Indeed. For AMDs new approach of using multiple inexpensive dies to work, they will have to of remedied many of the short comings with current multi gpu configs, such as micro stuttering and inconsistent scaling/compatibility problems. They claim that after Nvidia sees their new approach on the 4870x2, that they will be making their concurrent offerings in a similar fashion. I do agree that this method is best for us as consumers ( as Nvidia's method of disabling hardware capabilities is somewhat lame ) and it should result in more affordable, yet still high performance, gpus. This would suggest to me that they have indeed designed a vastly improved bridge and hopefully this is the case. I would love to see unanimous scaling in games with these cards.
I'm still holding my breathe on this though but its what we should want to happen for the sake of progress and competition. For this all to go mainstream they will have to do more than plain old crossfire as its the only way this will be universally adopted and their high level of confidence sure seems to suggest this. Hopefully its not just PR hype. The current inconsistency of multi gpu sucks to put it bluntly so anything that improves this is welcomed. Sure multigpu usually will offer improved performance but its often quite small given the substantial increase in cost. What we need is a highly efficient method of scaling multiple cores and SLI/Crossfire isn't it. I won't touch a 4870x2 if its merely plain jane crossfire. For consistency's sake I'd more likely get a single GTX280 and be done with it. Hopefully AMD will prove us all wrong.
Reply With Quote
Bookmarks