Who were those guys that said how great R600 was at G80 launch?

[sillyeagle]

What was up with those guys? They were claiming to have all this inside info on how great the R600 is and how it was going to destroy the G80, and it turns out they were making it all up.

Does that kind of stuff happen a lot? I haven't really seen anything like that before, but then again I don't visit many forums. Seemed those guys would have said anything to prove that their lies were in fact truths.

Pretty pathestic really, but I don't understand what their point was. Do they work for AMD or something? lol

[SH64]

Guys like those come & go with each new product/chip release .. i'm getting used to that heh ...

[sammy sung]

Idd i wonder where Shadow001 went,havent seen him post anything for a very long time:rolleyes:

[zer0]

nice avatar :)

maybe his x2900 died due to heat :D

[Trademark]

What was up with those guys? They were claiming to have all this inside info on how great the R600 is and how it was going to destroy the G80, and it turns out they were making it all up.

Does that kind of stuff happen a lot? I haven't really seen anything like that before, but then again I don't visit many forums. Seemed those guys would have said anything to prove that their lies were in fact truths.

Pretty pathestic really, but I don't understand what their point was. Do they work for AMD or something? lol
I call it... wishful thinking...

I knew it was going to be very hard for AMD to counter G80 and that is VERY clear. When G80 was released it was a whole new unexpected architecture, something that went beyond the limits, pretty much delivering twice the performance of nvidias old flagship. AMD probably never expected it and had to counter it quickly with something similar. There's a good reason for the power usage of the 2900XT, if AMD has made an optimised architecture instead of a (semi -) rushed one then they would be too late, because by that time nvidia would have already had something to counter that again, thus they made a less efficient card so t could atleast keep up with a G80, yeh it costs more power, but atleast we can keep up with a G80!, must have been their thoughts.

The idiots who spread those rumors were truly wishful thinkers, how can you counter a brand new architecture just like that? You can't, not AMD, not Intel, hell not even IBM could have done that. nVidia just took em by suprise and they took em good. The 8800GTS is alot more attractive than a HD2900XT, less noise, less power and sometimes beats the 2900xt aswell. They thought it would be another Radeon 9800XT vs Geforce FX5950Ultra situation, where the XT litterelly smashed the Geforce card witouth too much hassle. The Geforce FX series were also a new architecture, the only difference is is that the FX architecture failed horribly against the competition, while the G80 architecture doesn't.

Doesn't take away that these rumors still persist, because most people blame alot on drivers (such as getting beaten by a GTS with about 10 fps or so). These people could of course open their eyes and accept it get's owned untill drivers fix the "issue", if it's an issue of course. But we don't know, because we haven't seen ATI's latest and newer drivers.

Don't listen to rumours, they are always made up by nvidia/ati/amd/intel/whaever fanboys, just wait till the benchmarks roll out after it's released.

[sammy sung]

nice avatar :)

maybe his x2900 died due to heat :D

Thanks if you meant mine:) Thats a long-stemmed wine glass,pretty spectacular picture (and motive i might add:p )

[Xion X2]

The 8800GTS is alot more attractive than a HD2900XT, less noise, less power and sometimes beats the 2900xt aswell.

Says you. A lot of enthusiasts find the 2900XT a more attractive option because it overclocks better, scales better with an overclock (check the Oblivion benches on the other thread), is faster in most games, is built more on a DirectX10 path with its heavier shader architecture, etc. Not to mention the Catalyst Center is a far superior control panel to that buggy pile of crap Nvidia calls a "control panel." There's just no competition there.

And again, if power consumption is so important to you, then why not scale back that monstrous overclock you have on your X6800 a little? You're using up more power by overclocking your CPU like that than you'll ever see between running an XT compared to a GTS or GTX.

That argument is just all-around-stupid, IMO. You criticize the card for its power consumption while saying it doesn't perform on par w/ G80 on two-week-old drivers and not a single DX10 game is out on the market to test it with yet. Perhaps that extra power in both wattage and shaders will be needed a lot more when DirectX10 gets going in full force with its more demanding graphical effects.

So again, a lot of this comes down to speculation. Once a # of DirectX10 games are on the market and the 2900XT is on more mature drivers, we'll have a better idea of whether or not a lot of that power is going to waste or being put to good use. But right now you're doing an awful lot of assuming, as am I, if you or I are trying to draw a conclusion at this point.

