The Sun Fire X4450 server is optimized for world class performance, density and energy efficiency in a compact footprint. Sun's leadership in system design combined with the power of Quad-Core Intel Xeon processors, result in very space and energy-efficient 4-socket 2U server that sets the bar for the industry and runs Solaris OS, Linux, Windows, and VMware.
(Tue, 13 January 2009)
Designed To Do More: Leading Virtualization Performance and Power Consumption in ½ the Physical Footprint of Competitive Systems
The Sun Fire X4450 server, running VMware Virtual Infrastructure 3 software, posted the best VMmark score of 19.47@14 tiles among all 24-core results. This result represents the best posted score among all results based on Intel Xeon architecture. The Sun Fire X4450 server, equipped with four Intel Xeon X7460 processors and six processing cores each, provides a highly scalable virtualization platform, that, in combination with the VMware ESX Server 3.5.0 hypervisor, takes full advantage of Intel Virtualization Technology.
Today, nearly every IT organization is using, or planning to use, virtualization technology to run multiple operating systems, consolidate applications and increase server utilization. Virtualization allows for reduction in datacenter space utilization, cooling requirements and system management costs while increased server utilization has an added benefit of bringing the server's power supplies into optimal power efficiency range.
Currently there are no benchmarks or standard tests that allow comparison of server power consumption in a virtualized environment. To capture the true power consumption of a server you must measure it with the actual customer workload in a real environment. That is not always practical or possible. To aid customers in their choices, Sun has been reporting power measurements on critical benchmarks and posting that data side-by-side with the performance numbers. The power consumption measured, while running VMmark, was 933W on average, and was captured during the steady state of the benchmark.
VMmark is a free tool developed by VMware Inc. as a standard methodology for comparing virtualized systems. The benchmark allows hardware vendors, virtualization software vendors and other organizations to measure the overall performance and scalability of individual applications running in virtualized x86 environments. The benchmark suite captures some of today's most commonly executed workloads such as e-mail server, Java-based application server, web server, commercial database, file server, and an idle standby server. These six diverse workloads run concurrently on top of their respective operating system instances and are organized in a unit of work called a tile. Each virtual machine in a tile is tuned to use only a fraction of the system's total resources thus requiring the execution of multiple tiles simultaneously to achieve peak performance. With 14 tiles, the Sun Fire X4450 server effectively supported 84 virtual machines in 2 rack units (RU) of physical space.
Compared to the similarly equipped Dell PowerEdge R900 and HP ProLiant DL 580 G5 rack-mount servers, with the posted VMmark scores of 18.69@14 tiles and 18.56@14 tiles respectively, the Sun Fire X4450 server used up to 60 percent less system memory and occupied only half of the physical space (2 RU).
Additionally, Sun's server relied on very efficient and reliable external storage arrays—the Sun StorageTek 2540 Fibre Channel arrays with Sun StorageTek 2501 SAS expansion units. This family of arrays offer enterprise-class, reliable RAID-protected functionality in a high-density 2 RU package and is well suited for virtualization environments.
The combination of superior performance, best-in-the-industry physical footprint and leading power consumption positions the Sun Fire X4450 server as the best platform for virtualization workloads, allowing enterprises to consolidate their datacenters from both software and hardware perspectives, while saving a bundle on operational costs.
Sun's dense 2RU Intel Xeon-based Sun Fire X4450 server, which was selected as a finalist in the Intel Server and Storage Innovation Award ceremony at the Fall 2008 Intel Developer Forum in San Francisco, was also featured in the event's keynote as the best four-socket performer in the world on the SPECjbb2005 benchmark. The server, equipped with four Intel Xeon X7460 processors, features 24 processing cores and runs the latest version 1.6.0_06 Performance Release of Sun's Java Platform, Standard Edition software on top of the Solaris 10 Operating System (OS).
