A good and useful way of doing a sort of torture testing is to perform a double check. If it matches the first result, we can be pretty sure the machine has been reliable. If it doesn´t, well, the original LL might be faulty in the first place, so, especially if our run was error free, we may perform another test on a different exponent, and only then, if we still have no matches, perform the torture test. The good thing about this is that we are using the clock cycles to perform some real work. Every time I put a new machine to work I always start by doing a couple of DCs, to be on the safe side.

[QUOTE=lycorn;193964]A good and useful way of doing a sort of torture testing is to perform a double check. [/QUOTE]

Sure. OTOH, there is a reason the first screen gives you two options: "Join GIMPS" or "Just stress testing". Prime95 is famous in the OC community for its torture test capability. People aren't going to join GIMPS, and wait to complete a double check just to see if their m/c is stable.

I think the original question got lost in the shuffle, so I'm going to attempt my own answer.

Programming is hard, and that includes placing onscreen buttons. If you grey out buttons that can't be used, then it's probably less stress on the programmer than removing the button altogether every time it can't be used. Besides, if they removed the button, they'd either have to deal with the blank space or there'd be a new question,"Why the hell is there blank space in my Prime95 window?"

Hope that's helpful.