The X1950XT used in the review you linked had 256mb. But it didn't use any AA, so it might perform much better in parts of the game where AA bottle-necked the game. This review is not even remotely comparable to 5870 vs. 480 in Heaven benchmark where they made an apple to apple comparison.
NVIDIA GF100 & Friends speculation
- Thread starter Arty
- Start date
The X1950XT used in the review you linked had 256mb. But it didn't use any AA, so it might perform much better in parts of the game where AA bottle-necked the game. This review is not even remotely comparable to 5870 vs. 480 in Heaven benchmark where they made an apple to apple comparison.
yeah but still the price advantage isn't as great as before, if nV prices according to performance, they too can drop prices down too, unless we are talking about no profits..... ATi will also get almost no profits. Going into a price war when they don't really need to is stupid, what happened with the 4870's did they gain marketshare? Was the 20% difference of launch price worth the gain in marketshare? Was reason for the price war justified at the end? Price wars are only effective if you gain enough sales to out weight the loss of initial profits
Hm, what do you mean with "other game stuff"? Something like rendering and physics?
Maybe you should look at nVidia's game Rocket Sled: http://www.youtube.com/watch?v=gP4aWLkyj_M
http://hardocp.com/article/2009/03/02/asus_gtx_260_55nm_4870_dark_knight_1gb/4
http://hardocp.com/article/2009/03/02/asus_gtx_260_55nm_4870_dark_knight_1gb/5
here then to make my point why does in some parts of the game the ati's card does better and in some parts nv's card do better, leave at that. Same damn settings! As I stated you missed the point there are parts of a game or demo that is going to be more favorable to one of the cards but the Heaven demo was built and optimized on ATI cards and where there is no tessellation in parts of the demo the card still looses, in some parts with tessellation it still close, but over all tessellation seems to give the geforce card a huge lead. If someone states those 30 secs are representative of all other games I think that is naive, if you want me to point blank about it.
Last edited by a moderator:
Market share certainly matters if you are interested in gaining developer support for your software, or to counter your opponent's proprietary stuff.
The RV770's certainly gained a lot of points in favor of ATI amongst developers, and RV870 being out early for developers certainly helped them develop DX11 stuff and certainly won't hurt ATI's favor. This is in stark contrast to where ATI was in relation to Nvidia amongst relations pre-RV770
True but still ATi's dev rel program
lets not go there that's a different topic.Never stated ATi's rv870 would hurt them in anyway, but development of Dx11 stuff has been slow, its nice we have 2 games that really use some nice effects but still going to have to wait a year so for more intensive shaders, just like Dx10, it takes time for adoption. Its possible a bit slower because of consoles.
Jawed
Legend
The very start of the demo is looking upwards over the dense grass to the windmill. You can clearly see the FRAPS counter indicating 60+ fps. This corresponds closely with the graph.
Now notice that both cards have the same framerate here. Bearing in mind that HTX480 is supposed to be way faster, why isn't it any faster at this "easy", as he describes it, part of the benchmark?
60fps is not actually "fast" for these cards, it indicates they're doing a hell of a lot of work, per frame.
Jawed
Yes, I have seen your post later. Obviously I needed 3 minutes to write my one
AFAIK, tesselation is never really off during the benchmark runs, it only affects more or less of the scene.
So the part of the video where tesselation is turned on and off at the dragon is especially interesting. With tesselation the GTX480 has 42/43 fps, with tesselation turned off it rises to 56fps. From my feeling, that 56 fps is definitely something a HD5870 is able to reach there at 1920x1080 without AA and AF (at least during the normal benchmark run it exceeds that framerate quite a bit at the dragon, but this is of course not exactly the exact same condition). Could someone with a HD5870 try to re-enact the view from the video (1:48 through 1:58) and check this?
Last edited by a moderator:
DavidGraham
Veteran
IMO , running the benchmark without AA/AF is useless , even if done for showing architectural advantage , what matters is the final image quality with all the options activated , well .. for the average consumer anyway .
The 470 pictures I've seen show 10 GDDR5 chips, do you have contrary info?
We both know that in games, ATI or Nvidia might have a lead on parts of the game. A simple explosion can change the curves even on a game thats benched real time. Reviewers like [H] makes several runs and average them, since there can be variations.
On canned, static benchmarks like heaven, there is less variation (drivers and system components might change the benchmarks, but the benchmark itself is static).
Those 30 seconds where there might be less tessellation, might indicate how the performance difference is in those cases. In my opinion, Heaven demo is probably no indication on how either card performs in real games. But, you cannot say that this is how the 480 performs vs. the 5870 when it comes to tessellation and at the same time dismiss that people say that the 30 seconds of little to no tessellation shows how it performs in that scenario.
Yes, I have seen your post later. Obviously I needed 3 minutes to write my one
AFAIK, tesselation is never really off during the benchmark runs, it only affects more or less of the scene.
So the part of the video where tesselation is turned on and off at the dragon is especially interesting. With tesselation the GTX480 has 42/43 fps, with tesselation turned off it rises to 56fps. From my feeling, that 56 fps is definitely something a HD5870 is able to reach there at 1920x1080 without AA and AF (at least during the normal benchmark run it exceeds that framerate quite a bit at the dragon, but this is of course not exactly the exact same condition). Could someone with a HD5870 try to re-enact the view from the video and check this?
Sorry shouldn't have said turned off, but when the camera gets farther away from the town near the end of the demo there is no tessellation going on, why there does the geforce have a 100% + lead, I think its a good demo to show the potential of the card in Dx11 environments but to try to say those lulls where the cards are close and say it will perform the same as the cards we have today, in Dx10 or less games, is a mistake at this point.
Its a good example of how well tessellation works on these cards, ATi has been using it for how many months now?
Oh yeah those 30 secs you might want to check what is going on before you talk, the video goes along with that graph. There are parts where tessellation is going on.
Jawed
Legend
Version 1.0 of Heaven:
http://unigine.com/download/
NVidia is using the unreleased version 1.1 of this benchmark. So it turns out we cannot make a direct comparison.
Jawed
It doesn't matter what you should have said
, I'm talking about that part of the video where tesselation is turned on and off. It may allow to assess the performance drop for this obscenely tesselated case for both architectures.So anyone out there dares to tell us how a HD5870 fares there (scene from video at 1:48 - 1:58) with tesselation on and off?
Perhaps. Either that or Heaven is the only benchmark where the GTX 480's lead over the 5870 is significant enough to give NVIDIA something to brag about. We'll know soon enough.
Assuming similar yields, which is assuming a whole ******* lot, but we've had this discussion already.
Hmm then why would they state time and time again, its going to be the fastest graphics card
, I would think its more then just Unigine where they are leading.I've seen already the claim 1.1 doesn't do too much for the performance. At least not more than Catalyst 10.3 does
Similar threads