The Sun Fire X4450 server obtained a World Record score of 531,669 bops (132,917 SPECjbb2005 bops/JVM) for all four-socket systems, demonstrating an almost 15% improvement over the previous best result. The SPECjbb2005 benchmark emulates the design of real-world server-side Java applications and provides an accurate reflection of the business logic and objects, while stressing the implementation of the Java Virtual Machine (JVM) and the scalability of the system's processors and memory.
This benchmark result demonstrates the benefits of collaboration efforts between Sun and Intel where the enterprise-class, mission-critical Solaris 10 10/08, the latest update to the Solaris 10 OS, is enhanced with new capabilities and is specifically optimized for Intel's new Xeon processor 7400 series.
Record Breaking Single JVM x86 Performance on SPECjbb2005 Benchmark
As demonstrated by the SPECjbb2005 benchmark, the Sun Fire X4450 server obtained a single JVM x86 World Record score of 448,262 SPECjbb2005 bops (448,262 SPECjbb2005 bops/JVM). The SPECjbb2005 benchmark actually produces two equally important metrics — total system throughput (SPECjbb2005 bops) and JVM scalability (SPECjbb2005 bops/JVM), reflecting how different organizations may deploy their applications.
Running a single instance of the Java Virtual Machine per physical server continues to provide an effective way to isolate multiple Running a single instance of the Java Virtual Machine per physical server continues to provide an effective way to isolate multiple tasks in the environments where dedicated solutions are necessitated by business or regulatory requirements. Moreover, this mode of operation guarantees uniformed system utilization often desired by Enterprise datacenter managers and users alike. The Sun Fire X4450 server, equipped with four Intel Xeon X7460 processors, runs the latest version 1.6.0_06 Performance Release of Sun's Java Platform, Standard Edition software and is well suited for single JVM deployments.
The SPECjbb2005 benchmark emulates the design of real-world server-side Java applications and provides an accurate reflection of the business logic and objects, while stressing the implementation of the Java Virtual Machine (JVM) and the scalability of the system's processors and memory.
Powered by Solaris 10 10/08, the latest update to the Solaris 10 Operating System (OS), Sun's flagship Intel Xeon-based Sun Fire X4450 server delivers top-shelf single-JVM performance, showcasing new capabilities, including performance and energy efficiency enhancements, specifically optimized for Intel's new Xeon processor 7400 series.
World Record for 4-Processor Systems Running Unix on 2-Tier SAP Sales and Distribution (SD) Standard Application Benchmark
The Sun Fire X4450 server with four Intel Xeon processor X7460 (24 cores, 24 threads), running the SAP ERP application Release 6.0 with Unicode software and MaxDB 7.6 database on top of the Solaris 10 OS, was able to support 4,600 SAP SD Benchmark users, as of September 12, 2008.
SAP is one of the premier worldwide ERP application providers, and maintains a suite of benchmark tests to demonstrate the performance of competitive systems on the various SAP products. The SAP Standard Application Sales and Distribution (SD) Benchmark is a two-tier ERP business test that represents full business workloads of order processing and invoice processing, while demonstrating the ability to run both the application and database software on a single system. The SAP Standard Application SD Benchmark represents the critical tasks performed in real-world ERP business environments.
This new four-processor UNIX World record relies on the Unicode version of SAP software to deliver the breakthrough result and highlights the optimal performance of SAP ERP applications on Sun Fire servers running the latest version of the Solaris 10 OS. Unicode enables seamless multilingual support when running various SAP applications and requires a different version of the software. Sun has pioneered the usage of Unicode-based SAP software in several benchmark publications to date and has gained significant expertise in lowering the additional resource consumption normally associated with Unicode software, limiting the performance difference between Unicode and non-Unicode versions to 14%.
The benchmark results fully comply with the SAP Benchmark Council regulations and have been audited and certified by SAP AG. SAP certification number 2008051.
