Does anyone know whether TBB can work on the multi-processors machine? My laptop has a Core 2 Duo chip, and TBB works very well on my machine. I hope that I can know my code can scale to 4-core chip or 8-core chip, and get the experiment result. But I don't have access to 4-core chip or 8-core chip. If you TBB can work on the multi-processor machine, then the problem will be much easier.
For sure TBB works on multi-processor machines with shared memory.
Yes, if possible you should check if your code scales well to 4 and more cores. There are different aspects affecting scalability; and while TBB does its best to eliminate some of those, there are also some it can do nothing with (such as bus saturation for example). So if you have access to a multi-processor system (including one with a few multi-core processors), it is completely reasonable to try your TBB-based program on it.