Here are the results from my temperature testing on my new gaming rig, using Everest Ultimate, Prime95, and Fur (allowing cool down in between tests).

Please give me some feedback on the results.
I just want to know if I should do anything to cool it down more (like install HSF or GPU HSF) or just not worry about it?

Also, a question: Should I be looking at the CPU temp in Everest or the CPU/Core1, 2, 3, and 4 temps (which are much higher)?

Setup:
Q9450 (Stock cooling)
P5N-D mobo
8800GTs in SLI
2 HDDs
Cosmos 1000 case (without the VGA cooler panel installed due to SLI) - closed case
Room Temp = 23-25 celcius
Room Humidity = 55%

Idle Temps:
CPU Fan Speed = 2033 rpm
GPU1 Fan Speed = 1388 rpm
GPU2 Fan Speed = 1367 rpm
CPU = 45
Core1 = 61
Core2 = 56
Core3 = 60
Core4 = 62
Mobo = 49
GPU1 = 74
GPU2 = 72

Prime95 for 15 minutes (on blend setting):
Load Temps:
CPU Fan Speed = 2848 rpm
GPU1 Fan Speed = 1429 rpm
GPU2 Fan Speed = 1404 rpm
CPU = 64
Core1 = 75
Core2 = 70
Core3 = 74
Core4 = 73
Mobo = 50
GPU1 = 75
GPU2 = 73

Fur for approx. 10 minutes straight (6 tests in a row windowed mode @ 1600x1200):
Load Temps:
CPU Fan Speed = 2935
GPU1 Fan Speed = 4386
GPU2 Fan Speed = 4386
CPU = 56
Core1 = 70
Core2 = 63
Core3 = 66
Core4 = 68
Mobo = 54
GPU1 = 96
GPU2 = 75

For stock cooling that's fine. SLI doesn't appear to be working in your "fun" scenario based on the GPU temps. As for which temps to look at, the core temps are both more important and less important than the CPU temp, depending on context.

1) Core temps are closer to the actual temp of each core assuming they are relatively accurate and are being read correctly. There are plenty of reasons why they might not be being reported right.

2) CPU temp is important for controlling fan speed, and is the only temperature that Intel cares about on desktop processors. Try to RMA a chip because your core temps are being read incorrectly and expect an Intel rep to deny your RMA. Intel does not provide the core temps for you to read, they are there to control processor overheating prevention functions like throttling and critical temperature shutdown (instead of meltdown ).

3) CPU temperature is calibrated by the BIOS, so it is going to be reported differently from one BIOS version to another. Core temps are calibrated by Intel.

I recommend reading the Temperature Guide sticky which has recently been re-posted.

Thanks for the information, but:

Do I have to worry about these temps (CPU, mobo or GPU)?
Or should I be looking to lower them by using better cooling HSF, etc?

Also, SLI was and is engaged in the Nvidia control panel. Just not sure if Fur (not Fun), which is a program to stress test GPUs, takes advantage of SLI. Probably not because as you alluded to, one GPU is high while another is low.

Fur... fun... I think I must be blind.

Those temps are fine for stock cooling with artificial load (prime95). Unless you intend to run Folding@Home SMP clients you won't hit those temps under "normal" usage. Even if you did, they are hardly dangerous. Warm, yes, but not unsafe. Reseating the CPU cooler can do wonders for temperatures too, just make sure you replace the thermal paste each time you do that, cleaning off the old stuff well.

The GPU temp on the other hand is rather high, did you buy those 8800GTs when they first came out? If you did, they may not have the BIOS update which fixes the "stuck" fan speed.