Hey vx,
I havent had a chance to run prime95 in the past few months, so this also looked a bit weird to me. I proceeded to download the latest version of prime and test it myself to see if infact a change has been made for it to take 5 hours to finish one test. But unfortunately the problem isnt that easy to explain.
However, i can try to narrow down the problem for you.
1) Your CPU might be throttling and skipping many itterations therefore taking a long time to complete one test (but this is unlikely as i see from your sig that u have adequette cooling and your temps are fine)
2)You have many processes running in the background which alone take up most of your cpu power therefore not letting prime95 to perform at full speed.
3)Your settings in prime95 could have been altered mistakenly and therefore set to take this long. For us to help you out more with this you would have to post your settings here, and also inform us what type of test you conducted.
4) I have a huntch that it could have to do with those big numbers that were being tested. It looks like you are testing 560,000 Lucas-Lehmer iterations using only a 10K FFT length. When I tested my comp with prim95, it took me only a few min to complete one test testing 400 Lucas-Lehmer iterations. Maybe someone more informed on how prime95 works could help explain this better.
Im also curious as to what version of Prime95 you are using?
Raven
EDIT: From reading the undoc file that came with prime i came across this
You can control how many minutes each FFT size is tested by the torture test.
In prime.ini set:
TortureTime=n
Check that to see what your TortureTime is set to. That could explain a few things.
This can also be changed when you use the custom torture test options. Take a look at that and see what is there. Options > Torture Test > Custom