PDA

View Full Version : R600vsG80 Benchmark Comparisons From Any Game Welcome!


Pages : 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 48 49 50 [51] 52 53 54 55 56

sammy sung
08-26-07, 04:19 AM
Hehehe and the tug of war continues :p

fivefeet8
08-26-07, 05:44 AM
And here's another knee-jerk reaction. That is not what I was claiming[that lack of ROPs makes sense from a design perspective],


I'm probably miss-understanding what you're saying. And of course from a hardware performance perspective, shader based AA resolve will lower performance for the G8x. I don't think I've ever claimed otherwise.


My comment had to do with our line of discussion earlier about that. Thus, if developers decide to go the route of shader-resolve for HDR compatability reasons, G80s ROPs become effectively useless and it's brought down to the AA performance level of R600.


I agree, but I don't see many developers going to purely shader based AA atleast until dedicated hardware resolve is done away with in future Shader Model specs.


Sheesh you are one defensive cat. It is truly difficult to carry on a civil discussion with you.

I've never once said anything about you personally Xion. I've probably just miss-understood some things that you said. No need to get defensive. :p Lets continue with the screens. If I have some time this weekend, I'll post some screens from a few games using 8xS and 16xS Mixed SS/MS modes. ;)

Razor1
08-26-07, 10:56 AM
You're going off on one of your usual tangents here. Please try and stay on topic or at least half-way follow what I'm saying.

What I am saying is that none of these benchmarks have been comparing two cards handling AA resolve in the shaders. G80 has not had to do it in any game but CoJ in DirectX10, so it has had the advantage of dedicated hardware this entire time while R600 hasn't.

Looking back at it now for the r600 was it a good or bad decision by ATi going with only shader based AA?


You're acting like this "similar situation" is something new for R600 when it's not. It's been handling AA in the shaders the entire time. We know exactly how it performs when doing so, and if CoJ is any indication, we have a small glimpse of projected performance difference that we can expect if G80 is forced to do it, too.

Again same thing as above, shader based AA won't be predominent till much later after these cards are pretty much old news. Coming to early with only shader based AA is probably one of the huge down falls of the r600.

We certainly know, at least, that the significant AA advantage that G80 holds would be seriously diminished.

Its still compairable to the r600 even with games that use it, so all that extra ALU throughput for the r600 still doesn't help it.

And here's another knee-jerk reaction. That is not what I was claiming[that lack of ROPs makes sense from a design perspective], and you should know this by now because we just discussed it a few posts ago. You and I had already agreed upon dedicated hardware being more effective at AA than generalizing it in the shaders just a few posts back.

This is actaully not correct, if more transistors are used general hardware can work as fast as dedicated hardware. Its all about cost vs. neccessity

My comment had to do with our line of discussion earlier about that. Thus, if developers decide to go the route of shader-resolve for HDR compatability reasons, G80s ROPs become effectively useless and it's brought down to the AA performance level of R600.

The g80 still has less transistors then the r600, and still has higher performance. (well if nV and ATi count their transistors similiarly)

fivefeet8
08-26-07, 04:21 PM
Razor, I believe Xion was simply stating the obvious. ;)

Razor1
08-27-07, 05:53 PM
ah ok guess I was reading into it too much :o

Shocky
08-27-07, 08:38 PM
If it no longer has that outlier in specific titles in the future, then the performance gap between the cards, at least respective to what we've seen with CoJ, is destined to close quite a bit. And we've already established that it makes sense from a hardware perspective.

Rubbish, from the COJ benchmarks the performance of g80 in this title is horrible even without AA, i'm sure the performance gap will close slightly when using shader AA but nowhere near as much as your suggesting.

Xion X2
08-27-07, 10:29 PM
ah ok guess I was reading into it too much :o
I think you guys do a lot of times, which was my point. :)

I wasn't being "defensive," fivefeet8. It gets really tiresome being repeatedly misinterpreted and argued with when I'm already busy walking on eggshells around this place.

Anyway, no harm - no foul.

And for the record, I think ATI not going with dedicated hardware for AA was a mistake. Unless games soon start to go that route, that is, and the only game we've seen use it [CoJ] performs like crap on both cards, so it's not really a great sign of things to come so far.

ATi very well may have jumped the gun with it. I admit that I generally prefer their forward-thinking approach to Nvidia's which is focused more on the present-generation (at least in recent memory), but they may have jumped ahead into the future too far this time.

