G
Greg Sohl
I have some code which I'm testing both within a test harness WinForm app
and separately in its intended production environment which is an ASP.NET
web service. In both cases, the code set up and function is the same. The
code is always being run in a single managed Thread from the ThreadPool.
However the performance running within the web service is 7+ times slower
than running outside of web services within the test harness.
Note that the code includes database access (SQL Client), GDI manipulation
for printing and significant memory manipulation, also for the printing.
In profiling, the database access times are similar. The ThreadPool thread
is set to normal priority, as it should be.
My test environment is Win XP SP1. I've tried both Low and Medium isolation
mode (application protection) settings. Similar performance degredation is
being seen in our production server (Win Server 2000/2003) environments,
compared to the execution speed within the test harness app.
I'd understand and could accept if the performance were slightly different.
But I'm not sure where to start looking for this significant a difference.
I'd appreciate any ideas or experience you might have with similar issues
and will gladly address any questions to further investigate the
possibilities here.
Thank you.
Greg
and separately in its intended production environment which is an ASP.NET
web service. In both cases, the code set up and function is the same. The
code is always being run in a single managed Thread from the ThreadPool.
However the performance running within the web service is 7+ times slower
than running outside of web services within the test harness.
Note that the code includes database access (SQL Client), GDI manipulation
for printing and significant memory manipulation, also for the printing.
In profiling, the database access times are similar. The ThreadPool thread
is set to normal priority, as it should be.
My test environment is Win XP SP1. I've tried both Low and Medium isolation
mode (application protection) settings. Similar performance degredation is
being seen in our production server (Win Server 2000/2003) environments,
compared to the execution speed within the test harness app.
I'd understand and could accept if the performance were slightly different.
But I'm not sure where to start looking for this significant a difference.
I'd appreciate any ideas or experience you might have with similar issues
and will gladly address any questions to further investigate the
possibilities here.
Thank you.
Greg