Multi-Core Mythbusters

Bit late picking up on this but Ted Neward has written an article on the myths of running apps on multi-core systems. While he makes some good points, I think it falls slightly short in addressing some of the issues raised. This is quite pertinent for me as I work in a place where we seem to be buying up quad core billion processor machines with little regard to how well the applications being deployed will make use of multi-core CPUs. I think developing for multi-core machines in future is going to require much more thought and this could be one reason why people are looking at languages such as Erlang. The days of "we'll just throw more hardware at it" are at an end, while I am a firm believer in avoiding premature optimisation - I also think performance is an important architectural concern.

links: digg this    del.icio.us    technorati    reddit