Problem with sample_transcode* between SW and HW DLL and

Using the current tutorial sample_decode and sample_decode-d3d, this is my problem I am working on getting around.

The attached file (I named it t.264) runs very differently through the HW and SW DLLs.

 simple_transcode and -d3d (in the current tutorial) using only the SW DLL

   simple_transcode.exe  (SW DLL)
   Frame number: 691
   Execution time: 3.70s (186.58fps)
 
   simple_transcode-d3d.exe  (SW DLL)
   Frame number: 21
   Execution time: 0.19s (109.49fps)
 
Each same EXEs run using HW DLL v 10.18.10.3349  dated 3 Nov 2013

   simple_transcode.exe  (HW DLL)
   Frame number: 2
   Execution time: 0.02s (98.11fps)

   simple_transcode-d3d.exe  (HW DLL)
   Frame number: 2
   Execution time: 0.01s (142.31fps)

The t.264 source file was created by the IM SDK encoder.

The 691 frames is correct.

The only change to any of the source files (src, common, whatever) was the line

    fopen_s(&fSource, "bbb1920x1080.264", "rb");

was changed to

    fopen_s(&fSource, "t.264", "rb");

Now here is hoping I can attach the .h264 (715 KB only).  If not with this I give it a shot in a new post.