overheating - Laptop overheats, BSOD, turns off when playing games

Recently, my year-old laptop began BSODing, turning off unexpectedly and heating up very fast when playing games. This is regardless of the integrated graphics or the add-on NVidia graphics used.

My specs:

Model: Samsung RF711-S02-US 
OS: Win 7 Home Premium 64-bit SP1 
CPU: Intel Core i7 2630QM @ 2GHz (current temp 78°C - 81°C)
RAM: 4GB DDR3 @ 668MHz (9-9-9-24)
Motherboard: SAMSUNG ELECTRONICS CO., LTD. RF511/RF411/RF711 (CPU 1) (current temp 80°C - 84°C)
GPU: Intel HD Graphics Family + NVidia GT540M
HDD: 699GB Hitachi (SATA) (current temp 41°C)

It looks to me like a heat issue since it lasts longer when I use low game settings and focus an external fan on the left side of the laptop (where the grills are, the area where it becomes hottest).

This was never an issue on the first few months of usage as I was playing WoW/SC2 Starter most of the time and doing file transfers overnight, so the laptop is on almost 'round the clock.

Recently though, when my son started playing Skyrim, it would BSOD with DRIVER_POWER_STATE_FAILURE. More recently, it just turns off.

During those times, I would notice that the left side of the laptop, the grills under the power cord, puts out a lot of heat.

I am not sure how to get this fixed (should I have this serviced already?). I'm concerned because this is just a year old. Should I consider re-seating the CPU?

The temp readings I mentioned above are from Speccy and I'm not even playing any game. I have 2 Chrome tabs and 4 explorer windows.

Any idea how I can solve the issue? Other smart recommendations to make my laptop last longer?

Answer

Processor:

Download the Intel Tuning utility to see if the processor is actually throttling. Many temperature monitors are not precise and provide temps based on the Thermal junction max, which isn't always accurate. I've found it is best to monitor the throttling if I suspect a chip is over heating and not the core temperature. See below for more info.

According to Intel the formula for core temperature is: Tjunction max - DTS (digital thermal sensor) = current thermal junction temperature (which we see as core temperature in applications). DTS is based off of distance to tjmax, so if it reports 20 and the tjmax is 91 your current temperature is 71. Most monitoring software is calibrated to a tjmax of 91. Why is this an issue? Well take a look at the specifications from Intel, which states the thermal junction target is 100, not 91.

The developers of Real Temp go into more detail about why it is 91 degrees and the difference between TJmax and TJ target... This is why I generally monitor thermal throttling when trying to figure out if a processor is over heating.

Graphics Card:

You don't give any temperature readings for your graphics card so it's a possibility that it may be over heating. First get a utility to see the GPU temperatures so we have an idea of how hot it is getting. EVGA precision or MSI afterburner work well for this purpose because if it is over heating you can under clock memory/gpu. I've found that turning the clock down on the video memory frees up a little thermal room with no noticeable performance loss... Also note if it is shutting down while you are not gaming, as this is a optimus laptop the dedicated GPU 'shouldn't' even be on when your just browsing the web.