[lduguay]

Thanks if you meant mine:) Thats a long-stemmed wine glass,pretty specatular picture (and motive i might add:p )
Is that what it is... specatular indeed :p

[SLINROB]

Says you. A lot of enthusiasts find the 2900XT a more attractive option because it overclocks better, scales better with an overclock (check the Oblivion benches on the other thread), is faster in most games, is built more on a DirectX10 path with its heavier shader architecture, etc. Not to mention the Catalyst Center is a far superior control panel to that buggy pile of crap Nvidia calls a "control panel." There's just no competition there.

And again, if power consumption is so important to you, then why not scale back that monstrous overclock you have on your X6800 a little? You're using up more power by overclocking your CPU like that than you'll ever see between running an XT compared to a GTS or GTX.

That argument is just all-around-stupid, IMO. You criticize the card for its power consumption while saying it doesn't perform on par w/ G80 on two-week-old drivers and not a single DX10 game is out on the market to test it with yet. Perhaps that extra power in both wattage and shaders will be needed a lot more when DirectX10 gets going in full force with its more demanding graphical effects.

So again, a lot of this comes down to speculation. Once a # of DirectX10 games are on the market and the 2900XT is on more mature drivers, we'll have a better idea of whether or not a lot of that power is going to waste or being put to good use. But right now you're doing an awful lot of assuming, as am I, if you or I are trying to draw a conclusion at this point.


I wouldnt rip on him to hard, I dont think he was trying to diss on the R600. I myself try to look at things from a non biased point of view and in all reality the 8800GTS wins some benchmarks and the 2900XT wins some bench marks. Each card has its perks but the fact is the GTS is a pretty attractive option at its price point right now and easily holds its on with the 2900XT. Games such as Tombraider Legends, FEAR, etc the GTS wins in hands down especially when AA and AF are applied and then in Games like Oblivion they are pretty equal until you start overclocking the 2900XT, but not everyone likes to over clock so its really more of an enthusiast option.

As far as I know the 2900XT takes the prize over the GTS in R6 and Test drive unlimited.


As far as Nv's control panel being buggy yeah I can agree with that statement on the first couple sets of drivers but I dont recall any issues with it myself now for a long time, I dont even use ntune anymore because the CP works just fine for me.


As for which card will be better in DX10 I guess we will have to wait and see, So far we have 3 patched DX10 games to look at and the only one that the R600 was holding its own in was Call Of Juarez. Then again I would imagine there is a difference in a PATCHED dx10 app to one that is built DX10 from the ground up.


But as anyone would state its really to early to speculate anything just yet. As long as people are happy with what they have who cares what anyone else thinks or says, right?;)

[Xion X2]

I wouldnt rip on him to hard, I dont think he was trying to diss on the R600.

Trust me, he was. This guy has ripped on the card practically since launch on multiple threads around here. He and I have gotten into it quite a few times.

Games such as Tombraider Legends, FEAR, etc the GTS wins in hands down especially when AA and AF are applied and then in Games like Oblivion they are pretty equal until you start overclocking the 2900XT, but not everyone likes to over clock so its really more of an enthusiast option.

Where are you getting this? The GTS beats the XT in Tomb Raider, but not in F.E.A.R. necessarily. The two are very close at stock speeds, and all you have to do is overclock an XT slightly and it outperforms it with an equal overclock on a GTS (and most people on this forum overclock their cards.) And Oblivion runs better on the XT at even stock speeds. There's at least a 4-5fps difference there in heavy foliage w/ transparency AA enabled; just check the other thread where Ben (buffbiff21) and I were benching the two cards back to back. With a decent overclock, they're not even close. The XT is significantly faster.

And, hello? The 2900XT is on two-week old drivers compared to 7-month mature drivers for the 8800GTS? If Nvidia hadn't released their miracle drivers that have increased performance across the board by 30-40% in the last month this would be a non-issue on both counts. I was getting worse performance on a single 8800GTX back around launch on both Oblivion and Rainbow Six, similar performance in F.E.A.R., better performance in Tomb Raider, equal performance in Riddick. And in none of these tests done on my XT lately do I even have my CPU overclocked like I did back then w/ my GTX. I'll be generous and say that's about an even split between those games.

