990FX

25% was obviously TOO high.

Hold out for PetrolHead here to discover and reply. He has IIRC an 1100T. I skipped that generation and although it's basically the same for manually OC'ing he has the knowledge on Phenoms.

I'm pretty sure it's the setting. Those BIOS OC settings aren't guaranteed to work and the higher you go, the more likely your system becomes unstable or even unable to boot. If you put some work into a manual overclock, you should be able to reach higher stable OCs than with the BIOS OC settings, but there's no guarantee your system is overclockable by any certain percentage.

If you want to learn about overclocking your Phenom II CPU, Dolk's guide is probably the best place to start:

http://www.overclockers.com/step-guide-overclock-amd-phenom/

I'll help with what I can, but I would like to know the following details:

-Which 990FX motherboard do you have?
-What CPU cooler do you use?
-How have you arrenged your case fans?
-Is your Phenom II a black edition (BE) CPU or a "plain" Phenom II X4 960T?
-Which PSU do you have?
-How many memory modules does your system have?

Some of the programs you might want to use while overclocking:

-CPUIDs HWMonitor
-Prime95
-CPUIDs CPU-Z
-AMD Overdrive
-RealBench v2.41

P.S. Your GPU probably does bottleneck your system. If you are looking for an FPS increase in games, you probably won't get any noticeable increase by overclocking the CPU. You should either overclock your GPU, which could give you a relatively easy 10% increase, or then get a new GPU

Does the BIOS give you the option to increase the multiplier for the CPU clock speed directly? If not, then you don't have a BE. It's not a big issue, just makes overclocking a bit more involved. What you will be doing is increasing the FSB clock speed while decreasing the multipliers for CPU, CPU-NB and HT. Which multipliers and when depends on which way you want to go first. CPU clock speed is king, and the other clock speeds can be considered secondary, but for example Dolk leaves the CPU clock speed last, if I remember correctly. But before we start:

Your motherboard's VRM section should be able to handle some overclocking without extra cooling. It has a nice looking heat sink (the one that says Killer) and it's advertised as having 8+2 phases, although it may be that it's just a doubled 4+1. However, I've no idea how large a margin you have in reality, so I'd still consider even 10% a risk. It's advisable to use a smaller spot fan to provide airflow over the VRM heatsink - the more you OC, the hotter it will become. In fact, your VRM section may be running hotter than normal already. Your CPU had a TDP of 125 W to begin with and the fact that you have unlocked two extra cores means that may have gone past that (although the X6 has the same TDP). I'm assuming your larger coolermaster setup for the CPU is a tower cooler. This may be fine for the CPU itself, but it's also one more reason you need to consider extra cooling for the VRM section.

Considering the temperatures, there are two you need to monitor: The CPU temperature listed under the CPU in HWMonitor (sometimes labeled "package") and the CPU socket temperature listed under the motherboard in HWMonitor (usually labeled CPUTIN). For the Thuban generation of Phenoms, the CPU temperature should not exceed ~62 C, while the CPU socket should be okay as long as it stays below  ~70 C. The 40 C you have seen may be a good value or a bad value, depending on what you were doing on the computer when it occured.

By the way, disk corruption can occur if you have an unstable overclock. It can also happen slowly over time, even if your computer is stable enough not to crash. This is why it's a good idea to stress test the system properly when overclocking, although it should be said that stress testing isn't very kind on your system and can be harmful in itself - especially if you don't take your time when overclocking. I don't know if your previous overclock is the cause for the disk corruption, but you should be aware that something like this - or worse - can be the result when you overclock.

You should begin your overclocking by establishing a baseline for your system. Start with stock settings (keep the two extra cores unlocked, though), run HWMonitor and then run Prime95's Blend test on six workers. Keep an eye on the two temperatures I mentioned. If either one gets close to the figures I've stated, stop all workers immediately and report back. If this happens, there's an issue with your system's cooling. If the temperatures seem to level out at safe values, let the test run for at least a couple of hours. If everything runs great, make a note of the temperatures and post them here. If you lose any workers due to an error before two hours are up, you're not providing the CPU enough voltage. This is something I'm actually half expecting to happen. That X4 CPU of yours is basically a Thuban generation X6 that has failed AMD's tests, so it may be that all six cores won't work as well as they should on stock voltages.

In short:

1. Understand the risks.
2. Don't forget the OC affects other components as well, especially the VRM on your motherboard.
3. Determine a baseline before you start.

I ran it a few hours trouble free, no more heat than regular web surfing, I dropped the BIOs OC and increased it using AMD catalyst, I did keep the core unlocker up tho. BTW this is my chip. http://www.newegg.com/Product/Product.aspx?Item=N82E16819103995

I think what I've done is pretty much just open 2 extra cores, removed the overclocking and the BIOs are all set to auto control the voltage. I've been getting frustrated with the disk error startup, was able to find a BIOs update that was directed toward the M.2 harddrive, thats where I put my OS, I'm hoping that is dealt with, then I may get back to OC, or may just abandon that all together, it's not necessary.