Sieve C++ has been designed around the way YOU write software, not around new unproven concepts about how you should write parallel code.  Using Sieve C++ tools is a stepping stone into writing for multicore without changing your toolchain.

1.  Using your IDE, write of modify your C++ code using SieveThreads to offload the parts for execution on an accelator or coprocessor.  You can use Codeplays Analysis Tools to assist you.

2.  Compile the Sieve C++ code using the Sieve compiler.  Sieve C++ has full integration with Visual Studio 8 (integration with Eclipse and other IDE's can be done for you - contact us).  The Sieve compiler can also operate as a stand-alone command line tool or with a makefile.  Sieve will compile you program into auto generated C code through it's OutputC back end.

3.  Compile the generated C code with your current C/C++ compiler.

4.  Using your existing linker, link the object files with the Sieve runtime library for your processor architecture.

5.  Test and debug your software using your existing debugger.  Information from the Analysis Tools and Sieve Compiler will compliment the capabilities of your existing debugging process.