What do the amount of CUDA cores, shader clock speed, etc mean performance wise

[GuitarJozz]

I am looking into upgrading my video card, and was wondering what exactly the CUDA cores do, and basically what I should be looking for in the specs of each video card I'm considering. Would a card with lower core clock, but high CUDA core amount be better than a card with lower CUDA cores, but higher core clock? Also, would running two cheaper($100) cards in SLI mode be comparable or faster than one card in the $250-300 price range?

I'm just asking for a brief explanation of how exactly things like the CUDA cores, Shader speed, core clock speed, boost clock, and memory interface affect the performance of the video card, so I can better compare the videos cards in my price range myself.

Thank You in advance. I kinda just want to know what to look for and what the specs actually mean so I can understand what I'm buying, rather than just asking for what card is best, you know?

 

[cl-scott]

This example is a little on the oversimplified side, but it should give you enough to get started. After which, I'd suggest some of the fine articles on sites like Tom's to get more of the specifics.

If you think of your video card like a freeway, then CUDA cores would be analogous to the number of lanes in the road, clock speed would be the speed limit, and memory interface would be the number of lanes for exit/entry ramps. More lanes means more cars can be moving on the freeway at any given time, the higher the speed limit the faster any given car is moving, and then if there's say 2 exit lanes instead of 1, you can have more cars getting off the freeway, same as you can have more cars coming onto the freeway if there are say 2 entry lanes instead of 1.

Again, it's a bit of an oversimplified example, but it should be enough to get you started in understanding some of the articles explaining these things in greater detail.

 

[k1114]

In the end, the specs are irrelevant. What is important is game performance so you ignore the spec sheet and look at game benchmarks when choosing a card.

Tom's publishes a "best graphics card for the money" article each month and also includes a hierarchy chart at the end. http://www.tomshardware.com/reviews/gaming-graphics-card-review,3107-7.html

 

[GuitarJozz]

Best answer selected by GuitarJozz.

 

[Dylan_30]

This example is a little on the oversimplified side, but it should give you enough to get started. After which, I'd suggest some of the fine articles on sites like Tom's to get more of the specifics.

If you think of your video card like a freeway, then CUDA cores would be analogous to the number of lanes in the road, clock speed would be the speed limit, and memory interface would be the number of lanes for exit/entry ramps. More lanes means more cars can be moving on the freeway at any given time, the higher the speed limit the faster any given car is moving, and then if there's say 2 exit lanes instead of 1, you can have more cars getting off the freeway, same as you can have more cars coming onto the freeway if there are say 2 entry lanes instead of 1.

Again, it's a bit of an oversimplified example, but it should be enough to get you started in understanding some of the articles explaining these things in greater detail.

That is the best analogy I have ever seen.

 

[boxcrate32]

This is by far the best analogy that I have ever heard on this topic.

Thank you very much for providing this analogy, it helps me a lot.