Theoretically, there would be a small performance penalty. SATA SSC works by varying the clock rate from the maximum supported speed down a very small amount, and then back up to the max (3.0gb/s). Realistically, the difference in the clock rates between the maximum (with no SSC) and the lowest rate it runs at with SSC enabled is such a small percentage of the available bandwidth that you'd be hard pressed to notice any performance difference between SSC and non-SSC behavior.
SATA SSC works by slowing the clock rate by 5,000 parts per million, over a 30Khz period. This leads to an average slowing of the clock by 0.25%, so that would be the expected performance difference between the two settings.
SAS is a little different with regard to SSC. SAS1 compliant devices (associated with a maximum speed of 3.0gb/s usually) do not support SSC. SAS2 devices support two methods of SSC. The first is Down-spread, like SATA uses, which is subject to the same performance hit as SATA. The second is Center-Spread, which varies the clock frequency both up and down from the target link rate, so there is no performance hit in that case.
In a non-SSC configuration, all the EMI (Electro-magnetic interference) is radiated at a single frequency. Spread spectrum reduces EMI by varying the clock slightly to flatten out the frequency spike across a wider range.