Rob, have you grown so in love w/ your SLI setup that you can't see through your green-tinted glasses now? :p

[SLINROB]

Rob, have you grown so in love w/ your SLI setup that you can't see through your green-tinted glasses now? :p


Dude. Really.:rolleyes: No reason to slam me either. Everyone is due to their own opinions.

[J-Mag]

You criticize the card for its power consumption while saying it doesn't perform on par w/ G80 on two-week-old drivers and not a single DX10 game is out on the market to test it with yet.

What? CoD is DX10 now, so there is one full game and two demos. In all three circumstances the r600 is getting worked.

[Xion X2]

What? CoD is DX10 now, so there is one full game and two demos. In all three circumstances the r600 is getting worked.

One patched DX10 game and two demos. Two week old drivers.

I think that's all that needs to be said.

Now we can sit here and play the semantics game, but you and I should be able to reason the difference between a game that is coded in DX10 from the ground up and one that is patched, or two other games that are patched for DX10 and are in demo form and not even finalized yet. The testbed is pretty weak at this point.

Once Crysis, BioShock, World in Conflict and some of these other games get here, we'll have a better idea of what to expect.

[J-Mag]

One patched DX10 game and two demos. Two week old drivers.

Now we can sit here and play the semantics game, but you and I should be able to reason the difference between a game that is coded in DX10 from the ground up and one that is patched, or two other games that are patched for DX10 and are in demo form and not even finalized yet. The testbed is pretty weak at this point.

Once Crysis, BioShock, World in Conflict and some of these other games get here, we'll have a better idea of what to expect.

If hardware development progresses like I think it does, then claiming this "two week old driver" thing is wrong, because it is an arbitrary start point. ATI has had almost the same amount of time with the r600 that nvidia has had with the g80. It's not like ATI waited until the day of launch to whip some drivers together. They start work on them before initial spins even hit the fabs. Driver performance improvements tend to be application specific or minimal in their gains anyway...

I'll agree with ya that the newer games built with DX10 from the ground up will load the cards differently due to the developers using new shader techniques, so only time will tell.

[Xion X2]

If hardware development progresses like I think it does, then claiming this "two week old driver" thing is wrong, because it is an arbitrary start point. ATI has had almost the same amount of time with the r600 that nvidia has had with the g80.

Are you overlooking the fact that ATI underwent a merger during this period? Don't you think that shook things up a bit?

It's not like ATI waited until the day of launch to whip some drivers together.

Well as impossible as it would seem, that appears to be the case so far. Drivers are unstable, and we are seeing significant performance increases from one set to the next. Oblivion and Rainbow Six saw a 35-40% jump in going from the 8.37s to the 8.37.4.2s.

They start work on them before initial spins even hit the fabs. Driver performance improvements tend to be application specific or minimal in their gains anyway...

I absolutely don't agree with this. At all. There is a night and day difference in performance in G80 now from back when it first launched. Several of these games we have been benching like Rainbow Six, Oblivion, and F.E.A.R. are showing anywhere from 25-40% gains. That's far from "minimal."

Throughout the life cycle of a card you can see very significant performance increases with driver revisions.

I'll agree with ya that the newer games built with DX10 from the ground up will load the cards differently due to the developers using new shader techniques, so only time will tell.

Glad that we agree on that, but I can't help but disagree with you on your other points.

[Blacklash]

Yep in "Oblivion" overclocked XT (875|1950) vs an overclocked GTS (660|2000) in GPU bound situations with max quality settings from the same save point XT is faster-

GTS: 31, 33, 34

XT: 37, 38, 41

Xion just corrected a bench from test one and the page is here. (Test 1 did say 40 for the XT)

http://www.nvnews.net/vbulletin/showthread.php?t=92461&page=22

[J-Mag]

Are you overlooking the fact that ATI underwent a merger during this period? Don't you think that shook things up a bit?


Sure, but the turmoil for the managers, directors, AP/AR and sales people was far greater than that of the engineers. I don't have any inside info, but I guarantee you they didn't pull people off GPU projects to work on CPUs because most of these EE, CE, and CS people are specialized.


