ene
Subject: [l/m 2/13/96] NIST source and .orgs.(11/28)..c.be FAQ
Date: 11 Apr 1996 12:25:07 GMT

11.NIST source and .orgs....< This panel >
12.Measurement environments
13.SLALOM
14
15.12 Ways to Fool the Masses with Benchmarks
16.SPEC
17.Benchmark invalidation methods
18
19.WPI Benchmark
20.Equivalence
21.TPC
22.RFC 1242 terminology (network benchmarking)
23
24
25.Ridiculously short benchmarks
26.Other miscellaneous benchmarks
27
28.References
1.Introduction to FAQ chain and netiquette
2
3.PERFECT
4
5.Performance Metrics
6.Temporary scaffold of New FAQ material
7.Music to benchmark by
8.Benchmark types
9.Linpack
10.Network Performance

NISTLIB/NBSLIB -- the mail server daemon
It exists from an ONR project of indeterminate origin.

nistlib@cmr.ncsl.nist.gov
.send index
includes LLNL loops (aka Livermore Loops, Livermore Fortran Kernels (LFK),
.NAS kernels, Los Alamos benchmarks, parcbench, whetstone, dhrystone,
.mendez, and many others
no results -- they do not wish to get sued.


Reviews of various benchmarks:

LFK, Livermore Fortran Kernels, Livermore Loops, MFLOPS.
Frank McMahon's program is one of the more rigorous.
A personal favorite in some ways.
Problems include the sheer volume of information.


There are 14 and 24 loops versions.  The 24 loops version is the
"official one."  Do not trust vendors bearing 14 loop results.
Frank has decided that the output of his programs goes back into the
program in a Comment form.  This helps with the database communication problem.


David Bailey's NAS Kernels.
Uses higher order kernels than either Dongarra and McMahon.
Uses trig functions which can behave in strange ways.
Requires the use of 64-bit arithmetic.
Has interesting consequences for DOUBLE PRECISION COMPLEX types.
These programs are less portable than the less complex loops.
One advantage is an attempt to measure compiler differences.
There is an optional Level which people may code to improve optimization.


ORGs.

Performance mailing list: (oriented toward performance analysis
using quantitative modeling (Mean-Value Analysis [MVA], etc.)
to receive:.perform-request@vuse.vanderbilt.edu
where it comes from:.perform@vuse.vanderbilt.edu

ACM/SIGMETRICS (Perforance Evaluation Review)
.Check the ACM.


Computer Measurement Group

CMG
414 Plaza Drive
Suite 209
Westmont, IL 60559 

Bus: (708) 655-1812
Bus: (800) 436-7264
Fax: (708) 655-1813

They can be contacted at:  72322.3515@compuserve.com

And they put up a web page at:
http://admin2.bc.edu:80/cmg/
Mirrored at:
http://www.cmg.bc.edu/

By the way, CMG is losing it's IBM manframe bend.  Looking through last year's 
presentation, only 55 out of 132 presentations were IBM M/F based, and some
of them would have been applicable for other systems as well. 

Next National Conference: San Diego - December 8 - 13 1996

The US CMG is only one international association, with groups
in many other countries.

Keywords: Measurement,Capacity Management,System Management

                   ^ A  
                s / \ r                
               m /   \ c              
              h /     \ h            
             t /       \ i          
            i /         \ t        
           r /           \ e      
          o /             \ c    
         g /               \ t  
        l /                 \ u
       A /                   \ r
        <_____________________> e   
                Language
 
