Migration and exploitation considerations for z13 server functions

Start of changeThe following z13™ functions have considerations when you are planning for migration and exploitation. For PTF information, see the Notes for Table 1.End of change

  1. Large Systems Performance Reference (LSPR) method. LSPR is designed to provide comprehensive z/Architecture® processor capacity ratios for different configurations of central processors (CPs) across a wide variety of system control programs and workload environments.
  2. Simultaneous multithreading (SMT). Incremental throughput is achieved partly because the new processor chip offers intelligently implemented 2-way simultaneous multithreading. Simultaneous multithreading allows two active instruction streams per core, each dynamically sharing the core's execution resources. SMT is available on the z13 for workloads running on the Integrated Facility for Linux (IFL) and the IBM® z Integrated Information Processor (zIIP).
  3. CFCC Level 20 support. CFCC Level 20 supports the Coupling Facility use of Large Memory to improve availability for larger CF cache structures and data sharing performance with larger DB2® Group Buffer Pools. This support removes inhibitors to using large CF cache structures, enabling use of Large Memory to appropriately scale to larger DB2 Local Buffer Pools and Group Buffer Pools in data sharing environments.
  4. Coupling links. IBM z13™ introduces new PCIe (Integrated Coupling Adapter – ICA SR) based Short Reach coupling links using a new CHPID type, CS5. ICA-SR links can only connect a z13 CPC to a z13 CPC. IBM z13 CPC now supports 256 Links. A single z/OS® or a CF Image still supports a maximum of 128 Links. When displaying STP (DISPLAY ETR) from a z/OS image, information is provided for the entire CPC. If greater than 128 links are defined on an IBM z13, the z/OS support must be installed on all z/OS releases running on that IBM z13, so that greater-than-128-links information, including STP Timing information, can be displayed.
  5. SIMD. z/OS V2R1 is designed to support the new vector extension facility (Single Instruction Multiple Data, SIMD) instructions available on IBM z13 servers. SIMD provides a powerful framework for development of new Business Analytics workloads, porting math-intensive workloads from other platforms, and accelerating Business Analytics workloads on IBM z13. High Data Intensity, High Compute Intensity, Predictive IT Analytics, Advanced Security/Crypto, BI Reporting, Perspective Analytics, and Next-Gen Data Warehousing are some of the workloads that might benefit from Data Parallelism (SIMD). z/OS support includes enablement of Vector Registers (VR) on IBM z13, Mathematical Acceleration Subsystem (MASS), and Automatically Tuned Linear Algebra Software (ATLAS) support, as well as LE enablement for C runtime functions.
  6. Cryptographic Enhancements. Cryptographic enhancements for z/OS V2R1 on z13 servers include:
    • VISA Format Preserving Encryption (VFPE). Support for VFPE algorithms in CCA-based callable services helps legacy databases to contain encrypted data of sensitive fields without having to restructure the database or applications. This support will rely on the Crypto Express5S coprocessor.
    • Start of changeGreater than 16 Domain support. This support allows a cryptographic coprocessor to be shared across more than 16 domains, up to the maximum number of LPARs on the system. This support relies on enhanced firmware available with a minimum microcode level for the Crypto Express5S coprocessor.End of change
    • Trusted Key Entry (TKE) 8.0 Licensed Internal Code (LIC). This support includes: Crypto Express5S Coprocessor support, FIPS Certified Smart Card, Crypto Coprocessors with more than 16 domains, a full-function migration wizard for EP11 coprocessors, new master key management functions, a Smart Card Readers Available indicator, a Configure Displayed Hash Size utility, ECC Authority Signature Keys, print capability, new features in the Crypto Node Management (CNM) utility, ENC-Zero verification pattern for 24-byte DES operational keys, and usability enhancements.
    See Decide on the steps you will take for your migration to a z13 server in Actions you can take before you order a z13 server
  7. IBM zAware system. IBM zAware is designed to offer a near real-time, continuous learning, and diagnostics capability intended to help you pinpoint and resolve potential problems quickly enough to minimize impacts to your business. The new version of IBM zAware introduces a new generation of technology with improved analytics to provide better results. After you order a z13 you can prepare to set up an IBM zAware environment by defining the IBM zAware logical partition, defining and using OPERLOG log streams, and network definitions to connect the z/OS LPAR to the zAware LPAR. For more information, see IBM System z Advanced Workload Analysis Reporter (IBM zAware) Guide, (SC27-2623) and z/OS MVS Setting Up a Sysplex.
  8. Flash Express®. With this support, z/OS is designed to help improve system availability and responsiveness by using Flash Express across transitional workload events such as market openings, and diagnostic data collection. z/OS is also designed to help improve processor performance by supporting middleware exploitation of pageable large (1 MB) pages. This support requires the Flash Express hardware feature.
  9. 2GB Large Page support. This support includes support for 2 GB large fixed pages.
  10. New z/Architecture instructions: XL C/C++ ARCH(11) and TUNE(11) parameters. This function provides new hardware instruction support, SIMD (Vector support, including data), Decimal Floating Point packed conversion facility support, performance improvements (machine model scheduling and code generation updates), and Single Instruction Multiple Data (SIMD) instruction set and execution, including Business Analytics Vector ProcessingMASS and ATLAS Library. Unlike prior generations of servers, to use any of these functions you must download and install the z13 Exploitation Support for z/OS V2.1 XL C/C++ web deliverable.
  11. FICON® Express16S. FICON Express16S supports a link data rate of 16 gigabits per second (GBPS) and auto-negotiation to 4 or 8 GBPS for synergy with existing switches, directors, and storage devices. With support for native FICON, High Performance FICON for System z® (zHPF), and Fibre Channel Protocol (FCP), the new FICON Express16S channel is designed to work with your existing fiber optic cabling environment. The FICON Express16S feature running at end-to-end 16 GBPS link speeds will provide reduced latency for large read/write operations, and increased bandwidth compared to the FICON Express8S feature.
  12. FICON Dynamic Routing. With z13, FICON channels are no longer restricted to the use of static Storage Area Network (SAN) routing policies for Inter-Switch Links (ISLs) for cascaded FICON directors. You will need to ensure that all devices in your FICON SAN support FICON Dynamic Routing before implementing this feature.
  13. Storage Area Network Fabric I/O Priority. This new function on the z13 provides the ability for z/OS to specify an I/O priority for the SAN fabric to utilize. This capability allows z/OS to extend the z/OS Work Load Manager (WLM) to manage the SAN fabric, completing the management of the entire end-to-end flow of an I/O operation.
  14. Improved High Performance FICON for System z (zHPF) I/O Execution at Distance. High Performance FICON for System z (zHPF) has been enhanced to allow all large write operations (> 64 KB) at distances up to 100 km to be executed in a single round trip to the control unit, thereby not elongating the I/O service time for these write operations at extended distances.
  15. Improved Channel Subsystem (CSS) Scalability. The z13 server has improved the channel subsystem (CSS) scalability with support for six logical channel subsystems (LCSSs) which are required to support the eighty-five LPARs for z13, four subchannel sets (to support more devices per logical channel subsystem), and 32K devices per FICON channel, up from 24K channels in the previous generation. Additionally, a fourth subchannel set for each LCSS is provided to facilitate elimination of single points of failure for storage after a disk failure.
  16. Start of changePCIe and HCD Definitions. z/OS V2R1 HCD (FMID HCS7790) supports full exploitation for z13 processors (type 2964) with support of 6 channel subsystems, 4 subchannel sets. To support PCIe functions for systems that are running on zEC12, zBC12 and z13 processors, HCD has introduced a new dialog for defining PCIe functions and assigning the functions to LPARs.

