Risk Factors Intel®’s Roadmap Update presentations may contain forward-looking statements and projections. All statements and projections made that are not historical facts are subject to a number of risks and uncertainties, and actual results may differ substantially. Please refer to our recent Earnings Release and Form 10-Q for more information on the Risk Factors that could cause actual results to differ.3
What we will cover • Intel and Oracle Collaboration • What hardware is inside the Exadata • The difference of Exadata x2-2 and Exadata x2-8 hardware architectures • Intel’s server processor roadmap • Intel’s platform architecture • Overview of Intel’s cpu architecture and technology • Which Intel processors for which workloads4
Oracle-Intel Collaboration Model SYSTEM ARCHITECTURE DESIGNS and INFRASTRUCTURE COMPILERS and TOOLS OPERATING ENVIRONMENTS ENABLEMENT MIDDLEWARE and APPs DATABASE OPTIMIZATION OPTIMIZATION6
Database Workloads: Intel Technologies Increased Bandwidth: Intel® QuickPath InterConnect Increased Security: Dramatic Performance Technology Intel® Advanced Enhancement Enables internal scalability and 3X enhancement with Intel higher performance on database Encryption Standards Xeon processor 7500 series clusters like Oracle RAC New Instructions (AES- launch (Nehalem-EX) and IBM DB2 PureScale NI) Intel® Turbo Boost Improves database software Technology encryption and decryption by More Scalability: Multi- Enables dynamic increase of up to 5X Core & Hyperthreading frequency and performance of Intel Xeon processor E7 family individual cores offers up to 10 physical cores, Increased Performance running up to 20 threads per Tick-Tock model processorRoadmap balances advancements in transistor density with new microarchitecture innovations Higher Availability Memory Scalability with Consolidation 64-bit Itanium Intel Virtualization Technology Reliability: Enhanced Intel address multiple brings High Availability, failover, Ecosystem Support for terabytes of memory and consolidation of databases MCA-R Scalability with NUMA onto a single server Intel Xeon processor E7 family Enables high performance, low with advanced MCA-R helps Intel® VT-x Introduced Higher Performance I/O enable SAP HANA database latency, and highly parallelized Intel virtualization technology VT-x memory access with Solid State Disks recovery from memory errors introduced 35% higher and better performance with Intel SSD for OLTP environments7
Intel® Xeon® Processor 5600 Series 32nm Westmere High-k + Metal Gate Transistors Memory Controller M Core Core Core Core Core Core Q i u s e M c u i e s I c & O I U O & n c & Q P Shared L3 Cache** o r Shared L3 Cache** Q I e P I Transistor count: 1.17B Die size: 248mm222
Intel® Xeon® processor E7-8800/4800/2800 product families(Westmere-EX) Overview• CPU Refresh Next generation core design (“Westmere” uarch) C C C C C C C C C C O O O O O O O 32 nm process technology (new “Tick”) R R R O R R R R O R O R R Socket compatibility with Xeon® 7500 (“drop-in”) E E E E E E E E E E 1 2 3 4 5 6 7 8 9 1 10 Cores; 20 threads with Intel® Hyper-Threading 0 30MB of last level cache (shareable across all cores) 30 MB Shared Last Level Cache Same TDP power levels as Xeon® 7500 processor Integrated Memory Interconnect• Additional Memory Support Controller controller RDIMM up to 1066MHz LV-DIMM support (DDR3L (1.35V) ) 4 Intel® LR-DIMM support (Upto 32GB) Scalable Memory 4 Full-width Interconnects up to Intel® 6.4GT/s QuickPath• Key New Features Interconnects New processor & memory power management capabilities Security: Intel® Trusted Execution Technology (Intel® TXT) Mill Brook 2 Mill Brook 2 DDR3 AES-NI (AES New Instructions – crypto acceleration) RDIMM Mill Brook 2 Channel DDR3 RAS: New memory reliability features RDIMM Mill Brook 2 s DDR3 ChannelRDIMM s DDR3 RDIMM Channel RDIMM VT: Next-gen performance improvements and latency s RDIMMRDIMM Channel RDIMM reductions RDIMM s RDIMMRDIMM RDIMM Improved Performance, RDIMM RDIMM RDIMM Virtualization, Reliability, RDIMM Security, and Power Efficiency23
