K6-2 CPUs are generally considered "Super Socket 7", but technically there is no such thing as "Super" Socket 7, it's a marketing term for a bunch of features not directly connected to the CPU socket, such as SDRAM support, AGP, and 100MHz bus speed. Even the most advanced "Super" Socket 7 board is technically just a Socket 7 board.
So given every "Super" So7 board is an So7 board, there is no difference by definition.
As Munx says, voltage is the biggest limiting factor for K6-2 is voltage, as most K6-2 CPUs need 2.2V, and K6-2+ is specced for 1.8V. They can generally survive slightly higher voltages, but 'standard' 2.8V or worse 3.3V singl voltage So7 boards will roast them.
However, if your old So7 i430HX board supports 2.2V (as later revisions of Asus' P/I-P55T2P4 boards do), it will happily run a K6-2. Moreover, the later K6-2 cores (CXT - 400MHz onwards) interpret a 2.0x multiplier as 6.0x, so even if your motherboard only does 66MHz, it can run a later K6-2 at 6x66=400MHz, and with higher bus speeds even higher (6x75Mhz=450MHz, 6x83MHz=500MHz).
With K6-2, both L2 cache and RAM run at CPU bus speeds so 6x66MHz will perform worse than 4x100MHz. The difference will be smaller though with a K6-2+ or K6-3 as they have on-chip L2 cache, which runs at full CPU speed regardless.
AGP doesn't make a huge difference with period-correct cards, early AGP cards hardly used the advantage.
Finally, SDRAM outperforms EDO, but in fact the clock-for-clock fastest SDRAM So7 implementations weren't the last-generation "super" 100MHz chipsets but some of the previous generation such as the i430TX and the ALi Aladdin IV+.