January/April/October | Intel AS/400 RISC Server instruction

IBM Power Systems

Performance Capabilities Reference

IBM i operating system Version 6.1

January/April/October 2008

This

document is intended for use by qualified performance related programmers or analysts from IBM, IBM Business Partners and IBM customers using the IBM PowerTM Systems platform running IBM i operating system. Information in this document may be readily shared with IBM i customers to understand the performance and tuning factors in IBM i operating system 6.1 and earlier where applicable. For the latest updates and for the latest on IBM i performance information, please refer to the Performance Management Website: http://www.ibm.com/systems/i/advantages/perfmgmt/index.html

Requests for use of performance information by the technical trade press or consultants should be directed to Systems Performance Department V3T, IBM Rochester Lab, in Rochester, MN. 55901 USA.

IBM i 6.1 Performance Capabilities Reference - January/April/October 2008
© Copyright IBM Corp. 2008	IBM i Performance Capabilities	1

Image 1

Intel AS/400 RISC Server, 170 Servers, 7xx Servers manual January/April/October

Contents

January/April/October Toolcat Mkttools GET AS4ADV36 Package Table of Contents 107 Performance Observations and Tips117 System Application Server Instance 119 157 10.2 DB2 for i5/OS access with Odbc 155158 11.1 196 Supported Backup Device Rates 243199 200 IPL Performance 19.6 User Pool Faulting Guidelines 310 19.7 17.7 File Level Backup Performance 292284 285 334 Switchable IASP’s 327 Geographic Mirroring 329335 What the Estimator is Not 335 Disclaimer Notice Special Notices Ipds Purpose of this Document Introduction Page ISeries and AS/400 Risc Server Model Performance Behavior Interactive Indicators and MetricsOverview 3 V5R3 Disclaimer and Remaining Sections 4 V5R2 and V5R1 Server Model Behavior V4R5 V5R2Page 100 Server Model Differences Custom Server Model Performance Highlights of Model 7xx Servers Model 7xx and 9/98 Model Performance Highlights of Model 170 Servers Additional Server Considerations Performance Highlights of Custom Server Models Example Interactive Utilization Server Dynamic Tuning SDT Page Affects of Server Dynamic Tuning Managing Interactive Capacity Interactive/Server characteristics in the real world Is Interactive really Interactive? Page Migration from Traditional Models Number of interactive transactions per hour Interval end time hour and minute Upgrade Considerations for Interactive Capacity 13.2 V5R1 DSD Performance Behavior V5R1 Non-Domino Processing Domino-Complementary ProcessingInteractive Processing Running Linux on a Dedicated Server Logical Partitioning on a Dedicated ServerPage Effect of CPU Speed on Batch Batch PerformanceEffect of Dasd Type on Batch Batch Workload Description Expert Cache Tuning Parameters for BatchJob Priority Dynamic Priority SchedulingPage New for i5/OS V6R1 DB2 for i5/OS PerformanceI5/OS V6R1 SQE Query Coverage Query Attribute POWER5 Processor POWER6 Processor Records Inserted/Selected DB2 i5/OS V5R4 Highlights I5/OS V5R4 SQE Query Coverage I5/OS V5R3 SQE Query Coverage I5/OS V5R3 HighlightsCommon table expressions Deletes Derived tables I5/OS V5R3 SQE Performance Enhancements Like/Substring predicates DB2 Multisystem LOB columns Partitioned Table Support Materialized Query Table Support Fast Delete Support V5R2 Highlights Introduction of the SQL Query EngineSQE Optimizer SQE Database Statistics SQE Query Engine Statistic Usage Example SQE for V5R2 Summary IndexingEncoded Vector Indices EVIs DB2 for i5/OS Memory Sharing Considerations DB2 Symmetric Multiprocessing feature IntroductionDecision Support Queries SMP Performance Summary Journaling Journaling and Commitment ControlRemote Journal Function Smapp System-Managed Access Path Protection Smapp Commitment Control DB2 Multisystem for i5/OS Copyright IBM Corp Triggers Referential Integrity Description