Intel Performance Enhancements Intel® Hyper-threading Intel® Turbo Boost Technology Technology Increases performance for threaded applications delivering Increases performance by increasing processor frequency greater throughput and responsiveness and enabling faster speeds when conditions allow <8C Turbo 8C Turbo Normal Frequency Core1 Core8 Core0 Core1 Core8 Core0 Core1 Core0 All cores … All cores … Fewer cores … operate at operate at higher may operate at rated frequency frequency even higher frequencies Higher Performance Higher Performance For Threaded Workloads on Demand Available in both E7 and E5 processors24
E7 Protecting Data Integrity Memory Demand & Patrol ScrubbingParity Checking And ECC Ensures that all memoryDetects and corrects soft errors locations contain valid datain cache & data paths tomemory Intel QPI Protocol Protection Via CRC (8bit Or 16bit Rolling)Corrupt Data Containment Mode Coverage of data paths forKeeps data errors from transient errorspropagating across processors,memory, I/O Memory Thermal Throttling Reduces errors by controlling memory temperature Reduces Circuit-level Errors, Detects & Corrects Data Errors Across The System, Limits The Impact Of Errors25
Intel® Xeon® Processor E7 Family Complete Portfolio of RAS Support • Parity Checking And ECC Reduce circuit-level errors • Memory Thermal Throttling • Memory Demand & Patrol Scrubbing Detect and correct data errors • Corrupt Data Containment Mode • Viral Mode • Intel QPI Protocol Protection Via CRC (8bit Or 16bit Contain uncorrected errors Rolling) • MCA Recovery With SW layer Support Heal failing data connections • Intel® SMI Lane Failover • Intel® SMI Clock Fail Over • Intel® SMI & QPI Packet Retry Support redundancy and failover for key system • QPI Clock Fail Over components • QPI Self-healing • Enhanced DRAM Single Device Data Correction • Enhanced DRAM Double Device Data Correction Recover from uncorrected data errors • Fine-Grained Memory Mirroring • Memory DIMM And Rank Sparing • Dynamic CPU And Memory Migration Help predict failures before they happen • Electronically Isolated (Static) Partitioning • MCA Error Logging (CMCI) With OS Predictive Failure Analysis Maintain partitions instead of systems • Memory Board Hot Add • OS Memory On-lining* • CPU Board Hot Add At QPI Proactively replace failing components • OS CPU On-lining Advanced RAS For Highly Available Deployments26
E7 Machine Check Architecture Recovery How It Works MCA Recovery System works in conjunction with OS, VMM, System Normal Status or DBMS to recover or restart processes and Recovery With Error continue normal operation with SW Prevention Error information passed to SW layer Bad memory location flagged so data will not be Error Error Error used by OS or applications Contained Detected1 Corrected Un-correctable Errors Errors HW Un-correctable HW Correctable Errors 1 Errors detected using Patrol Scrub or Explicit Write-back from cache Allows Recovery From Otherwise Fatal System Errors27
E7 Advanced Reliability Requires An Ecosystem Mission Critical OS/VMM/Application Software OEM Innovation Systems Silicon Integrated Solutions To Deliver Maximum Results28 * Other names and brands may be claimed as the property of others.
Which Intel® Xeon® Processor Family is Best for Which Workload? Workload / Usage Business Processing (DB, ERP, CRM, Batch)1 Workloads Xeon E7 Larger Decision Support (Data Warehouse, Business Intelligence) 4S & 8S Large-Scale Consolidation (Including Virtualization & Multi-tier) Servers Application Development High Performance Computing Collaboration 2S Mid/Smaller Workloads Volume Web Infrastructure Servers IT Infrastructure Development/Quality Assurance Fewer/Smaller More/Larger Instances/Users Instances/Users 1 For directional guidance only. This is not a server selection guide. Actual server sizing is a relatively complex effort involving workload characterization including such considerations as type of application, size of workload, number of users, type of transaction, SLA response times, targeted utilization, level and estimation accuracy of workload baseline/peak/growth, physical or managerial constraints, need to maintain a single state at all times to ensure all users see the same results at any given time,30 cost to migrate to a scale-out alternative, and system availability requirements.