You will have 512MB for CF.
The clocks are shared, and it goes to the lowest card.
The clocks are shared, and it goes to the lowest card.
Memory and crossfire.
1 - 4 of 4 Posts
Quote:
Well, I look at it like mixing RAM modules or hdd's in RAID. the speed and capacity is always defaulted to the lower number. I was unsure how this worked with GFX cards closkc and RAM amounts. I know SLI and xfire do their jobs differently, so I wasn't sure if a card that OCed higher than another was matched up with said card in xfire, what that would mean.
Originally Posted by Drakan290 This is a strange question to ask, as it's not conventional logic in how the cards work. Basically they both will work at the same speeds they do now... HOWEVER. The faster card will finish it's rendering quicker than the slower card, which forces the faster card to 'wait' to display what it has rendered. Following me here? The slower card is the benchmark for the faster card, so there is no real reason to overclock if you can't get both of the cards at the same speed, at least, in a Crossfire solution. Also, you will get a total of 512mb addressable space for Crossfire. Think of the two cards as one, instead of on a PCB they are together via a soft 'bridge'. |
1 - 4 of 4 Posts
