Soldato
- Joined
- 6 Feb 2010
- Posts
- 14,582
So you are saying AnandTech's data are wrong and Tech's are right then?without aa, at 2560x1600(which is MUCH closer to 1920x1200 4xaa performance) its 25% behind IIRC.
just because this happens to be the first in my firefox history from techpowerup.
http://www.techpowerup.com/reviews/ASUS/GeForce_GTS_450_TOP_DirectCU/10.html
84 to 94fps, not a huge difference, no AA
Crysis, low res, no aa 55fps to 65fps, a 20% lead for the 5850,
slightly higher res, 2xaa 34 to 44fps, 30% lead to the 5850, thats 1280x1024 with 2xaa, hardly mind blowing stuff.
1680x1050, 26 to 34fps, coming up to almost a 40% advantage to the 5850,
1920x1200 4xaa, 21 to 28fs, again almost a 40% advantage, then on to top res and the 5850 finally gets further than 40% ahead.
The most demanding game in the test probably, the 5850 even at low res no aa offers a pretty massive advantage, move up beyond the res we all used on monitors a decade ago(for me, almost two decades) and the advantage is between 30-40%, its not close, and the 460gtx 1gb is no where near 30% cheaper, which makes the 5850 a MUCH better value card and its MASSIVELY more powerful.
Dawn of War 2(a game I expected to be more cpu limited, but then its a BS RTS compared to the first game) and 1920x1200 4xaa has a 30% advantage.
Dirt 2, 1920x1200, 4xaa whats this, another 30% advantage, 61 vs 78fps.
Far Cry 2, actually beats the 5850, but basically matches it throughout, no realistic difference between them. Being honest, I swear when I played it(sure witha 4870x2 back then) I was blistering along in 1920x1080 with 4xaa at much higher FPS than that review shows, has a patch changed performance massively or something? It seemed ridiculously fast and smooth given it looks pretty damn decent, or that was my impression of it anyway.
Hawx, first game the 460gtx 1gb has a lead in, pretty decent at low res, at 1920x1200 they are within about 5%, at higher res the 5850 leads marginally.
Metro are pointless results, they used settings that would give unforgiveable low minimums at 1024x768, weird results throughout, the 460gtx starts lagging first at 1920x1200(you'd think due to the worse memory performance on Nvidia vs AMD, IE memory limits for 5 years have ALWAYS hit Nvidia first, and continue to do so, yet at higher res the memory limit has hit all amd cards but not the 1gb 460gtx, DOF/Tess kill performance more on AMD, no idea which card is faster without both enabled.
The problem with tabulated results is they include, WOW, Supreme commander, very very old 3dmark, they are all so cpu limited that tha 480gtx is performing the same as a 5770, if you go back far enough and include ENOUGH games, then the results will even out far more than if you compare non cpu limited games.
In the vast majority of games the 5850 is between 20-30% ahead, in a LOT its 40% ahead from low to high res, aa or without. In several games they perform the same as 260gtx's and 470gtx's, because the games are massively cpu limited, these effect the results massively but don't remotely indicate the cards are the same speed. Throw out the cpu limited results and that 10% across all games, will get much MUCH closer to 30%.
The cards aren't in the same ballpark and with the 460gtx not 20%, let alone 30% or 40% cheaper, its a much worse value card.
This will only get worse as AMD bring the same performance level down to the £170 price point.
http://www.anandtech.com/bench/Product/180?vs=164


If graphic card can poo, the fanboys would probably argue the poo from their graphic card smille nicer than the poo of other side 