So we have Cuda from Nvidia, OpenCL from Open community supported by ATI, Apple and others, and later this year Intel will introduce beta of its own parallel language called CT. This will definitely won't help healthy growth of the parallel programming community, as most of developers won’t be sure what to use as its base for parallel programming.

Since this looks like the dawn of parallel computing, we will have to wait and see, Nvidia is pushing hard on Cuda and it has a huge advantage as it started a year before everyone else and Intel’s CT won’t launch in beta before the end of 2009, which was confirmed by Pat Gelsinger at IDF Beijing Keynote.

Intel’s Larrabee is the chip Intel needs for CT as it is based on many parallel cores and it should do this parallel job well. Intel already said that Larrabee supports OpenCL so if industry goes this way, Intel should be ready.

Intel told the world that Intel CL is a natural extension to C++ and it hopes for a success and we are sure the time will tell if CT, OpenCL or Cuda are the right way to go. Intel showed some promising results in parallel applications done in CT but since the beta software, it won’t come until late 2009. The timeframe is crucial, as only in late 2009 Intel actually should have Larrabee at least in beta phase running. Until this happens, you won’t see the language either.