Anyway, Crysis is a little less than 3 months away, so we'll soon find out for certain whether or not they blew it. And I think if R600 crawls on that game, which is the most anticipated PC game in history for high-end users, then they truly would have blown it.

Xion X2
08-27-07, 10:32 PM
Rubbish, from the COJ benchmarks the performance of g80 in this title is horrible even without AA, i'm sure the performance gap will close slightly when using shader AA but nowhere near as much as your suggesting.
Ah bull, Shocky. R600 is about even with GTX in a lot of titles that don't use AA because of G80s advantage w/ its ROPs.

In some it's still not close, but in many it is.

Obie
08-28-07, 02:57 AM
Review COJ (http://enthusiast.hardocp.com/article.html?art=MTM4NCwxLCxoZW50aHVzaWFzdA==)

Fotis
08-28-07, 03:08 AM
There is a nice little R600vsG80 review testing Bioshock over at Gamespot (http://www.gamespot.com/features/6177688/p-6.html).
To sum up the results the HD2900XT beats the 8800GTX in Windows XP but it looses terribly in Vista.

Great Vista driver support AMD! :rant:

fivefeet8
08-28-07, 03:42 AM
Review COJ (http://enthusiast.hardocp.com/article.html?art=MTM4NCwxLCxoZW50aHVzaWFzdA==)

Seems drivers may have changed the performance between the various cards in that particular dx10 bench. The 8800gtx/ultra are a good 30% ahead of other cards with 4xAA at 1280x1024 in dx10. The 8800gts and x2900xt are pretty much neck and neck with the same settings.

fivefeet8
08-28-07, 03:48 AM
There is a nice little R600vsG80 review testing Bioshock over at Gamespot (http://www.gamespot.com/features/6177688/p-6.html).
To sum up the results the HD2900XT beats the 8800GTX in Windows XP but it looses terribly in Vista.

Great Vista driver support AMD! :rant:

Something's not right with their FX60 CPU test. How can a Athlon FX 60@2.0 ghertz get the same fps as the same CPU at 2.6? The game seems to love Intel Core2Duo's.

It's a bit confusing looking at the CPU test. They suggest the game is purely GPU limited and very CPU limited at the same time.

Shocky
08-28-07, 08:00 AM
Ah bull, Shocky. R600 is about even with GTX in a lot of titles that don't use AA because of G80s advantage w/ its ROPs.

In some it's still not close, but in many it is.

Yes, everybody knows 2900XT is good without AA, but the fact is the poor performance in COJ with g80 has little to do with AA so it's a very poor example.

Review COJ (http://enthusiast.hardocp.com/article.html?art=MTM4NCwxLCxoZW50aHVzaWFzdA==)

How about that, 8800GTS still faster then a 2900XT with AA in COJ when using shader based AA in DX10 and GTX/Ultra again ahead, I guess nothings really going to change at all is it.

So now all four DX10 games are faster on a 8800GTS, oh dear.

Xion X2
08-29-07, 09:04 PM
ATI's DirectX10 drivers are crap at the moment. No argument here. I'm hoping they're able to work that out soon.

Considering most DX10 games perform poorly on all hardware, though, it's not a huge drawback for me.

Redeemed
08-30-07, 06:17 AM
ATI's DirectX10 drivers are crap at the moment. No argument here. I'm hoping they're able to work that out soon.

Considering most DX10 games perform poorly on all hardware, though, it's not a huge drawback for me.

Bioshock plays wonderfully on my rig with literally everything cranked- in DX10. ;)

Xion X2
08-30-07, 11:35 AM
Bioshock - DirectX10

1680x1050, maximum settings, 8xAF (no AA)

2900XT @ 850/950

2007-08-30 04:03:31 - Bioshock

Frames: 5364 - Time: 92405ms - Avg: 58.048 - Min: 34 - Max: 99


http://img261.imageshack.us/img261/9963/bioshock200708300255212xw0.jpg
http://img254.imageshack.us/img254/3410/bioshock200708300256067kz0.jpg
http://img254.imageshack.us/img254/440/bioshock200708300257471pk8.jpg
http://img254.imageshack.us/img254/1669/bioshock200708300307430zf7.jpg
http://img261.imageshack.us/img261/2482/bioshock200708300315442zu3.jpg
http://img254.imageshack.us/img254/5576/bioshock200708300322303om9.jpg
http://img261.imageshack.us/img261/3384/bioshock200708300327110zf1.jpg
http://img261.imageshack.us/img261/5873/bioshock200708300330216qx6.jpg
http://img261.imageshack.us/img261/3286/bioshock200708300334515ry4.jpg

The big daddy causes the biggest hit I've seen yet. Down to 34fps.

