RE: Some new results

["Joe Pluta" <joepluta@xxxxxxxxxxxxxxxxx>]

> From: William Washington III
> 
> SO, the optimal (highest performance) technique might be to use native
I/O
> to a
> file with access path maintenance to *DLY, then change the file back
after
> the
> process has finished.  Just a thought in need of an experiment.

I suspect you're right, William, and I also suspect that there will be a
zillion different little variations like this.  That's why I wanted to
build an automated framework to do the testing.

Right now I run all ten tests with a single command:

RUNSUITE RUNID('SIMPLEIO-1')              
         CFG(MYCONFIG)                    
         SUITE(SIMPLEIO)                  
         DESC('Running Simple I/O Suite')

SIMPLEIO is the SUITE ID, and identifies the suite.  A suite expands to
a list of tests.  Each test specifies the program to call as well as
various runtime parameters (for example, two of the SQL tests call the
same program, but with a different number of rows).  The results are
logged to a file that can then be queried.

The command as shown also allows you to add a "run ID", which is just a
10-character key for reporting purposes, as well as a description.  The
last parameter, CFG, is used to identify the configuration the test is
running under, which specified things like OS version and the amount of
disk/RAM.  I haven't worked out all the kinks yet, but I think we've got
enough to get us started.

Joe