In the last 11 years, a lot has changed in this field. That's to be expected in a technology that is driven by Moore's Law, where complexity doubles every 18 months or so. So we have a factor of 7 or so, but actually the rate of change is increasing.
Where we are today is machines built with multiple cores on one chip, each core being a complete computer. The same communication schemes discussed in this book are used, with the same trade-off's, shared memory or communication channel. We now have the technology to put many cores on one piece of silicon. Along the way, a parallel technology developed, that of graphics processing units. These are found in graphics interfaces to produce screen displays. There was a realization that general purpose computing could use these streamlined devices, and the community developed extensive code libraries for standard graphics cars. Now, multicore architectures are not necessarily ALU based; they contain multiple GPU's. There is a huge amount of computation available, compared to 11 years ago, but it is now chip-sized, not box-sized.
