I'm confused. You said "I succeeded in compiling a simple "hello world" application on the SCC using visual studio."
One doesn't actually compile on the SCC. Standard usage is to compile on the MCPC using a cross compiler. The resulting executable does not run on the MCPC. It is loaded onto the SCC cores with pssh or rccerun. Or one can ssh to a core and execute the file directly. Running a program on the MCPC is just running on a standard Ubuntu system.
There's been some early work on an SCC plugin for Visual Studio. This creates a baremetal SCC application. By baremetal we mean you can run it on SCC cores and not have an OS running on the cores. Non-baremetal applications require an SC Linux to run on the cores; we provide that SCC Linux.
Visual Studio, of course, runs on a WIndows system. It results in a file on the Windows system. The MCPC must read that file and load it on the cores. There is some non-publicly available sw that lets the MCPC so this. I don't know how WinGDB fits into this mix. Can you be more specific?
I'm sorry, I was wrong. I've thought when I could run the application on the mcpc, it can automatically run the application on the cores.
At any rate, I think , perhaps we can also build an application using WinGdb on the SCC, because there is some options that supports the case of two linux machines : (debug host , and remote target ). It might match our case (debug host : mcpc and remote target: SCC ), but in this case we need to have a TCP connection from our win pc to the SCC cores!
Thanks a lot for your helping.