- May 7, 2005
- 5,161
- 32
- 86
Xbitlabs Crossfire review.
- Thread starter Cookie Monster
- Start date
thumbs up for XBit, even though they make mistakes some times (and who wouldnt when they do mammoth tests over 30+ games), they are quite thorough....a good source of research for me
Awesome review. Wow, they do a lot of testing!
Crossfire had 0 issues during their entire testing and performance is there. Good for ati. Hopefully this will prompt nivida to get more efficiency in their AA performance.
Crossfire had 0 issues during their entire testing and performance is there. Good for ati. Hopefully this will prompt nivida to get more efficiency in their AA performance.
that is because 14xAA wasn't true AA sampling. Nvidia's 16xAA is true AA sampling and hence its not a fair test.
Though thumbs up for ATI's algorithm in 14xAA. pretty effecient while providing good result (but not nearly as good as Nvidia's true 16xAA) but better than 8xAA.
Wow, this review shows Crossfire in a very good light. Thats bound to make some NV fans not happy. They'll have to get new arguments to belittle ATi with now.
Quality looks the same, or even in favor of ATi. So thats a moot point. Its the highest ATi and NV has to offer. Even ATi's 14x runs faster than NV's 8x, most of the time.
Quality looks the same, or even in favor of ATi. So thats a moot point. Its the highest ATi and NV has to offer. Even ATi's 14x runs faster than NV's 8x, most of the time.
Anyone notice the "white out" of the water in the HL2 comparisons? ATI's looks all whited out, but nVidia's looks like actual water. Am I seeing things? Someone check behind me on this.
Also was suprised by how the single X1800XT compared to both the GTX and GTX512 in Fear, COD2, and BF2. In the 2 higher resolutions it handily beat the GTX and came within a frame or 2 of the GTX512!! Pretty impressive.(Also 3D Mark 05 score).
Also same for Half Life 2 and Far Cry(beat GTX512 at 1600X1200 in Far Cry).("ATI" games)
Also same for Half Life 2 and Far Cry(beat GTX512 at 1600X1200 in Far Cry).("ATI" games)
Yreka
Diamond Member
- Jul 6, 2005
- 4,084
- 0
- 76
I am dissapointed, but not surprised there are no Everquest 2 Benchmarks. Would it be safe to say that there is likely little to no benifit with xfire since that is the case with SLI ?
Honestly, I would be all over xfire if :
1. It worked with SLI X16 motherboards.. Will never happen, I know but I am in love with my A8N32.
2. Get rid of the dongle and mastercard requirements. its been said 1000 times, and I could definatly get over this if #1 was met, but since this is just a wish list, there it is.
Either way, I will likely go single card-ATI next round if it stays close. For me high AA,AF and IQ is more important that raw speed.
Honestly, I would be all over xfire if :
1. It worked with SLI X16 motherboards.. Will never happen, I know but I am in love with my A8N32.
2. Get rid of the dongle and mastercard requirements. its been said 1000 times, and I could definatly get over this if #1 was met, but since this is just a wish list, there it is.
Either way, I will likely go single card-ATI next round if it stays close. For me high AA,AF and IQ is more important that raw speed.
Perhaps moot if you're comparing the results. I don't know as I haven't even looked yet.
But there is a big difference in the ways things are done, here. ATi's 14x mode is pure multisampling while nV's 16x is a mix of supersampling and multisampling.
It's not because ATi is somehow magically better.... there's a real reason for the performance delta.
That's a great review-much better than the last one I saw of crossfire.
I'm probably going to go the crossfire route over SLI since I'll not be playing @ less than 8xAA and 16xAF.
How does everyone else feel?
Rogo
I'm probably going to go the crossfire route over SLI since I'll not be playing @ less than 8xAA and 16xAF.
How does everyone else feel?
Rogo
I like the fact that the implementation of Crossfire is hardware based. I like this style as games will just work. No need to create a profile for a paticular game. I like that. Simplicity is what I'm all about. 
Nice performance numbers too although I think I'll wait for X1900XTX.
Nice performance numbers too although I think I'll wait for X1900XTX. I know there is a reason. But since the results are for NV sometimes, and for ATi sometimes, and overall results seem to favor ATi, it doesnt matter how its done. What matters is the end result, and that is that NV's SLI AA is a lot slower than ATi's Super AA.
read. in the article they SPECIFICALLY said that super 14xaa is NOT AS GOOD AS Nvidia's true 16x sampling.
No one ever said it looks graphically better unless there is something wrong with their eyes. its theoratically impossible.
please stop your ATIfantism and stop mixing your opinion with reality.
Ati's superAA is not pure multisampling. It uses different sampling patterns for each card, and then the 2 results are blended in HW by the compositing chip on the master card. That in effect makes it partially supersampled. I'm not sure how exactly it's done in SLI, but I know the reason Ati takes a smaller hit is because the 2 resulting images are blended in HW. This certainly wasnt the case in x800 CF, and even then the IQ was comparable to NV. I know for a fact that SLI AA wasnt available when SLI was first introduced, which leads me further to believe that it was added as as afterthought, without the proper hardware to make it efficient.
perhaps there is some kind of bug in 14axx sampling that caused the water to look that way? You do have to sacrafice something doing the shortcut. Since super 14AXX is not true AA, it is possible.
TRENDING THREADS
-
Discussion Zen 5 Speculation (EPYC Turin and Strix Point/Granite Ridge - Ryzen 9000)- Started by DisEnchantment
- Replies: 24K
-
Discussion Intel Meteor, Arrow, Lunar & Panther Lakes Discussion Threads
- Started by Tigerick
- Replies: 20K
-
Discussion Intel current and future Lakes & Rapids thread- Started by TheF34RChannel
- Replies: 23K
-
-
Facebook
Twitter