Variable Length Fields Performance Expectations Description of Function Reuse Deleted Record SpaceComparison to Normal Inserts Performance References for DB2 Copyright IBM Corp Communications Performance CCIN3 System i Ethernet Solutions Hardware Communication Performance Test EnvironmentSoftware Dasd Performance Communication and Storage observationsLarge data transfer FTP Streaming Performance TCP/IP non-secure performance TCP/IP Secure Performance RR & Acrr Performance 1167 565.4 530.0 479.6 462.1 202.2 SSL Performance249.7 53.4 48.0 31.3 27.4 478.4 55.7 53.3 36.9 31.9 VPN Performance SSL Relative Performance Performance Observations and Tips VPN Relative Performance27.5 10.9 15.4 18.7 88.8Page APPC, ICF, CPI-C, and Anynet Page Test Type Hprip Link Hprip Link Speed AnyNet HPR and Enterprise extender considerationsSpeed = 10Mbps = 100Mbps For network information select Networking Additional Information Web Server and WebSphere Performance Http Server powered by Apache Relative Performance Metrics Http Server powered by Apache for i5/OS V5R4 Relative Capacity for CGI Http Server powered by Apache for i5/OS Page Page Page Page PHP Zend Core for System ConfigurationZend Core for Database and Workload Description OS / DB Database Access Database Persistent versus Non-Persistent Connections Zend Platform Database Isolation Levels PHP System Sizing What’s new in V6R1? WebSphere Application ServerThi Sho Using direct map native Jdbc Tuning changes in V6R1Using ‘threadUsed=false custom property toolbox Jdbc Use of unix sockets toolbox JdbcPage Topology of the Trade Application Page WebSphere Application Server Trade Capacity Results Trade3 Measurement Results Trade 6 Primitives Trade Scalability Results Description of Trade primitives in Figure WebSphere Trade 3 Primitives Accelerator for System Trade Jdbc Response Time 1 User Time seconds Page WebSphere Application Server V51 Express WebFaced applications, one simple, one IBM WebFacingModerate, and one complex. See .4, for More in line with the other two machines Version 5.0 of WebfacingMany 5250 applications have been implemented With best performance techniques, such as Display File Record I/O Processing Tuning the Record Definition Cache Cache Size Effect Cache Management Definition Cache Content ViewerDefinitions would be retained in the cache with few cache Misses for not commonly used records Cache Viewer Button operations Cache Management Record Definition Loader Button Operation Compression Record Definition Loader Button operations Operation Webfacing Compression Enabling Compression in IBM Http Server powered by Apache Additional Resources PartnerWorld for Developers Webfacing website What’s new with V5R4 and Hats WebSphere Host Access Transformation Services HatsHats Customization Hats Customization CPW/User IBM Systems Workload Estimator for Hats System Application Server InstancePage WebSphere Commerce WebSphere Portal WebSphere Commerce Payments Performance Tips and Techniques Connect for iSeries Performance Observations/Tips Connect for iSeries Connector TypesPage Introduction Java PerformanceWhat’s new in V6R1 IBM Technology for Java 32-bit and 64-bit Garbage Collection Native Code JIT Compiler Classic VM 64-bitPage Garbage Collection Bytecode Verification Functional Considerations Determining Which JVM to Use Performance Considerations General Guidelines Capacity Planning Java Performance Tips and Techniques Classic VM-specific Tips I5/OS Specific Java Tips and Techniques Java Language Performance Tips Consider the special property os400.jit.mmi.threshold Minimize synchronized methods Java i5/OS Database Access Tips Instead, the above procedure should be written as Pool Database Connections Resources IBM Software Solutions System i Cryptographic SolutionsIBM Hardware Solutions Cryptography Performance CSP API Sets Cryptography Performance Test Environment Measurement Results Software Cryptographic API Performance Bytes/Second