fatslob-:O on 21 January 2014
The point of AMD kaveri is to accelerate GPGPU workloads in cases where benefits may be seen with low latency access towards the GPU with moderate amount of concurrency happening in a program.
The problem that arises with AMD kaveri is that there is almost no software developers willing to take advantage of it but I suspect some things will change once openCL 2.0 arrives to the scene.
All in all it doesn't make sense right now and I doubt software developers will be taking advantage of it any time soon so even in the short term it still I would still say that it does not make any sense.
