Now u can see the difference.
I had to prove why HTML5+XHTML structure is better than non-semantic HTML structure.
Here is the difference.

Render latency? Is that much MS delay visible?

Both pages have same content, even semantic.xhtml has four more paragraph. Both files are very light, just look at Size column.
I had to prove that HTML5+XHTML is faster than non-semantic HTML

@MarkMed cool

@MarkMed I don't think anybody on this planet will care about 3ms difference for a document. If you scale it up by, say, 10 or 100, and the difference persists, and is consistent, on all major browsers and platforms, then you have a case.

@AndSoWeCode the difference is half time taked. semantic.xhtml took 2ms and nonsemantic.html took twice the semantic.xhtml time. If both documents were more heavy, semantic.xhtml would load in 500ms and nonsemantic.html would load in 1000ms.

@MarkMed when it comes to 2-5 milliseconds, don't trust the readings. They don't scale well. It could be that you're right, and there is a 100% overhead. It could also be that this is a fluke (are the results consistent?), or that it has just a slightly longer warm-up phase that doesn't scale at all with document size (is the difference x and 2x, or is it x and x+y where y is constant? Or is it something else entirely?)

Bottom line:

If N is document size, and T0(N0) = 2 * T1(N0) (times for different methods depending on the same document size), it takes a TON of assumptions to conclude that the same is true for T0(N) = 2 * T1(N) for any N value.

Also who says the extra 3ms comes from the longer filename nonsemantic vs semantic

I remember when XHTML was the future. Remember many a day using XHTML strict and getting things to validate. Didn't even know you could use it with HTML5.