So what are you saying here DBZ? Should I dismiss the whole idea of maybe SLi-ing the two card I have and be satisfied with just using the 4GB's one?
SLI still has the advantage of double GPU power- or rather, halving the workload (frames rendered) of each GPU...it's just with dissimilar specced cards you don't achieve the
full benefit of the higher specification
Because if one card is going to render fast and the other card is going to take some time, I don't know if I quite like that idea.
Nvidia's driver won't allow you to run an asymetrical specification for that exact reason. My comment:
so there is nothing to be gained by one card being faster than it's partner(s), it just induces lag - one fast frame rendered followed by a pause before the slower card completed it's frame.
...simply notes the reasons why Nvidia (and AMD) approach their optimizations in this way.
Hit me with some truth here, if this was your build, and knowing everything that is on the part list, would you run the two video cards at once, or would you just run the eVGA 4GB's only?
You're probably asking the wrong person. I've been a near-continuous user of SLI since 3dfx owned the IP, and probably biased towards the you-can-never-have-enough-horsepower camp. Triple monitor setups require a dual (or more) GPU setup- you just wont realize the full potential of one card (for vRAM). Dissimilar clocks can be overcome by applying a small overclock to the slower card.
Short answers:
As per Blkfx1's adviceP: Corsair H80 = noisy. no better than mid-high end tower (air) cooler. H100 = a little better than high end air cooling. Don't run too much voltage through the CPU and it should be fine...and ditch the supplied fans.
They aren't that good
Bespoke water cooling much better but out of your price range if you're shopping ~$100.
Even a 2x120mm kit will run you 50% more than that budget. If it interests you.
If you're looking for more information on watercooling and/or what to look for in cooling fans feel free to peruse the guide I wrote >>
here<<
I understand the concept of downclocking. I do not understand the concept behind a 2GB and a 4GB card limiting you to only 2GB when there is a total of 6GB between the cards. Is memory being duplicated for both cards, thats the only scenario I can think of. If so why the duplication?
OK, I can see the confusion. Say card #1 renders odd numbered frames, and card #2 renders even numbered frames. If you have more resources (faster clock, larger frame buffer) available to (say) card #2 then it can render it's portion of the gameplay faster. You'd end up with ( with 3 dashes for the slower card and 2 for the faster card to exaggerate the difference- the dashes being the time to render the frame)
Card#1: 1---3---5---7---9---11---13---15---17---19---21---23
Card#2: --2--4--6--8--10--12--14--16--18--20--22--24--26
In this example card#1 would be behind in the render and out of sync by frame 7, so you would end up with lag (stuttering) or card#2 would just send out it's next frame(s) at the expense of card#1's -this happens in the normal course of SLI/Crossfire rendering but isn't the norm as would be the case in the above example.