The proliferation of multicore processors has reinvigorated an 
interest in alternative models for programming that make it easier to 
express concurrency and utilize parallelism. Multicores are relatively 
easy to architect but they significantly increase the burden on 
the software stack to deliver the promise of Moore's Law to end users. 
The burden on software will increase as the processor landscape is 
likely to feature a wide range of systems that include homogeneous and 
heterogeneous multicores as well systems with reconfigurable 
accelerators and FPGAs. 

In this talk, I will draw from my experience with the Cell processor 
to highlight the challenges in programming multicore processors. I 
will argue that the existing software stack is an impediment to 
programmer productivity, and argue that while new programming models 
are a necessity, it is also necessary to reevaluate the software 
stack, and blur the many existing boundaries between software and 
hardware. 

I will conclude the talk by presenting Liquid Metal, an end-to-end 
system from language design to co-execution on hardware and software. 
The goal of the Liquid Metal project at IBM Research is to allow 
hybrid systems to be programmed in a single high-level and object 
oriented language that maps well to both CPUs and FPGAs (and 
everything in between). While at first glance it may seem that these 
different systems have conflicting requirements in terms of 
programming features, it is our belief that many of the features turn 
out to be highly beneficial in both environments when they are 
provided at a sufficiently high level of abstraction. By using a 
single language we open up the opportunity to hide the complexity of 
crossing domains from software into hardware, and facilitate a fluid 
movement of computation back and forth between different types of 
computational devices, choosing to execute code where it is most 
efficient to do so.