Delta V Software Forums

I am in the process of evaluating messaging frameworks for use in our products.

In the simple scenario (Hello world example) I tried repeated calls from the client side, but even with
Win32NamedPipeEndpoint and calls without parameters I could not get more than 2400 messages through.
There seems to be a fixed latency with more than 400 µsec inside.

Even if there are other usage scenarios (asynchronous, publish/subscribe), I hoped to get more calls on the
LAN than I get here on a single machine.

Are there any configuration options to optimize this timings ?

The main thing is to make sure you have built the source code in release configuration, not debug... With client and server on the same machine you should quite easily be able to get over 10000 requests/second.

Can you describe your build and what platform you're running on?

Thanks, you are right.
I did the first tests in debug mode. After switching to Release-Build I achieved over 10000 messages without
optimization and >20000 Messages with Full optimization.

I am testing on Windows 7, x86 Build with MSVC 2005, Intel Xeon 3.3 GHz.
Tests consists of repeated sending of empty message and std::vector<float> in different sizes.