[Tinyos-devel] nesC concurrency semantics, volatiles, etc.

[Philip Levis]

On Aug 12, 2006, at 10:03 AM, John Regehr wrote:

> I hear you, but application-specific performance metrics are pretty
> difficult to deal with.  It would be nice if those of us doing tool  
> work
> had a better sensornet performance metric!  Actually this would be  
> more
> broadly useful, for example help help folks decide how fast to run  
> the CPU
> or to decide between AVR, ARM, MSP430, etc.

Good point. I think that rather than app-specific performance, the  
issue can be simplified a bit to the rate at which you can  
successfully handle events. E.g., the packet transmission rate at  
which a null AM handler drops packets, the continuous ADC sampling  
rate at which you miss interrupts, or the sequential ADC sampling rate.

Establishing such a set of metrics could lead to very directed  
improvement and optimization (e.g. spec). This has the benefit of  
leading to really deep technical excellence, but also has the  
drawback of sometimes leading a community to make those metrics  
themselves the end goal.

Phil