I absolutely don't agree with this. At all. There is a night and day difference in performance in G80 now from back when it first launched. Several of these games we have been benching like Rainbow Six, Oblivion, and F.E.A.R. are showing anywhere from 25-40% gains. That's far from "minimal."


Maybe I was being too hasty, do you have any links showing these results?

I sold my 8800 a few months ago, but when I had it I never noticed any major performance differences among the various drivers.

[SLINROB]

I sold my 8800 a few months ago, but when I had it I never noticed any major performance differences among the various drivers.


I have become very interested in this arguement and in a day or two I am going to see if I can find the original G80 drivers and do a bench comparison with the latest drivers. This should put this arguement to rest.

While I agree that the drivers have gotten a bit better FPS wise across the board I do not believe the diff is as big as someone else is making them out to be. Stalker on the other was the one and only game that I saw get a huge performance boost under.

I am at work right now and will be pulling 12hr shifts for the next 2 days so just hang tight.;)

[Xion X2]

Sure, but the turmoil for the managers, directors, AP/AR and sales people was far greater than that of the engineers. I don't have any inside info, but I guarantee you they didn't pull people off GPU projects to work on CPUs because most of these EE, CE, and CS people are specialized.

I don't find that line of reasoning too hard to agree with, so maybe we can meet in the middle here and say it probably had a little effect on the timeline/driver situation with R600.

And honestly, I'm not saying it just because I own the card or anything, but these early drivers really show signs of immaturity. The new 8.38s are unstable as hell and are giving me BSODs left and right just as Nvidia's release drivers for the G80 did. They're also fluctuating in performance in a big way with certain games (namely, Oblivion.) Drivers that have been in the works for 6-7 months normally don't do that.

Maybe I was being too hasty, do you have any links showing these results?

Here are some benches w/ F.E.A.R. from back in March on Guru3D:
http://www.guru3d.com/article/Videocards/391/25/

Here's one from HotHardware at around the same time:
http://www.hothardware.com/Error.htm?aspxerrorpath=/viewarticle.aspx

As you can see from those links, they're pretty consistent: one says 79, the other 80fps @ 1600x1200 w/4xAA, 16xAF. The average from users around here who have benched the game recently at the same settings is anywhere from 90-95fps, depending on the processor. And this is even w/ transparency AA enabled (disabled in the guru3d bench):
http://www.nvnews.net/vbulletin/showpost.php?p=1277332&postcount=228

Here is a bench that I did back around the same time on Oblivion. Notice how it's dropping to 27fps in this locale w/ 4xAA @ 1680x1050 res:
http://img161.imageshack.us/img161/4035/oblivion200702120858251kg1.jpg
http://www.nvnews.net/vbulletin/showpost.php?p=1204469&postcount=31

Now guys are reporting the game runs at 45+ fps at those same locations in all the benches we have been performing lately on the "Anyone Brave Enough" and "R600 benchmarks" topic here in these forums.

And, not to toot my own horn here, but I owned the 8800GTX for 6 months and know exactly how it performed on launch day. There have been incredible performance improvements with the last one-two driver sets. Nvidia deliberately sat on their @sses and waited for R600 to launch before releasing them as a surprise tactic against ATI. It appears that it might've worked.

[J-Mag]

Now guys are reporting the game runs at 45+ fps at those same locations in all the benches we have been performing lately on the "Anyone Brave Enough" and "R600 benchmarks" topic here in these forums.


With all of the examples you show there is one major issue; you aren't controlling for multiple variables. I try not to compare frame rate or scores among different people's posts because of forgers, cheaters, driver/game setting differences, OS differences, hardware differences, etc...


And, not to toot my own horn here, but I owned the 8800GTX for 6 months and know exactly how it performed on launch day.

I recall all the problems you had with your hardware during that time frame. Can I really trust your results during this hectic time?

[Xion X2]

With all of the examples you show there is one major issue; you aren't controlling for multiple variables. I try not to compare frame rate or scores among different people's posts because of forgers, cheaters, driver/game setting differences, OS differences, hardware differences, etc...

Look at all of the different users in that thread, though. Everyone among about 5-6 people are reporting the same. I don't think anyone is cheating. Even moderators (MUYA) have posted benchmarks in correlation to the rest, and we have all been very careful about setting our graphics options up the same.

