I decided to buy a cheap temperature gun (Ryobi IR002, Home Depot) and test my card to see what kind of temps I was getting after the new BIOS:
 
After running 20 minutes of the latest Furmark release, I started taking measurements while the stress test was still running.  
 
-Highest reading I could get from the back of the board was 84C. (83.9) No thermal pads. Just the new BIOS.
 
Just thought I would share this for others who have non-back-plate boards. 
 
Edit: will post pics once photobucket is done with their "maintainance".
 
This is 20 minutes into Furmark and still running. It only seems to let me use one photo, so here is the hottest measurement I found anywhere on the card:

 
 
 
 
 
 
 
 

How loud is the card now?  I have the same card and was so happy with how quite it was.
 
Just a bit more and you can boil water.  ^_^

Athena
How loud is the card now?  I have the same card and was so happy with how quite it was.
 
Just a bit more and you can boil water.  ^_^

I think my dual-fan CPU liquid cooler makes more noise when gaming. My 780 with the standard blower motor was a LOT louder.
 
So the card isn't as quiet as it was, but it's still a pretty quiet card.
 
Although my case has good air-flow, I hear Furmark is the worst possible thing you can ask of your GPU, so I'm very happy with the results, I really think the cards with back plates and no pads are the hottest ones. 
 

Thanks for the quick reply and good info!!  ^_^

FWIW, The gun has an accuracy rating of +/- 5 degrees And I think it was reading on the high side.
 
Every time I shot the GPU I was reading a few degrees higher than the number displayed on-screen from the built-in sensor.

Still to high....if the back of the board is already at almost 90c, the actual VRMs are running well over 100c+

These cheap guns are quite inaccurate. Tomshardware used a thermal camera thats worth 4k€. I doubt a lot of people have such a camera at home, tho. 

pawelblyskal
Still to high....if the back of the board is already at almost 90c, the actual VRMs are running well over 100c+

 
Um, no.
 
The internal GPU sensor was showing 68 and when I shot the back of the card right where the GPU is located....The gun was reading temps in the low 70's. So, while the accuracy of the gun may be in question, it consistently showed GPU temps a little higher than the reading the internal sensor was sending to software.
 
I suppose you could argue that the GPU sensor was ALSO faulty and reading very low, or that the area on the PCB behind the GPU does the opposite of what the rest of the PCB does, or that the gun was changing calibration from reading high over the GPU to reading crazy low behind the VRM's, chokes, and VRAM chips....every single time I alternately read the different areas?
 
But none of those scenarios are likely.

NeroRay
These cheap guns are quite inaccurate. Tomshardware used a thermal camera thats worth 4k€. I doubt a lot of people have such a camera at home, tho. 

I agree.
 
That's why I measured the back of the GPU with the gun and compared the gun's measurement to the internal sensor's readout.
 
Gun measurement on the back of the board was  inaccurately high compared to the internal sensor on the GPU. (Even then it was close enough to know that I don't need to freak out about my VRM or VRAM temps.)
 
I didn't check, but I don't recall Tom's test showing any great discrepancy between GPU's internal sensor and their FLIR readings. 

The back of the card will always be hotter than the front.
 
Why? The HSF is on the front, not the back.