Not sure EXACTLY what you want to know about TestComplete and loadtesting. TestComplete's LoadTesting feature operates much on the same principal as any other web application/web service loadtesting tool. There's a transponder that records HTTP requests and responses into a "task" in TestCOmplete. You then create a population of virtual users to which you assign the task. When you execute the test, then, those recorded requests are transmitted and the responses received are checked against the recorded responses. Statistics are measured on KB/sec, request time, response time, processing time, etc, and reported in a test log that includes graphs over time for the requests.
As for Windows based applications, if the windows application operates off of HTTP requests and responses, you can use this tool for that. However, if not, while you can still probably use other features of test complete to "load test" your windows application, they would not be the actual LoadTesting feature of TestComplete.
As to how TestComplete, as a tool, differs from LoadRunner, that I cannot answer. In fact, there's a general policy by the AQA staff that they do not post direct comparisons between their tool and competitor's tools. What I will say is that LoadRunner is a loadtesting ONLY tool while TestComplete includes loadtesting, functional testing, unit testing, manual testing, regression testing, etc. It is a "Complete" test tool. That's the primary difference between TC and LoadRunner.
My suggestion is that, if you're interested in TestComplete as a tool, they have a fully functional 30 trial license that you can download and explore. I'd highly recommend that.
Automated Testing Evangelists
Definition expert - noun - Unknown drip under pressure