[Search for users] [Overall Top Noters] [List of all Conferences] [Download this site]

Conference ilbbak::ibi_focus

Title:FOCUS, from INFORMATION BUILDERS
Moderator:ZAYIUS::BROUILLETTE
Created:Thu Feb 19 1987
Last Modified:Mon May 05 1997
Last Successful Update:Fri Jun 06 1997
Number of topics:615
Total number of notes:1779

178.0. "FOCUS vs. RDB Benchmarks" by NCPROG::KERNS (Fast - maybe, Good - maybe, Cheap - never!) Mon Feb 06 1989 12:07

Has anyone done any benchmarking on FOCUS and RDB V3.0? For example, does FOCUS slow down the data access rate when using RDB?
How about RALY and TEAMDATA vs. FOCUS?

If no one has done any benchmarks, does anyone have any suggestions on how to do them?

Steve
T.RTitleUserPersonal
Name
DateLines
178.1focus benchmarkMEMV01::COPPERSMITHSYSTEM-W-RELFLT, reality faultMon Feb 06 1989 16:1021
    We benchmarked a FOCUS application using DTM, a Fortran benchmarking
    subroutine from the VTX Toolshed (modified for use in Fuselib) and
    SPM.  DTM was required for automated scripting because the FOCUS
    application was driven by FMS screens.
    
    This approach pushed the uncertainty principle a little bit, because
    we were using software on a system to measure the effect of other
    software on the system, but since the goal was to derive minimum
    system configurations for the application, this extra load actually
    served to provide the estimate with a margin of safety.
    
    The benchmark also progressed at machine speed rather than human
    speed (of input), but the benefit of the automation was that the
    benchmark was reproducible across different systems.
    
    The experimental design and results are available from MEM Finance
    IS.  The system measured was the SMIRS system.
    
    
    Curtis Coppersmith
    232-2246 DTN
178.2improvements to RDB interfaceFDCV10::REVRPTMon Apr 03 1989 15:5216
    
    Our group is using FOCUS to go against an Rdb database and encountered
    some performance problems and decided to explore. We used a simple
    fex to navigate the entire relation sequentially and sum a $ field.
    The target relation has about 640,000 entries. We also built a FOCUS 
    database containing identical data. A simple RDO request took 3
    mins. of CPU to sum the field. FOCUS with a FOCUS db took 4 mins.
    of CPU time to produce the total. FOUS with Rdb took 19 mins CPU. 
    The FOCUS hotline told me that the technical group is currently
    working on the FOCUS/Rdb interface and should have a new beta version
    (5.22 ?) available in 2 weeks. The current interface instructs Rdb
    to returns each field to FOCUS where the sum takes place. The next
    release will allow Rdb to produce the sum and pass the one field
    back to FOCUS.