Strange Results testing .Net vs. COM vs. Classic ASP ...

[Guest]

Hello everyone!

I have been running some tests of a component that simply returns an image
path based upon some logic and a simple select statement to an Access db. The
component had to be available to Classic ASP apps and ASP.Net apps from a
central location. Performance is a great concern.

In developing this solution I created three seperate mini-apps to try to
ascertain what the best approach could be.

The Three mini-apps:
1) A classic ASP page using an include file encapsulating component logic.
2) A ASP.Net page referencing a .NETclass that encapsulates component logic.
3) A classic ASP Page creating a COM object from a A COM interface created
around a copy of the class referenced in app #2.

When testing processing times I was surprised to see that the Classic ASP
apps 1 & 3 had roughly the same results while the ASP.Net app (#2)
refrencing the assembly in the GAC took almost twice as long to doexactly the
same job.

Here is logic for component:
Create Class with no parameters (or call function for include file)
Set two parameters (write only properties)
Call MakeImage Method (This method creates a connection to an access DB
using a referenced assembly) (also note the include file contains all db
access code)
Destroy object

Why does the Assembly built with COM interface out perform the norrmal .NET
assembly when they are virtually the same?

Any ideas?

Thanks,
RichDef

 

[Marina]

Did this happen after you just recompiled the app? Because the first hit
takes a while. After that, it is the normal speed.

 

[Guest]

I ran each app a number of times after recompiling and the results were the
same - the native .net app took about twice as long to process.

Does the vb script timer() function work the same in vb.net? I'm using the
timer method in each app to determine how long it takes to process the code.