Nobody buys a compiler to calculate the values generated by the whetstone benchmark[1], everybody already knows the answers.
The point is not to make a mug, the mug is a benchmark. The point is that it takes a nontrivial amount of time, the model readily available, it is repeatable, and it has some very interesting properties WRT complex curves and overhanging structures.
(Like all benchmarks, we can argue from now to forever whether it is a good benchmark.)
The point is not to make a mug, the mug is a benchmark. The point is that it takes a nontrivial amount of time, the model readily available, it is repeatable, and it has some very interesting properties WRT complex curves and overhanging structures.
(Like all benchmarks, we can argue from now to forever whether it is a good benchmark.)
[1] http://en.wikipedia.org/wiki/Whetstone_%28benchmark%29