Hardware Cryptographic API Performance Encryption Threads Cipher Encrypt Performance4764 Transactions/second Bytes/second CCA CSP Cryptography Observations, Tips and Recommendations Supported number of 4764 Cryptographic Coprocessors ISeries NetServer File Serving Performance ISeries NetServer File Serving Performance Conclusion/Explanations Response TimePage Jdbc Performance Tuning Tips 10.1 DB2 for i5/OS access with JdbcDB2 for i5/OS Jdbc and Odbc Performance References for Jdbc 10.2 DB2 for i5/OS access with Odbc Odbc Performance Settings Page Microsoft Odbc webpage References for Odbc Domino on Mail and Calendaring Users MCU Domino Workload Descriptions Domino Domino Web Access formerly known as iNotes Web Access Domino Domain Monitor Additional memory was added for this test Domino Web Access client improvements with Domino Response Time and Megahertz relationship Collaboration Edition Collaboration Edition and Domino Edition offeringsDomino Edition Performance Tips / Techniques Page Page Domino Subsystem Tuning Domino Web Access Main Storage Options Performance Monitoring Statistics Minimize Thousands V5R3 Main Storage Options Response Times Sizing Domino on System Performance Data Collection Tools Lpar and Partial Processor Considerations System i NotesBench Audits and Benchmarks Performance Improvements for WebSphere MQ V5.3 CSD6 WebSphere MQ for iSeriesQueue Manager Start Following an Abnormal End Checkpoint Following a Journal Receiver Roll-over Conclusions, Recommendations and Tips Test Description and Results Other Sources of Information Linux on iSeries Performance Basic Requirements -- Where Linux RunsSummary Key Ideas Linux on iSeries Architecture Linux on iSeries Technical Overview Linux on iSeries Run-time Support Basic Configuration and Performance Questions Computational Performance -- C-based code General Performance Information and Results Computational Performance -- Java Number Processors PartitionWeb Serving Performance Network Operations TCP/IP Function Megabit Ethernet LAN Virtual LANGcc and High Optimization gcc compiler option -O3 Gcc Compiler, Version Virtual Disk Value of Virtual LAN and Virtual Disk 13.7 DB2 UDB for Linux on iSeries Top Tips for Linux on iSeries Performance Linux on iSeries and IBM eServer Workload EstimatorPage Page Page Internal Native Attachment Dasd Performance Devices & Controllers Direct Attach Native Hardware Characteristics Intensive Workload Performance Comparison 14.1.2 iV5R2 Direct Attach Dasd Intensive Workload 14.1.3.1 571B RAID5 vs RAID6 10 15K 35GB Dasd 14.1.3 571B14.1.3.2 571B IOP vs Iopless 10 15K 35GB Dasd Dasd 14.1.4 571B, 5709, 573D, 5703, 2780 IOA Comparison Chart 14.1.4.4 14.1.5.1 14.1.6.2 Comparing 571E/574F and 571F/575B IOP and IOPLess 1000 2000 3000 4000 Ops/Sec 14.1.7.4 14.1.8.1 Performance Limits on the 571F/575BPage 14.1.9.2 14.1.9.1 Direct Attach 571E/574F and 571F/575B Observations 100 200 300 400 500 600 Ops/Sec New in iV5R4M5 14.2.1 9406-MMA CEC vs 9406-570 CEC Dasd RAID Hot Spare 14.2.3 12X Loop Testing Non Encrypted ASP vs Encrypted ASP New in iV6R1M0 Encrypted ASP Non Encrypted ASP vs Encrypted ASP 100 350 600 850 1100 1350 1600 1850 14.3.2 57B8/57B7 IOA POWER6 520 57B8/57B7 6 RAID5 Dasd in CEC 14.3.3 572A IOA Page SAN Storage Area Network External 14.5 iV6R1M0 -- Vios and IVM Considerations Generic Concepts General Vios Considerations Generic Configuration Concepts Page Page Chdev -dev hdisk03 -attr pv=yes will assign a Pvid to hdisk3 Page Vios and JS12 Express and JS22 Express Considerations 001 10000 20000 30000 40000 50000 60000 Vios CPU Vios Average Dasd Response Time Seconds BladeCenter S and JS12 Express Page Page System Workload Response Time Seconds IBM i operating system 5.4 Virtual Scsi Performance Page Introduction Virtual Scsi Performance Examples Virtual Scsi Bandwidth-Multiple Network Storage Spaces Native vs. Virtual Performance Nwsd Read Scaling Virtual Scsi Bandwidth-Disk Scaling Sizing