FOCUS AREAS

• Base SAS
• Graphics
• Enterprise Management Integration
• Migration
• Scalability & Performance
  • Papers
  • Grid Computing
  • Scheduling
  • Platform Suite for SAS
  • SAS/ACCESS Software
  • SAS/CONNECT Software
  • SAS Metadata Server
  • SAS OLAP Server
  • Scalable SAS Procedures
  • SPD Engine
  • SPD Server
  • ARM Tools
  • FULLSTIMER Option
• Beautiful Reports
• Statistics & Operations Research

Hot Topics

• Bank of America avoids gridlock in credit-risk scoring, forecasts using SAS
  Bank of America uses a SAS grid environment to reduce loan default calculation times from 96 hours down to just 4 saving valuable time and money.

Scalable SAS Procedures

From both in-house as well as customer experience, several SAS procedures have been identified as performance critical, especially for long-running applications. In SAS 9 some of these procedures have been modified to take advantage of faster I/O and SMP hardware. These modifications include support for multi-threading. With the appropriate hardware configuration and these enhanced procedures, bottlenecks in your long-running applications can be minimized.

For the selected SAS procedures, internal modifications have been made to take advantage of threads in order to parallelize some or all of the computationally intensive portions of the procedures and reduce the real time of execution. In addition, the appropriate hardware configuration can increase the effective transfer rate of data into the procedures to also help minimize overall execution time.

The procedures that have been modified so far in SAS 9 to take advantage of SMP hardware include:

• in BASE SAS
  • PROC SORT *
  • PROC SQL (group by, order by)
  • PROC SUMMARY/MEANS
  • PROC REPORT
  • PROC TABULATE
• in SAS/STAT
  • PROC ADAPTIVEREG
  • PROC BCHOICE
  • PROC FMM
  • PROC GLM
  • PROC GLMSELECT
  • PROC HPGENSELECT
  • PROC HPLOGISTIC
  • PROC HPLMIXED
  • PROC HPCANDISC
  • PROC HPFMM
  • PROC HPREG
  • PROC HPSPLIT
  • PROC HPNLMOD
  • PROC HPPRINCOMP
  • PROC IRT
  • PROC LOESS
  • PROC MCMC
  • PROC NLIN
  • PROC QUANSELECT
  • PROC QUANTLIFE
  • PROC REG *
  • PROC ROBUSTREG
• in Enterprise Miner
  • PROC DMREG *
  • PROC DMINE *

(* Supports parallel read with SPD Engine)

The SPD Engine and several of the SAS/ACCESS engines support parallel read capabilities in SAS 9. With the appropriate hardware configuration, these engines can perform parallel partition reads to deliver more MB/sec to SAS procedures.

It is important to note that performance gains on SMP machines with these procedures are not limited to SPD Engine data sources. Assuming your hardware provides sufficient I/O bandwidth, the threading changes incorporated into these procedures will scale against standard base data sets. Our testing of these procedures using existing base data sets has shown excellent scalability, especially on machines that provide hardware striping.

We at SAS have created the Scalability Community to make you aware of the connectivity and scalability features and enhancements that you can leverage for your SAS installation. The success of this community depends on you. Send electronic mail to scalability@sas.com with your comments, requirements, and suggestions.