This is just speculation and reminiscing, but in my older days (PDP 11/45 - mid 70's), when we had problems with a certain program that we couldn't pin-point, we found loading programs in different order, hence forcing it to a different memory location (bank), sometimes help isolation. Yes, I did a lot of work with the US Navy and that was a "Navy sentence."
Many times we had to run memory diagnostics that were "relocatable". They would load in high memory and check low memory or the opposite (of course this is a simplified explanation, but I think it makes the point -- loading the diagnostic into the area where the memory problem was, doesn't work very well.) We had three types of memory on those old machines (core, Bipolar, and MOS) all with different timings, and we had to worry about crosstalk, timing, core heating (you could actually burn up a core memory board with a tight loop) and "memory boundaries." Never have figured out why we called those "the good old days", but we did.
Anyway, there are some good modern pc memory diagnostics that will check all your memory, including where your operating system would reside -- I think "Memtest86+" is one, but can't remember.
My advice is to try to isolate the problem without going into you computer. Of course you may have to if the problem is over-heating due to dust, etc. Do the low risk stuff first. Then try to check one thing at a time. We used to have a lot of "re-seating" problems, but if you swap, you are "re-seating" too, so which one was the real problem, the swap or the re-seat?
So, when you ask a question like "Computer Shuts down during rendering", you are going to get the heat, insufficient power, bad memory module, and many other responses that can cause the problem you encountered. Terragen has a large enough user base that you can pretty much figure that a lot of users would be reporting this problem consistently if it was a coding fault. The "stress" answer is the popular answer because heat is a common cause.