I recall all the problems you had with your hardware during that time frame. Can I really trust your results during this hectic time?

For the majority of the time around when I first bought my GTX, I had it running on an Asus P5B Deluxe that never gave me a single issue. It was only months later when I went SLI on the 680i platform that I had problems (and none of these were related to performance; the board ran great when it worked and chipsets weren't fizzling out on me).

But for the first few months that I owned a GTX, it was perfectly stable on the Asus P5B platform. My 3DMark06 score was among the best back then, 12,200.

And Rob (SLINROB), I don't know what in the hell is going on with you lately, but it's beginning to piss me off. You and I have even discussed this over the phone before, so I'm having a really hard time with you denying all of this. I know that we've talked specifically about Rainbow Six Vegas slowing to 45 fps in the past on a single GTX when soldiers were on-camera (whereas you showed now it drifts along at 70+fps on these new drivers) and also about Oblivion slowing down to ~30fps in the pink grass/forests (whereas you've shown now it stays at 45+fps.) Add to this the fact that myself and some others have clearly shown the XT outperforming the GTS on Oblivion just recently, while you say that it doesn't, and I'm not sure what to say.

I'm a little disappointed in you, to say the least. I don't expect you to agree with me on everything just because we're friends; it's perfectly fine if you don't. But you're being a bit hypocritical based on our past conversations about this issue and ignoring facts that are staring you right in the face. You say that you're trying to be non-biased with all of this, but your actions sure don't speak that way.

[sammy sung]

Is that what it is... specatular indeed :p

Well sue me,i was tired and spelled it like a 8 year old:)

[J-Mag]

Look at all of the different users in that thread, though. Everyone among about 5-6 people are reporting the same. I don't think anyone is cheating. Even moderators (MUYA) have posted benchmarks in correlation to the rest.


Yeah I doubt there will be cheaters in this circumstance, but I'd like to see one user run tests on one machine with both drivers. Looks like Rob is doin' that soon...


But for the first few months that I owned a GTX, it was perfectly stable on the Asus P5B platform. My 3DMark06 score was among the best back then, 12,200.


Your early 3dmark score was the best vs what, your later score or other people's scores?

[Xion X2]

Your early 3dmark score was the best vs what, your later score or other people's scores?

Among the best involving others' scores of that time. 12,234, to be exact:

http://www.nvnews.net/vbulletin/showpost.php?p=1120878&postcount=478

[Blacklash]

The only time I got 40FPS in the heaviest grass areas in "Oblivion" with all maxed and iMinGrass 80 was when AA + HDR was broke under Vista. So you'd just get HDR and AF when you tried to use it.

The only way I know to help boost FPS in very heavy grass in "Oblivion" would be no transparency AA, decreasing grass max viewing distance, and running something like iMinGrass 110. Of course no AA at all would help too.

With 8xAA + HDR, Transparency Multisample AA, 16x AF, max grass viewing distance, and iMinGrass 80 or default, I go low as 32FPS, even overclocked. Stock = like Xion's shot 26 or 27 FPS in the worst areas.

Examples-

First the below save with Vista x64 1680x 16xAF iMinGrass 80, all maxed, all on, No AA or Transparency AA = 41FPS @ 652|2054

41FPS
http://m1.freeshare.us/149fs892936_th.jpg (http://m1.freeshare.us/view/?149fs892936.jpg)

From the other thread same save etc etc,

I max 32FPS at the save loaded orientation with Vista x64
1680x 8xAA, Multisample Transparency AA, 16xAF iMinGrass 80, all on and sliders to the right.

32FPS @ 652|2054
http://m1.freeshare.us/149fs874223_th.jpg (http://m1.freeshare.us/view/?149fs874223.jpg)

Stock as above = 26FPS
http://m1.freeshare.us/149fs877117_th.jpg (http://m1.freeshare.us/view/?149fs877117.jpg)

The save:
http://members.cox.net/malficar3/Save%20199%20-%20Corvexius%20Lydon%20-%20The%20Great%20Forest,%20Level%208,%20Playing%20 Time%2004.12.32.ess

So overclocked I can break 40FPS in very heavy grass with 16xAF all maxed and it's a no no with 8xAA + HDR and Transparency AA in addition.