Sizing when using Dedicated Processors CPU milliseconds to process virtual Scsi I/O transaction Sizing when using Micro-Partitioning Disk 1GB Sequential Read Sizing memory AIX Virtual IO Client Performance Guide Supported Backup Device Rates Save/Restore Performance Use Optimum Block Size Useoptblk Save Command Parameters that Affect PerformanceData Compression Dtacpr Data Compaction Compact Database File related Workloads WorkloadsUser Mix User Mix 3GB contains 12,300 objects Comparing Performance Data Medium & High Performing Backup Devices Lower Performing Backup DevicesUltra High Performing Backup Devices Use of Multiple Backup Devices Parallel and Concurrent Library Measurements Large File Concurrent Save and Restore Rates Large File Parallel Save and Restore Rates User Mix Concurrent User Mix Concurrent Runs Number of Processors Affect Performance 450 440 430 420 100 Dasd and Backup Devices Sharing a Tower Virtual Tape 200 Data To Save Hours Parallel Virtual Tapes 500 Concurrent Virtual Tapes 1500 1000 Large File Virtual Tape Scaling Save Write to Virtual tape Save and Restore Scaling using a Virtual Tape Drive Large File Save User Mix Saves PCI-X ™ In a 5094/5294 tower use slot C08 or C09 High-End Tape Placement on System 700 250 SLR60 15.19 5XX Tape Device Rates Measurements in GB/HR all 8 Dasd in the system ASP Source File 1GB 15.21 5XX DVD RAM and Optical Library High Software Compression ASP 15.23 9406-MMA DVD RAM 1GB 15.24 9406-MMA 576B IOPLess IOA IV5R4M5 July What’s New and Tips on Performance What’s New IV6R1M0 MarchIV5R4 January August IPL Performance Considerations IPL PerformanceIPL Test Description 16.3 9406-MMA System Hardware Information Software ConfigurationActive Database 16.4 9406-MMA IPL Performance Measurements Normal 1 iV5R4M5 Normal IPL Power-On Cold Start16.5 9406-MMA IPL Performance Measurements Abnormal 1 iV5R4M5 Abnormal IPL Partition MSD MSD Affects on IPL Performance Measurements 16.7.2 5XX Large system Hardware Configuration 16.7.1 5XX Small system Hardware Configuration16.7 5XX System Hardware Information Database 16.8 5XX IPL Performance Measurements Normal 1 Normal IPL Power-On Cold Start16.9 5XX IPL Performance Measurements Abnormal Measurement units are in hours, minutes and seconds IPL Tips 16.10 5XX IOP vs IOPLess effects on IPL Performance Normal V5R4 iSCSI Host Bus Adapter iSCSI HBA Integrated BladeCenter and System x PerformanceIntegrated xSeries Adapter IXA Integrated xSeries Servers IXS Effects of Windows and Linux loads on the host system17.2.1 IXS/IXA Disk I/O Operations ISCSI Disk I/O Operations Extended Write Operations Crtsbsd SBSDQGPL/QFPHIS POOLS1 10000 ISCSI virtual I/O private memory pool 17.2.5 IXS/IXA IOP Resource Virtual Ethernet Connections ISCSI attached servers IXS and IXA attached servers Machine Pool MBytes Base Pool Qfphis Private Pool Total Disk I/O CPU Cost CPW per 1k Disk Operations Further notes about IXS/IXA Disk Operations ISCSI Target, IXA Capacity Comparison MB per Second Disk I/O Throughput VE Capacity Comparisons Virtual Ethernet CPU Cost and Capacities PP TCP Stream i5/OS to Windows VE CPW Cost Windows CPU Cost File Level Backup Performance Flbu SAV / RST Rates Additional Sources of Information Microsoft Hardware Compatibility Test URL See V5R3 Information Logical Partitioning LparV5R2 Additions General Tips V5R1 Additions Considerations Performance on a 12-way system Lpar Performance ConsiderationsPage Lpar Throughput Increase Lpar Measurements 5920 26 % Over a Standalone 12-way5700 21 % TPC-C Commercial Performance Miscellaneous Performance InformationSAP Performance Information VolanoMark SPECjbb2000 Dynamic Priority Scheduling Delay Cost Terminology Priority Mapping to Delay Cost Curves Qdynptyscd = ‘1’ on Qdynptyscd = ‘0’ Performance Testing Results Main Storage Sizing Guidelines Memory Tuning Using the Qpfradj System Value Additional Memory Tuning Techniques Setobjacc Set Object Access User Pool Faulting