Threading Technology complementstraditional multi-processing by offering greater parallelismand performance headroom for threaded software. Dual-processing server applications in the areas of Webserving, search engines, security, streaming media, depart-mental or small business databases, and e-mail/file/print canrealize benefits from Hyper-Threading Technology usingIntel Xeon processor-based servers. Multi-processing serverapplications in the areas of customer relationship manage-ment, media servers, site management, enterprise data-bases, business intelligence, collaboration/e-mail and supplychain management can also reap tangible benefits fromHyper-Threading Technology using the Intel Xeon processorMP in 4-way, 8-way and above servers. All of these typesof solutions, and more, can take advantage of the fasterresponse times and higher transaction rates offered byHyper-Threading Technology. As software applications continue to be optimized to take greater advantage ofprocessor parallelism, Hyper-Threading Technology provides an additional boost for newer capabilities and the growing needs of today’s businesses.This paper explains how Hyper-Threading Technology workswithin the processor, and describes how server platforms,operating systems and server applications take advantage ofit. While this focuses exclusively on servers, Hyper-ThreadingTechnology is not restricted to servers. Intel has also intro-duced Hyper-Threading Technology into workstations,business desktops and consumer desktops. For a discus-sion of how Hyper-Threading Technology is implementedon workstations, and the benefits realized, please refer to the “Intel’s Hyper-Threading Technology: DeliveringPerformance Boosts for Workstations” document.Overview of Hyper-Threading TechnologyHyper-Threading Technology is a form of simultaneousmultithreading technology (SMT), where multiple threads of software applications can be run simultaneously on one processor. This is achieved by duplicating the archi-tectural state on each processor, while sharing one set of processor execution resources. The architectural statetracks the flow of a program or thread, and the executionresources are the units on the processor that do the work: add, multiply, load, etc.Let’s use a manufacturing plant as an analogy of a processor.In a traditional manufacturing plant there are multiple assem-bly lines (execution resources) and one supervisor (architec-tural state) that keeps track of orders (threads) and keepsthem moving through the plant. Our supervisor can keep theassembly lines within the plant working in parallel, but ifhe/she has to wait for a part, an assembly line may slowdown while waiting for the part as well. In the case of a proces-sor, an execution unit may have to wait for data from disk.With Hyper-Threading Technology, each manufacturing plantis enabled to have two supervisors (or two architectural states).They share the same common set of assembly lines withinthe plant. When one of the supervisors is waiting for a partto be delivered, the other supervisor can make use of theassembly lines – resulting in a more efficient, higher perform-ing manufacturing plant. Now the plant can turn out moregoods per hour since the assembly