Sli vs Xfire

[Deleted member 24505]

what would xfire be like on my p5b-dlx board with the second x1950pro in the x4 pci-e slot? has anyone tried it?

 

[yogurt_21]

I wasn't talking about comparing scores, I wasn't comparing scores with him. I was comparing how much of a benefit SLI was giving vs. Crossfire vs. single cards.

exactly the benefit varies depending on how large of a cpu bottleneck you have once the second card is added. so a stock opteron bottlenecks heavily against 2 x1950's and an e6600 at 3.6GHZ with 2x 7600gts' doesn't bottleneck at all, and with 2x 7900gts's not that much.

 

[newtekie1]

exactly the benefit varies depending on how large of a cpu bottleneck you have once the second card is added. so a stock opteron bottlenecks heavily against 2 x1950's and an e6600 at 3.6GHZ with 2x 7600gts' doesn't bottleneck at all, and with 2x 7900gts's not that much.

Well first of all, I don't see where you are getting that his Opteron is stock, he said his graphics cards are stock, but didn't say anything about his processor.

Secondly, while you have some valid points, if he has a CPU bottleneck with his x19x0 crossfire setup, that just goes along with my first point that something isn't right with his numbers.

Also my 7600GTs scores were not in my rig with my e6600, check the specs in my sig, and my 7900GTs only lose about 500 points when I put them in with my Athlon X2, that was my original rig before I got the other ones. Without his exact specs we can't really say. However, having a CPU bottleneck would again just make his original statement incorrect, which still is my point.

 

[ElXtronic]

I have 2x 7600 GT's which i ran previously @ stock speeds, these cards fail to be stable and show no performance gain 60% of the time, when i do CS-S benchmarking i find that it maxes another 20 fps in performance where-as i start off with 105 with 1 card. Everything was set up properly and tested serveral different times with driver set after different driver set(all manufacturer) and still unstable with 20ish fps increasement at given times on all drivers. This continues to lose my trust in SLi and is very disapointing since my 7300 GT's worked perfectly with no Bridge(as u should not need to be using on older SLi versions). I went from 60-65fps to 90-95fps with my 7300GTs. I see in idea why SLi is so hard to give accurate views on, some find it to work where others do not, i even looked into motherboards which could cause issues, P5ND2-SLI and P5NSLI being a high end ASUS boards shouldnt have issues of this type with base cards such as 7600 GTs, couldn't be that due to testing on both and showing same results. Tested both individually and both worked 100% in single mode, called BFG,eVga, and PNY on these issues but all say SLi is there for a performance increase and thats urs, all i can say is GG SLi and ur unstable architecture.

Haven't supported SLi since, Xfire ftw. Though i still prefer nVidia cards in single mode

 

[newtekie1]

Odd, I have 3 SLI rigs, and haven't experienced any of the "instability" you claim to get with SLI on any of them with my 7900GTs or my 7600GTs and see performance gains for at least 40% over a single card setup.

Of course there are some games that don't really benefit from it, and you don't really benefit from it at lower resolutions(1680x1050 is the lowest I would go when using either SLI or Crossfire). But this is pretty much the same case with Crossfire. Though in the games that don't benefit from SLI you can always set it up so the second card is used to do the AA in the game, which at least takes the AA load off the first card.

My guess would simply be that your resolution wasn't high enough to see a benefit from the extra 7600GT, though it was probably high enough to see a benefit from an extra 7300GT since the 7300GT is shit.