EDIT:
You were right about EEDI2 using OpenMP and being dynamically linked to the libraries. I just recompiled the source with the OpenMP libraries statically linked:
EEDI2 64bit Multithreaded
Also, where did you run across RemoveGrain64? I'd like a copy (and the source if possible) so I can figure out what exception it's throwing.
The only thing I can say for sure is that converting RGB error is hard coded in there for now, as there was a decent amount of assembly involved in converting the routine, and I just plain didn't feel like it. I'll get back around to it, put that on my todo as well.
I'll start checking into those other two plugins, I'm working on MVTools at the moment . . . that would be a huge win.
The other issue with the code paths was that I was basically telling ICC to target just my host CPU. When I tried to get it going on any other computer I realized my mistake. The current DLL for download has code paths for all EMT64 supporting Intel processors.
Once again, don't know if ICC cripples AMD chips or not . . .