These are with an AXP @ 11x200 and 768 megs of RAM @ 200 / 2.5-3-3-8
My initial benchmarking here
That set of benchmarking was with the 5900XT and mostly to help specifically BFV people with how different settings affect performance. It also explains some of the details of how I benchmakred, the most important detail is that I forced 80FPS as the MAX FPS possible by truncating any periods above 80FPS at 80FPS, this is because I was looking for the BAD areas, and you can see by the benchmarks there are certain settings that can give VERY large swings in performance. It also explains the path I took with FRAPS for benchmarking.
Recently I got a Radeon 9700 from eBay for $103, the previous owner appeared to have already changed the clock speeds to 325/310, so I got a 9700Pro for $103. Given how close the 5900XT and 9700 Pro are in most benchmarks, I thought this data interesting and worth sharing in the AT forums.
These forums do not appear to allow any kind of preformatted text or
My initial benchmarking here
That set of benchmarking was with the 5900XT and mostly to help specifically BFV people with how different settings affect performance. It also explains some of the details of how I benchmakred, the most important detail is that I forced 80FPS as the MAX FPS possible by truncating any periods above 80FPS at 80FPS, this is because I was looking for the BAD areas, and you can see by the benchmarks there are certain settings that can give VERY large swings in performance. It also explains the path I took with FRAPS for benchmarking.
Recently I got a Radeon 9700 from eBay for $103, the previous owner appeared to have already changed the clock speeds to 325/310, so I got a 9700Pro for $103. Given how close the 5900XT and 9700 Pro are in most benchmarks, I thought this data interesting and worth sharing in the AT forums.
These forums do not appear to allow any kind of preformatted text or
Code:
option, so you'll have to bear with the CRAPPY looking tables. If someone knows an option to allow multiple spaces between characters like every other set of forums on the planet has, please let me know... I couldn't find it.
First I tried the settings I normally use for the 5900XT, and tried turning up the 'graphics' setting:
[b]5900XT vs. 9700 Pro -- In Game Settings (1024x768)[/b]
Card. . Graph . .Text . .Shadows . .AF. . .AVG FPS
. . . . . . . . Geom . .LMaps. . . AA. . . . . . . . . . MIN FPS
5900xt . .M . . .H. .H .On. .Off .Off . Off. . . 64.5 . .38
5900xt . .H . . .H. .H .On. . On .Off .Off. . . 58.4 . .29
5900xt . .HS . .H. .H .On. . On .Off .Off. . . 43.6 . .10
9700p. . .M . . .H. .H .On. .Off .Off .Off. . . .78.7 . .57
9700p. . .H . . .H. .H .On. .Off .Off .Off. . . .75.3 . .47
9700p. . .HS . .H. .H .On. .Off .Off .Off. . . .62.9 . .12
I saw the data from the first run and was like OMGBBQ!!1!1!
Even though the 9700Pro is clearly superior here, it still has HUGE problems with 'graphics' on highest when zooming in. You can see that the average is well into the 'acceptable' range, but when zooming the min dropped down into the teens. Something is clearly screwed up with BFV on highest. No way should you normally be at 60+ FPS and then zoom in on someone only to have your rates drop into the teens. Just ain't right.
So next I wanted to check what more I get out of the card and be happy with the settings. I had already looked at the route of increasing 'Graphics' quality, next I looked at resolution.
[b]5900XT vs. 9700 Pro -- Resolution (settings m-h-h)[/b]
Card . . . Res . . . . AVG FPS . MIN FPS
5900xt .1024x768 . . 64.5 . . . . 38
5900xt .1152x864 . . 55.3 . . . . 31
5900xt .1280x960 . . 47.4 . . . . 26
9700p . 1024x768 . . 78.7 . . . . 57
9700p . 1152x864 . . 75.0 . . . . 48
9700p . 1280x960 . . 70.1 . . . . 42
So at 1280x960, the 9700 Pro is still faster than the 5900XT at 1024x768. WAY bigger difference than I'd expect
[b]5900XT vs. 9700 Pro -- AA and AF[/b]
Card . . . Res . . . . AA . . AF . . AVG FPS . MIN FPS
5900xt. 1024x768 .Off . .Off . . . 64.5 . . . . . 38
5900xt. 1024x768 .Off . . 8x . . . 44.4 . . . . . 23
5900xt. 1024x768 . 4x . .Off . . . 55.5 . . . . . 35
5900xt. 1024x768 . 4x . . 8x . . . 40.8 . . . . . 24
9700p . 1024x768 .Off . .Off . . . 78.7 . . . . . 57
9700p . 1024x768 . 4x . .Off . . . 67.6 . . . . . 44
9700p . 1024x768 . 4x . . 8x . . . 61.0 . . . . . 39
9700p . 1280x960 . 2x . . 4x . . . 57.3 . . . . . 34
So here, the 9700Pro at 1024x768 4xAA/8xAF is about equivalent to the 5900XT at that res with no AA or AF.
A couple notes on the AA/AF runs:
- Text with AA on looks WAY WAY better with the ATi cards. It's actually readable. The text quality is not even close between the two, the 9700Pro is usable with AA on, and I would not conisder the 5900XT to be usable with AA on.
- At 1280x960 2xAA/4xAF, though the frame rates looked good, the game got REALLY choppy around the temple. I'm not sure what was going on. It was way worse than the FRAPS data suggests. Unplayable bad. It was like go for a fraction of a second, stop for a fraction of a second, go , stop, go, stop. Horrible. It was a lot like what happened when I tried my GeForce2 just to check it out. FRAPS framerates looked fine, but the actual game was unplayable.
Overall I was totally shocked at the performance of my new Radeon 9700 (Pro!) It was planned for the backup machine, but it looks like it'll get called up for duty in the main machine. I figured the ATi cards would run a little better in BFV, but I was also shocked at how MUCH better they were. I had seen the HardOCP reviews of the x800s and it hammered the 6800 in comparison [b]for BFV[/b] so I guess that should have been some indication, but somehow I didn't think it would be quite that large a difference.
Driver versions were 56.64 for nVidia and Catalyst 4.6 for ATi. Not sure on the details of how much those come into play as far as performance.
Facebook
Twitter