While it was just a small percentage points better in my testing of a divx encoding it did mean about 7-8 minutes which adds up....It may be more inn gaming...ask thugsrook as i rarely game
1.6@2.64ghz w/ 440mhz cas 2.5,3,3,6 = 1.6@2.72ghz w/ 340mhz cas 2,2,3,6
1.6@2.736ghz w/ 342mhz ddr cas 2.5,3,3,6 didn't beat or equal the time of the 1.6@2.64ghz....
I estimated I would have need 2.752-2.768 to equal it...so about 2-3 precent with divx encoding...noticeable only cause there was a clock otherwise maybe not noticeable or neglible in most things...