[1] Two-tier SAP SD Standard Application Benchmark, as of September 15, 2008: 4 Intel Xeon Processor X7460, 24 Cores / 24 Threads, Solaris 10, MaxDB 7.6, SAP ERP 6.0 with Unicode. Results: 4,600 SAP SD Benchmark users, 23,120 SAPS, certification number 2008051. IBM System x3755, 4 AMD Opteron model 8356, 16 cores / 16 threads, Windows Server 2003 Enterprise Edition, DB2 9.5, SAP ECC 6.0. Results: 3,540 SAP SD Benchmark users, 17,720 SAPS. Certification number 2008032. HP DL585 G5, 4 AMD Opteron model 8360SE, 16 cores / 16 threads, Windows Server 2003 Enterprise Edition, SQL Server 2005, SAP ECC 6.0. Results: 3,540 SAP SD Benchmark users, 19,020 SAPS. Certification number 2008041. IBM System p550, 4 POWER6 4.2GHz , 8 cores / 16 threads, Red Hat Enterprise Linux 5, DB2 9.5 , SAP ECC 6.0. Results: 3,104 SAP SD Benchmark users, 15,630 SAPS. Certification number 2008002
(Wed, 10 September 2008)
Industry-Leading General-Purpose Java Server — The Sun Fire X4450
The Sun Fire X4450 server, equipped with four Intel Xeon X7460 processors, leads the pack on SPECjvm2008, a general-purpose multi-threaded Java benchmark. Sun continues to demonstrate performance leadership on SPECjvm2008 with each new benchmark submission, while relentlessly pushing the performance of the Java platform forward.
SPECjvm2008 is a multi-threaded Java benchmark suite measuring the performance of a Java Runtime Environment (JRE). The benchmark focuses on executing a single application and reflects the performance of the operating system, the underlying processor and the memory subsystem as the system executes the JRE. This multi-threaded Java benchmark stresses specific aspects of JVM performance, such as threading, locks, just-in-time (JIT) compiler, and the virtual memory management system. It contains a broad collection of real-world applications and is applicable to measuring general-purpose Java performance on a wide variety of both client and server systems.
Benchmark Outcome
The Sun Fire X4450 server, equipped with six-core Intel Xeon processors, obtained a world record score of 283.79 SPECjvm2008 Base ops/m, the highest result ever posted on this benchmark.
The Sun Fire server features 24 processing cores and runs the latest version of Sun's Java Platform, Standard Edition software (1.6.0_06 Performance Release) on top of the Solaris 10 Operating System (OS).
Sun was the first vendor to publish on this benchmark and this record score further showcases it leadership in Java application performance.
The winning combination of Sun's x64 systems, Solaris OS and open-sourced Java HotSpot software, offer a proven and stable environment for customer deployments that rely on JRE to execute a single application on a given system.
Leading 4-Socket Performance on SPECjbb2005 Benchmark
Sun and Intel continue to reap the benefits of collaboration efforts that advance Solaris, an enterprise class, mission critical UNIX OS, and Intel Xeon processor-based systems further into the datacenter. Sun's flagship Intel Xeon-based Sun Fire X4450 server, equipped with four Quad-core Intel Xeon X7350 processors, delivered top-notch four-socket performance on SPECjbb2005 benchmark that emulates the design of real-world server-side Java applications and provides an accurate reflection of the business logic and objects, while stressing the implementation of the JavaVirtual Machine (JVM) and the scalability of the system's processors and memory.
Fueled by the latest version 1.6.0_06-p of Sun's Java Platform, Standard Edition software, the compact Sun Fire X4450 server obtained a World Record score of 464,355 SPECjbb2005 bops (58,044 SPECjbb2005 bops/JVM) for all four-socket systems. This result bests the previous record held by the Dell PowerEdge R900 server by 4% and accomplishes that in half the physical space. The table below compares Sun's result to the top scores posted by the competition.
SPECjbb2005 4-Chip Results by Various Competing Platforms
Sun Fire X4450
IBM p 570
Dell R900
HP ProLiant ML570 G4
Space (RU)
2
8
4
6
Chips/Cores
4/16
4/8
4/16
4/8
Operating System
Solaris 10
AIX 5L
Windows 2003
Windows 2003
JVM
Java HotSpot(TM) 1.6.0_06-p
J2RE 1.5.0
BEA JRockit(R) 6.0 P27.4.0
BEA JRockit(R) 5.0 P27.1.0
Performance (SPECjbb2005 bops)
464,355
346,742
446,209
217,334
Performance per JVM (SPECjbb2005 bops/JVM)
58,044
86,686
55,776
54,334
Performance / RU (Higher is Better)
232,167
43,343
111,552
36,222
This table clearly demonstrates the leading performance across the board and signifies the optimization work that has been ongoing since Sun and Intel entered the landmark agreement.