    z/OS V1R13 HCD (HCS7780) supports limited exploitation for z13 processors (type 2964). For PCIE functions, it is possible to define a complete 2964 processor, including PCIe functions and the new channel types and capabilities. However, a processor configuration can be dynamically changed or activated only if it contains neither PCIe function nor a PNETID attribute. For a software activate with hardware validation, PCIe functions and the PNETID attribute are ignored.

    IBM recommends that you define and activate all the new hardware definitions on a z/OS V2R1 system with the appropriate HCD PTF (APAR OA44294) installed and perform only software activates (with hardware validation) on lower-level systems.

    End of change
  17. Start of changeConsider the changes in the CPU Measurement Facility counters. The number of CPU measurement facility counters for z13 remains at 128. Though the structure of the SMF 113 record does not change, the values, interpretations, and frequency of certain sections do change; therefore, current tools that use the data must be updated for the z13.
    For example, consider the following SMF record field:
    • SMF113_2_CtrVN2 identifies how to interpret the MT-Diagnostic, Crypto, and Extended counter sets. As described in The IBM CPU Measurement Facility Extended Counters Definition for z10 and z196, SA23-2260, this field is set to 1 (for z10™), 2 (for z196 or z114), 3 (for zEC12 and zBC12), or 4 (for z13).
    Note: Start of changeAs of z/OS V2R1, if you use the CPU Measurement Facility, IBM recommends that your installation collect SMF 113 subtype 1 and 2 records. IBM also recommends that products process SMF 113 subtype 1 records when available because that is where future enhancements will be made. If subtype 1 records are not available, products can process subtype 2 records.End of change
    End of change