Core Clock, Memory Clock, and OCing. etc.

[brolli313]

Hey, so what's the difference between the GPU's core clock and memory clock? I'm comparing cards and I see,
Card 1: Core Clock 580Mhz ; Memory Clock: 1400Mhz
Card 2: Core Clock 600Mhz ; Memory Clock: 1300Mhz

Yet card 2 is costs more, and the other specs are the same, i.e. same memory size, shader model etc, pixel pipelines...

Is it sometimes that a card similar to each other differ in price because of the manufacturer like XFX , EVGA , or ASUS may have the same card but yet different prices?

Also, what are pixel pipelines and the memory bit. I hear people talk about pipelines such as when overclocking or "unlocking" them. And memory bits like when I hear the "8600 is only 128-bit"

Also when overclocking a card, when raising the core clock, is there some kind of way to know where you should then set your memory clock? I don't OC atm and I want to start but don't wanna f up my card.

Thanks. Sorry for all the questions, just some stuff I had to get off my mind.

 

[jonnyGURU]

You can't only go by just core clock and memory clock.

Take an Evga 8600 GT, for example. Reference is GDDR3, their uses DDR2 of the same clock speed. It's essentially an 8500 card with an 8600 GPU and performs only slightly better than an 8500 card. A reference 8600 is significantly better than an 8500, but Evga's costs A LOT less thana reference 8600 (although more than an 8500, of course.)

And then you have over clocked cards that may use better RAM (like Samsung vs. Nanya) than it's "cheaper" counter parts. A big misconception is "I'll just buy the non-OC card and OC it myself" but that's not always possible.

If you like bleeding edge, and buy a card when it first comes out, like when the 8800 GT's come out, you might as well buy based on price because all of the cards are pretty much the same, but if you wait, you need to be careful because all of the graphics card companies are trying to find ways to either make the card cheaper (cheaper RAM, thinner PCB like 6 layer vs. 8 layer, cheaper coolers, etc.) or make the card faster (better faster RAM, additional VR phases) or cool better (better designed/implemented coolers), but this may not be too obvious to the end user.