Record Breaking Single JVM x86 Performance for SPECjbb2005 Benchmark
When deploying their Java applications many organizations and users choose to run a single instance of the Java Virtual Machine per physical server. That creates a dedicated solution, allows for an additional degree of workload isolation and often simplifies system management while keeping system utilization at uniform level. Interestingly, the SPECjbb2005 benchmark, actually produces two equally important metrics - total system throughput (SPECjbb2005 bops) and JVM scalability (SPECjbb2005 bops/JVM), reflecting how different organizations may deploy their applications.
As demonstrated by a single JVM x86 World Record score of 389,208 SPECjbb2005 bops (389,208 SPECjbb2005 bops/JVM) the latest version 1.6.0_06-p of Sun's Java Platform, Standard Edition software in combination with the Sun Fire X4450 server, are well suited for single JVM deployments.
The SPECjbb2005 benchmark emulates the design of real-world server-side Java applications and provides an accurate reflection of the business logic and objects, while stressing the implementation of the Java Virtual Machine (JVM) and the scalability of the system's processors and memory.
Powered by Solaris, an enterprise class, mission critical UNIX OS, Sun's flagship Intel Xeon-based Sun Fire X4450 server, equipped with four Quad-core Intel Xeon X7350 processors, was able to deliver top-shelf single-JVM performance, showcasing the extensive software optimization work that has been ongoing between Sun and Intel.
Virtualization Redefined: Leading Performance and Power Consumption in ½ the Space of Competitive Systems
Nearly every IT organization is using, or planning to use, virtualization technology to run multiple operating systems, consolidate applications and increase server utilization. Virtualization allows for reduction in datacenter space utilization, cooling requirements and system management costs while increased server utilization has an added benefit of bringing the server's power supplies into optimal power efficiency range.
However, at present, there are no benchmarks or standard tests that allow comparison of server power consumption in a virtualized environment. To capture true power consumption of a server is to measure it with the actual customer workload in a real environment, and that is not always practical or possible. To aid customer in their choices, Sun has been reporting power measurements on critical benchmarks and posting that data side-by-side with the performance numbers. Today, Sun is the first vendor to disclose actual power consumption measured while running VMmark, the industry's most popular virtualization workload. The Sun Fire X4450 server, running VMware Virtual Infrastructure 3 software, posted one of the best scores among all 16 core results - 12.23 @ 8 tiles, with an average power consumption of only 830W measured during the steady state of the benchmark.
VMmark is a free tool developed by VMware Inc. as a standard methodology for comparing virtualized systems. The benchmark allows hardware vendors, virtualization software vendors and other organizations to measure the overall performance and scalability of individual applications running in virtualized x86 environments. The benchmark suite captures some of today's most commonly executed workloads such as e-mail server, Java-based application server, web server, commercial database, file server, and an idle standby server. These six diverse workloads run concurrently on top of their respective operating system instances and are organized in a unit of work that is called a tile. Each virtual machine in a tile is tuned to use only a fraction of the system's total resources thus requiring the execution of multiple tiles simultaneously to achieve peak performance.
The Sun Fire X4450 server, equipped with four Quad-Core Intel Xeon X7350 processors, provides a highly scalable virtualization platform, that, in combination with the VMware ESX Server 3.5.0 hypervisor, takes full advantage of Intel Virtualization Technology.
