The presentation fails to mention if the Direct3D 12 improves on the FPS or score of the 3DMark other than dividing the workload evenly between CPU cores, it could be that 3Dmark demo is pretty much over-saturated the GPU. One thing I dun like about 3DMark is that they dun reflect real games. 3DMark just throw zillions of triangles to stress the GPU whereas games are optimized and reduce the triangles sent and rendered, and balance the act of achieving same level of aesthetics with less triangles and better FPS.
@Charles: Keep C++ TV video length at 40+ mins, ok? Because I always couldn't wait till I get home to watch the STL videos. I usually watch it at workplace during lunchtime. 15mins lunch and 45 mins of STL goodness!
Hi STL, excellent video on shared_ptr. I always look forward to your STL videos. I have some STL questions I hope you can answer. My user suggested to me to implement iterators for iterating elements in my open-source xml library. Is it possible to implement my own iterators to work with the STL algorithms in VC8, VC9 and VC10? It will be best if my custom iterators can work with algorithms in all STL implementations out there.
I am thinking of rewriting my own next_combination algorithm to take advantage of bidirectional and random access iterators. Right now, it only uses the least-common denominator iterators (bidirectional iterators) which is slow because the algorithm has to increment íterator 1 by 1, instead of 'jumping' to the required iterator. Is it possible to write my algorithm to work with all VC(8,9,10) STL iterators or all STL implementations? Are the iterator trait tags the same?