30 Apr
2009
30 Apr
'09
2:57 p.m.
I have figured out an optimization that in simple testing makes integer_sort faster than std::sort on already-sorted data (and comparable on mostly-sorted data), and improves performance slightly on unsorted data. As the worse best-case performance was the main drawback to the algorithm based upon my prior results, I was wondering if it was appropriate to delay the review until I can get this change fully implemented and tested? If it is proper to delay the review, I apologize for wasting people's time. I thought I was done optimizing it, but apparently not. I think I can finish all this up by monday.