Additionally, it's worth mentioning that compared to the similarly equipped Dell PowerEdge R900 and HP ProLiant DL 580 G5 rack-mount servers, with the posted VMmark scores of 12.23 @ 8 tiles and 11.54 @ 8 tiles, respectively, the Sun Fire X4450 server can support up to 48 fully fledged VM instances (8 tiles) per server while occupying only half of physical space (2 RU). Furthermore, to support 8 tiles per server, efficient and reliable external storage is required. To satisfy that need, the Sun StorageTek 2540 Fibre Channel arrays with Sun StorageTek 2501 SAS expansion units were used. This family of arrays offers enterprise-class, reliable RAID-protected functionality in a high-density 2 RU package and is well suited for virtualization environments.
Best Integer Performance ever Recorded on SPEC CPU2006 Benchmark
The Sun Fire X4450 server is the world's fastest system on the integer suite of the SPEC CPU2006 benchmark. Based on real world applications, the SPECint2006 suite of the benchmark measures the performance of the processor, memory and compiler on the tested system by executing a range of integer-intensive tasks and capturing the time it takes to complete each of them. Then a geometric mean of all integer benchmarks is calculated where a higher score means "better performance" on the given workload. The Sun Fire X4450 server, equipped with the Quad-Core Intel X7350 processors, produced World Record result with an attention-grabbing SPECint2006 score of 25.0. This result was achieved using Intel's compiler version 10.1 and the systems was running SUSE Linux Enterprise Server 10 operating system.
X64 Leader in Floating Point Performance on SPEC CPU2006 Benchmark
The Sun Fire X4450 server is the fastest x64 system on the floating point suite of the SPEC CPU2006 benchmark. Based on real world applications, the SPECfp2006 suite of the benchmark measures the performance of the processor, memory and compiler on the tested system by executing a range of floating point-intensive tasks and capturing the time it takes to complete each of them. Then a geometric mean of all floating point benchmarks is calculated where a higher score means "better performance" on the given workload. The Sun Fire X4450 server, equipped with the Quad-Core Intel X7350 processors, produced x86 World Record result with an attention-grabbing SPECfp2006 score of 21.7. This result was achieved using Intel's compiler version 10.1 and the systems was running SUSE Linux Enterprise Server 10 operating system.
The Standard Performance Evaluation Corporation (SPEC) is a non-profit corporation formed to establish, maintain and endorse a standardized set of relevant benchmarks that can be applied to the newest generation of high-performance computers. SPEC develops suites of benchmarks and also reviews and publishes submitted results from their member organizations and other benchmark licensees.
SAP Standard Application Benchmarks test and prove the scalability of mySAP Business Suite. The benchmark results provide basic sizing recommendations for customers by testing new hardware, system software components, and Relational Database Management Systems (RDBMS). They also allow for the comparison of different system configurations. The benchmarking procedure is standardized and well defined. It is monitored by the SAP Benchmark Council made up of representatives of SAP and technology partners involved in benchmarking. The SAP Standard Application Benchmarks can also be used to test and verify scalability, concurrency and multi-user behavior of system software components, RDBMS, and business applications. All performance data relevant to system, user, and business applications are monitored during a benchmark run and can be used to compare platforms and as basic input for sizing recommendations.
Sun, Sun Microsystems, the Sun logo, Java, J2EE, Sun Fire and The Network Is The Computer are trademarks or registered trademarks of Sun Microsystems, Inc. in the United States and other countries. SPEC and the benchmark names SPECweb, SPECint, SPEComp, SPECfp, SPECjAppServer, SPECjvm, SPECpower, SPECmail, SPECsfs and SPECjbb are registered trademarks of the Standard Performance Evaluation Corporation. Sun's results have been submitted to SPEC. Competitive data obtained from http://www.spec.org as of the date located next to the respective claim. See the website for latest results. For comparison purposes, the terms CPU, chip and processor are used interchangeably. Each socket can accommodate one chip. SAP, R/3, mySAP reg TM of SAP AG in Germany and other countries. For the latest results and additional information visit www.sap.com/benchmark. TPC Benchmark C, tpmC, TPC-C, TPC Benchmark H, TPC-H, QphH are trademarks of the Transaction Performance Processing Council (TPC). More info http://www.tpc.org.