Raymond is definately making some great headway. Having targetted the single-character at a time conversion problem in version 4 of his program he's 1.84 times faster than the previous version.
Unoptimized Managed port of v1
As before I'm including the function costs in tree view, but only showing functions with an inclusive cost of at least 5% to keep the size of the report manageable.
Function Name (Sanitized)
So looking at this new program it seems like its time to start targetting string management and the allocations. We're spending 24% of our time just allocating strings. Now what's interesting is that at this point things are fast enough that the tear-down at the end is getting to be signficant -- at 18.84% it's a good chunk of the cost.