Not remotely overkill if one has working with a 500GB GIS dataset, or a 100K wide
sat image (IDEX imaging), or a CT/MRI volume set, or any number of very complex
applications that gobble huge amounts of RAM.
What I'd like to see is a pro GPU option that permits scalable memory. Comes
with 6GB default, scale to 600 or whatever.
^^^^
Memory and Bandwidth
> 12GB GDDR5 (6GB per GPU)
Output Connectivity
> Passive Option: 1x Mini DisplayPort and
1x DVI outputs
> Active Fan Option: 4x Mini DisplayPort
and 1x DVI outputs; also ships with
1x Mini Display Port to dual-link DVI
(active) adapter
> AMD Eyefinity multi-display technology is
only enabled on S10000 with active fan
also, if it is using 2 GPU's it looks like the heatsinks are not shared. What will this do for the cooling if the cards are sharing a single airflow channel? it seems to be designed for airflow to enter into the back of the card which means that the first GPU, will preheat the air for the second GPU and cause higher temperatures or worst, cause throttling and inability to overclock.
@razor512: It looks like they're trying to adjust for hot air from the first GPU by changing the fin density (fewer fins on the first GPU and more fins on the second [last in air flow] GPU).