GuidelinesLarge Memory Systems Batch Interactive 19.7 AS/400 NetFinity Capacity Planning AS/400 NetFinity Software Inventory Performance Conclusions/Recommendations for NetFinity General Performance Tips and Techniques Adjusting Your Performance Tuning for ThreadsHistory Coming Change Solution Problem In-lining General Performance Guidelines -- Effects of CompilationOptimization Levels Theory -- and Practice Typical Storage Costs System Level Considerations Brief Example Order1 Orderj Ordert OrderN Short but Important Tip about Data Base Which is more important? Final Thought About Memory and Competitiveness HMT Described Hardware Multi-threading HMT HMT and SMT Compared and Contrasted HMT Feature SMT FeatureSome key similarities and differences are Models With/Without HMT POWER6 520 Memory Considerations Aligning Floating Point Data on Power6 Page Switchable IASP’s High Availability Switchable Resources ConsiderationsHigh Availability Performance Workload Description System Configuration Workload ConfigurationHardware Configuration Cabling Map Geographic Mirroring Switchover Tips Workload Description Hardware Configuration Cabling Map Large System Configuration Geographic Mirroring Measurements Switchable Towers using Geographic MirroringSynchronization on an idle system Active State Geographic Mirroring Tips IBM Systems Workload Estimator Merging PM for System i data into the Estimator Estimator AccessWhat the Estimator is Not Page Commercial Processing Workload CPW Appendix A. CPW and CIW Descriptions CPW Application Description Compute Intensive Workload CIW Page Page How Collection Services works Performance Data Collection ServicesStarting Collection Services Batch Modeling Tool Bchmdl Page Page V6R1 Additions October CPW values for the IBM Power Systems IBM i operating system V6R1 Additions August CPW values for the IBM Power 595 IBM i operating systemV6R1 Additions April V6R1 Additions January Table C.3.1. CPW values for Power System Models IBM System i using the POWER5 processor technology IBM System i using the POWER6 processor technologyV5R4 Additions July V5R4 Additions January/May/August 2006 and January/April Model Edition Accelerator IBM ~ i5 Servers V5R3 Additions May, July, August, October 2004, JulyValue Express CPU ISeries Model 8xx Servers V5R2 Additions February, May, July750 540 V5R2 Additions Model 810 and 825 iSeries for Domino FebruaryModels share L2 cache between 2 processors Standard Models 8xx Servers Table C.9.2.1 Standard Models 8xx Servers 10 V5R1 Additions Table C.9.1.1 Model 8xx Servers Model 8xx Servers 10.3 V5R1 Dedicated Server for Domino Model 2xx ServersTable C.10.2.1 Model 2xx Servers Table C.10.3 .1 Dedicated Servers for Domino CPW Values and Interactive Features for CUoD Models Capacity Upgrade on-demand ModelsActive Stand-by 840-2352 840-2353 840-2354 16500 Table C.10.4.1.1 V5R1 Capacity Upgrade on-demand Models 11.1 AS/400e Model 8xx Servers 11 V4R5 AdditionsTable C.11.1 Model 8xx Servers All new Condor models Table C.11.2.1 Model 2xx Servers Dedicated Server for DominoTable C.11.3.1 Dedicated Server for Domino 12 V4R4 Additions SB Models12.1 AS/400e Model 7xx Servers Table C.11.4.1 SB Models Table C.12.2.1 Current Model 170 Servers Model 170 Servers Current 170 Servers Previous Model 170 Servers AS/400e Dedicated Server for Domino2407 2408 2409 120 Table C.12.2.2 Dedicated Server for Domino 14 AS/400e Custom Servers 13 AS/400e Model Sxx Servers15 AS/400 Advanced Servers Table C.13.1 AS/400e Servers Table C.15.1 AS/400 Advanced Servers V4R1 and V4R2 16 AS/400e Custom Application Server Model SB1Table C.15.2 AS/400 Advanced Servers V3R7 Table C.16.1 AS/400e Custom Application Server Model SB1 17 AS/400 Models 4xx, 5xx and 6xx SystemsTable C.17.1 AS/400 Risc Systems Table C.17.2 AS/400e Systems 18 AS/400 Cisc Model Capacities Table C.18.7 AS/400 Advanced Systems Cisc Table C.18.6 AS/400 Cisc Model 9406 SystemsTable C.18.8 AS/400 Advanced Servers Cisc