As many comments suggest, there are those for whom parallel programming isn't hard, but that is a small minority of the programmers in the world. Most of the code is use today was not written by those elite few. Unless we want to limit the continuation of Moore's Law to only applications written by those few programmers we need a better way. RapidMind has created a development platform that simplifies development of parallel applications for vector and multicore processors. Using RapidMind, the programmer masses can develop parallel applications using their existing skills, compilers and IDE's. The RapidMind runtime platform takes care of the parallelization and load balancing across any number of available cores. In addition, code written using the RapidMind platform will run unchanged on NVIDIA and ATI GPGPUs, IBM Cell BE and Multicore Intel and AMD processors. Even the elite few that can parallel program would have to port and refactor to support multiple or future processor architectures. Just as higher level programming languages like Visual Basic and now Ruby have enabled laymen to be productive programmers, a similar level of abstraction will be needed in the parallel world just to keep existing programmers productive. Check it out, it is free to download and try @ www.rapidmind.net
Slashdot Mirror
As many comments suggest, there are those for whom parallel programming isn't hard, but that is a small minority of the programmers in the world. Most of the code is use today was not written by those elite few. Unless we want to limit the continuation of Moore's Law to only applications written by those few programmers we need a better way. RapidMind has created a development platform that simplifies development of parallel applications for vector and multicore processors. Using RapidMind, the programmer masses can develop parallel applications using their existing skills, compilers and IDE's. The RapidMind runtime platform takes care of the parallelization and load balancing across any number of available cores. In addition, code written using the RapidMind platform will run unchanged on NVIDIA and ATI GPGPUs, IBM Cell BE and Multicore Intel and AMD processors. Even the elite few that can parallel program would have to port and refactor to support multiple or future processor architectures. Just as higher level programming languages like Visual Basic and now Ruby have enabled laymen to be productive programmers, a similar level of abstraction will be needed in the parallel world just to keep existing programmers productive. Check it out, it is free to download and try @ www.rapidmind.net