Redeemed
08-30-07, 02:08 PM
But still very playable- and it's DX10. ;)

Obie
08-30-07, 05:43 PM
Looks like Bit-tech (http://www.bit-tech.net/gaming/2007/08/30/bioshock_gameplay_graphics_and_performance/8) and Gamespot (http://www.gamespot.com/features/6177688/p-6.html) show a more accurate representation of how the HD 2900XT performs in Bioshock then what Firingsquad showed us. I took a few Fraps FPS at 60 seconds each and never fell that low in frame rates running through a cleared level like the one firingsquad reported. Even when the level is cleared you get a few splicers here and there but the frame rates never looked like that on FS. However, I believe that ATI still needs to work on their dx10 drivers.


Since the game doesn’t include a built-in utility for benchmarking, we’re testing BioShock performance with FRAPS, as we do with many other games we test with such as STALKER, Battlefield, and Oblivion. In this case, we manually run through the medical pavilion level of the game after it’s been clear of all the baddies. Our test sequence starts towards the beginning of the level, this area is where the frames per second is at its lowest. This is likely because this area uses shadows extensively. From here we run into the medical pavilion foyer, up the left stairs to surgery, and then hook another left to go to the crematorium entrance. From there we got up another set of stairs to the eternal flame, and that’s where we conclude our manual walkthrough.

source (http://firingsquad.com/hardware/bioshock_directx10_performance/page2.asp)

Xion X2
08-30-07, 11:18 PM
But still very playable- and it's DX10. ;)
Yea, it's not too bad, and I enjoy the extra effects of DX10 for sure. The new RivaTuner really helps by allowing triple-buffering in DX10.

Redeemed
08-31-07, 12:50 AM
Yea, it's not too bad, and I enjoy the extra effects of DX10 for sure. The new RivaTuner really helps by allowing triple-buffering in DX10.

Now to see if Crysis performs as well. :p

nightmare beta
08-31-07, 06:07 PM
I'm pretty sure Crysis will perform as well.

Redeemed
08-31-07, 06:35 PM
I'm pretty sure Crysis will perform as well.

Actually, there isn't anything in BioShock that I have to tun down. I literally have everything maxed (save for AA) and it played perfectly smooth. So smooth, that I'm confident even with AA it'd still be very playable. I do doubt I'll be able to say the exact same with Crysis.

Xion X2
08-31-07, 07:06 PM
Crysis is another animal altogether; it looks much better than Bioshock could ever dream of. It will be much, much harder on graphics cards due mostly to the high-res textures. Many of Bioshock's textures are low-res because it was ported from the 360.

I don't think there's a single card setup out there that'll max out Crysis at a smooth framerate. If it's optimized well for SLI or Crossfire, then those systems may come close.

I have everything turned up on Bioshock in those screens, btw. AA is the only thing not enabled, but it isn't working on any configs in DX10 right now.

tEhMazSaMm
08-31-07, 10:30 PM
Crysis is another animal altogether; it looks much better than Bioshock could ever dream of. It will be much, much harder on graphics cards due mostly to the high-res textures. Many of Bioshock's textures are low-res because it was ported from the 360.

I don't think there's a single card setup out there that'll max out Crysis at a smooth framerate. If it's optimized well for SLI or Crossfire, then those systems may come close.

I have everything turned up on Bioshock in those screens, btw. AA is the only thing not enabled, but it isn't working on any configs in DX10 right now.


Actully teh 8800gts will max out crisis as the devs had already stated. Teh Rsixhundred on the over hand will not be doing so well seeing as it can alrady bairly play bioshok. magor hardware flaws plague the r6hundered and when aa is on yoiu can forget bout it.


Plus dont forget a bout that huge electricity bill! lolololol

Rummy
08-31-07, 11:22 PM
Actully teh 8800gts will max out crisis as the devs had already stated. Teh Rsixhundred on the over hand will not be doing so well seeing as it can alrady bairly play bioshok. magor hardware flaws plague the r6hundered and when aa is on yoiu can forget bout it.


Plus dont forget a bout that huge electricity bill! lolololol


:headexplode: