In my spare time, I have been playing a bit with GCD, and the R package rMAT. My first impression is that it’s very easy to use and works as advertised. I have actually used it to parallelize the normalization processing of multiple tiling arrays. This will soon be incorporated in the R package rMAT, available from Bioconductor. In comparison to standard parallel computation in R done via, for example, snow and/or snowfall, with GCD you do not have to worry about cores/processors initialization. In particular, you do not have to specify the number of cores to be used, the optimum number will be decided at run time depending on resources available. The downside, is that currently, such optimization can only be used on Snow Leopard (and of course a Mac). If you do not have Snow Leopard, nor a Mac, then don’t worry, the code should still compile on your machine using the serial version. A machine specific compilation can be done using #ifdefs such as:

#if defined(__APPLE__) || defined(macintosh)

#ifdef __MAC_OS_X_VERSION_MIN_REQUIRED

// Use Grand Central on Snow Leopard

// note use of 1060 instead of __MAC_10_6

#if __MAC_OS_X_VERSION_MIN_REQUIRED >= 1060

The bottom line is that it’s great, and it’s very likely that future code/packages coming out of our lab will take care of such optimization. So perhaps you should buy a Mac .