Can't track source of overheating Intel i7 2700K

I am sorry for the long tale, but I don't want to accidentally leave something important out. This is also my first build so please forgive me for any near-offensive incompetence on my part.

So my original rig was as follows:

• 
  Cooler Master Centurion 5 Case
  Intel i7 2700K (w/ Stock Fan + Heatsink)
  Intel DZ68BC Motherboard
  Nvidia 550 GTX Ti Graphics Card

The build itself is 4 years old, and throughout its life I do remember periodic full system freezes under Ubuntu 12.04 and 14.04. Over time it seemed to become more and more frequent until the past few weeks where I could barely even run simple matlab and matplotlib visualizations. (I should mention I mainly use it for my engineering work, but I occasionally play games on a separate windows hard drive).

My first thought was that it could power supply going bad (for whatever reason). So I ordered an EVGA 500W. While I was shopping I thought why not upgrade the graphics card too, so I also ordered a Nvidia 750 Gtx Ti.

Now here is where my mistakes compounded on each other.

Mistake 0: I had never cleaned out the dust out of the machine since first building it, nor did I ever check the operating temperatures of the CPU

So I cleaned all the dust out of all the corners and even removed the heat sink and fan to dust around the cpu.

Mistake 1: I forgot to reapply new thermal paste before remounting the heatsink. It is also possible that I was a little rough removing the heatsink because I completely forgot about thermal paste since I first built it.

So after installing the new power supply and graphics card I unsurprisingly start having thermal trips after ~ 1 minute max running.

Mistake 2: Not looking at the motherboard manual to figure out what all the beeping meant. In this time I tried reverting the old GPU and then to no GPU, etc.

After finally figuring out what was happening, I go and buy the Arctic Silver 5 thermal paste, cleanly remove the old paste from the CPU and heat sink with rubbing alcohol and coffee filters, and apply the new thermal paste. I am quite confident I have done this correctly.

At this point I am starting to become perplexed, as I was able to complete a memtest86+ last night (which contained no errors), but shortly after I could not even stay in BIOS for long without another thermal trip. I even waited overnight to try again to let the system quiesce. The ambient air of the room is just under 70 F.

In the time I am now able to stay in BIOS, I see the CPU is operating at a blistering 98 C with almost no load. The CPU is not overclocked, it is receiving .94 V, and all BIOS settings are at default values. The CPU fan is also running at 2000 RPM and the rear case fan is also running (3-pin).

I have just ordered a new heatsink +fan (Cooler Master Hyper 212 EVO) along with a new 4-pin rear case fan, which should hopefully arrive tomorrow.

I am starting to think this story ends with me getting a new CPU which I would hope to avoid, but before doing that I thought I'd finally ask more seasoned folks for their thoughts and advice.

Thank you all so much for your time in advance, any help would be appreciated.

As sure as I can be, the rig feels flimsy in general. I can see that, and am starting to suspect that as well.

Great! I'll post my update tomorrow when I get the new cooler.

Hey all,

Sorry, shipping was delayed a day.

So I've installed the new 212 evo, and even threw in a new PWM rear fan. I've been running the real-time sys monitoring in BIOS for about 30 minutes now and the CPU is at about 51 degrees C.

Is it just me or does that still seem a little warm?

Thanks in advance

So great news. I just decided to go for it and installed my OS, Ubuntu 14.04. Using lm-sensors and cpuburn I've found the cpu is actually running at around 35 (actually 30 this morning after letting it run all night).

Stressing all 4 cores with cpuburn also only ever puts me up to ~51.

Although, I still don't know why the BIOS was just that far off.