6 gigabit/s is faster than 3 gigabit/s. It is just like GHz.
No it isn't, and even GHz doesn't mean faster.
When we talk about Gb/s we are talking about either bandwidth or a marketing term for compatibility. When you see a hard drive that says SATA 3.0, that doesn't mean it will run at SATA 3.0 speeds, it means it is compatible with SATA 3.0, however what looks better to someone purchasing a drive with no knowledge of what it means - SATA 2 or SATA 3? All that the bandwidth dictates is the maximum amount of data that can go through that port at any one time.
Consider two hose pipes, one that can have up to 2 litres of water per minute put through it and one that can have 3 litres per minute. If you had the 2 litre pipe running at capacity, so every minute 2 litres of water is coming out of it, but then the same amount coming from the 3 litre per second pipe, they are still having the water flow at the same rate, 2 litres per minute, even though you could put more water through the larger pipe in the same amount of time, up to 3 litres per second.
This is true for the SATA 2.0 vs SATA 3.0 as well, just because the SATA 3.0 with 6Gb/s bandwidth can run at 6Gb/s, doesn't mean it will, the speed of the drive along with the maximum bandwidth of the port will dictate that. Which ever the lower of the two is will be the deciding factor. If you have a hard drive that has a maximum throughput of 2Gb/s, it will perform the same on a SATA 2.0 port as it does on a STA 3.0 port, because the port is not running at capacity, even with SATA 2.0 having a lower maximum bandwidth.
For what you said about GHz, higher clock speed doesn't mean higher performance. For the same family of chip, for instance taking a 3.2GHz Sandy Bridge i5 vs a 3.4GHz Sandy Bridge i5, the 3.4GHz will be faster as they are both on the same architecture. Take a 3.0GHz Pentium 4 though and compare it to a 3.0GHz i5, and the i5 will win out, even in single threaded applications where the additional cores don't work to the processors advantage.
This is due to other factors in the architecture, for instance transistor count, the layout of components, cache size, cache speed etc.
Across families the only way to determine if CPU A is faster than CPU B is to run unbiased benchmarks and look at the numbers that are produced.
