So DT was right

[Extelleron]

Originally posted by: Matt2

Originally posted by: Extelleron

Originally posted by: Matt2

Originally posted by: Genx87
Is the new buzzphrase ATI fanboi's going to use "Proper Review"?

I thought the war cry was, "G80 is not optimized for DX10!".

I think at the moment it's, "Wait a few more months for the driver bomb!".

To put it into perspective. How much performance did G80 gain from release drivers to current? I'm not talking stability, I am talking pure performance.

FEAR 2560x1600 4xAA/8xAF

GeForce 8800GTX Release: 32 FPS
GeForce 8800GTX @ Ultra launch: 49 FPS

Oblivion HDR + AA 2560x1600 4xAA/8xAF

GeForce 8800GTX Release: 17.0 FPS
GeForce 8800GTX @ Ultra launch: 22.6 FPS

Two titles, gains of 53% and 33%.

From Firingsquad.

What about something more like 1920x1200 or 1600x1200? I'm really not interested in 2560x1600 numbers because I dont have a 30" LCD.

Oblivion, Release - > Ultra release sees a 4% performance improvement @ 1920x1200 and a 2% improvement @ 1600x1200.

Fear, Release - > Ultra release sees a 6% performance improvement @ 1920x1200 and and another 6% improvement @ 1600x1200.

CoH, Release - > Ultra release sees a 35% performance improvement @ 2560x1600, a 18% improvement at 1920x1200, and a 20% improvement @ 1600x1200.

It seems most of the performance increases come at 2560x1600, but CoH sees some significant increases all around the board.

Unlike the GTX though, the HD 2900XT shows clear PROBLEMS with performance. In the Inquirer review it performed faster at 2560x1600 than 1920x1200 in one or two tests. Clearly there's something very weird going on.

 

[Matt2]

Originally posted by: cmdrdredd
well, there's no point in doing those tests. Reviewers test the cards at near MAX resolution and settings to see what they are capable of at the highest level of quality.

Faster at Oblivion HDR + AA 2560x1600 4xAA/8xAF means it's faster period.

Yeah, but people are claiming that drivers will save R600 because G80 had significant performance improvements due to drivers.

2560x1600 numbers mean nothing to me. A 53% increase in FEAR and a 33% increase in Oblivion at 2560x1600 is not going to tell me what the gains were at lower resolutions.

So, if a GTX was getting 50fps at 1920x1200 at release, am I to believe that it now gets 76.5fps?

No, that would be dumb.

 

[Extelleron]

Originally posted by: cmdrdredd

Originally posted by: Matt2

Originally posted by: Extelleron

Originally posted by: Matt2

Originally posted by: Genx87
Is the new buzzphrase ATI fanboi's going to use "Proper Review"?

I thought the war cry was, "G80 is not optimized for DX10!".

I think at the moment it's, "Wait a few more months for the driver bomb!".

To put it into perspective. How much performance did G80 gain from release drivers to current? I'm not talking stability, I am talking pure performance.

FEAR 2560x1600 4xAA/8xAF

GeForce 8800GTX Release: 32 FPS
GeForce 8800GTX @ Ultra launch: 49 FPS

Oblivion HDR + AA 2560x1600 4xAA/8xAF

GeForce 8800GTX Release: 17.0 FPS
GeForce 8800GTX @ Ultra launch: 22.6 FPS

Two titles, gains of 53% and 33%.

From Firingsquad.

What about something more like 1920x1200 or 1600x1200? I'm really not interested in 2560x1600 numbers because I dont have a 30" LCD.

well, there's no point in doing those tests. Reviewers test the cards at near MAX resolution and settings to see what they are capable of at the highest level of quality.

Faster at Oblivion HDR + AA 2560x1600 4xAA/8xAF means it's faster period.

Well yes, testing at 2560x1600 w/ AA and AF tests a cards true performance and which card is really stronger, but it doesn't always reflect real-world gaming scenarios. The vast majority of people don't have a 2560x1600 display and for them, those results don't mean a thing. It's nice to know how powerful a card is but if you're not going to see a benefit at your resolution who cares.

 

[Ackmed]

Extelleron is correct. Newer drivers have shown in reviews that there is a increase on performance. Its plainly obvious that better drivers for the 2900XT will increase performance. I doubt to catch up to the 8800GTX, but enough to be much closer. It doesnt make you a fanboy to want such drivers, or to think it will happen.

Originally posted by: Matt2

If you dont like my posts then ignore them. Just like I ignore 90% of your posts. I dunno, I try not to call you out on your bias whenever possible, but you love to call me out on mine. Maybe you should grow up?

Maybe my posts are all the same in every thread because there are 100 different threads on the same topic?

No you dont, you always answer me.

Also, Ive said I am bias towards ATi, so uh...

You act like a kid, saying "neener neener" without doing anything to benefit the thread. Often times just trolling off topic. There is no need to post in such a manner. I dont know why this is so hard to understand.

 

[Matt2]

Originally posted by: Ackmed
Extelleron is correct. Newer drivers have shown in reviews that there is a increase on performance. Its plainly obvious that better drivers for the 2900XT will increase performance. I doubt to catch up to the 8800GTX, but enough to be much closer. It doesnt make you a fanboy to want such drivers, or to think it will happen.

Originally posted by: Matt2

If you dont like my posts then ignore them. Just like I ignore 90% of your posts. I dunno, I try not to call you out on your bias whenever possible, but you love to call me out on mine. Maybe you should grow up?

Maybe my posts are all the same in every thread because there are 100 different threads on the same topic?

No you dont, you always answer me.

Also, Ive said I am bias towards ATi, so uh...

You act like a kid, saying "neener neener" without doing anything to benefit the thread. Often times just trolling off topic. There is no need to post in such a manner. I dont know why this is so hard to understand.

That's why I said "90%" of the time. The other 10% falls into the category of when you come into a thread like this one, which you were not a part of, just to call me a troll.

Let me ask you Ackmed, what have YOU done to benefit this thread?

 

[apoppin]

Originally posted by: Wreckage
Everyone was bashing the early R600 benchmarks from Daily Tech. However, many sites are pretty much backing what they posted.

I think overall DT has been a pretty good source of rumor material. Far better than the Inq and a lot of people are eating crow today.

you're the OP ... post the list

i see more positive then negative ...
but then you only see negative things

your's is a *hate* thread
:Q

i certainly *don't* care to convince you or ANYone for that matter

you DON'T like it ... buy something else .... but you won't *fool anyone* that you are "impartial" or even

.... "fair and balanced'

--i know because i am

 

[Ackmed]

Called you out, exposing you're trolling and useless posts? Other than that, nothing. You always reply to me, dont try the "I dont read your posts" trick.

Also, I didnt call you a troll, I asked if thats all you do. Because thats all you have done today. Taking a page of gastanfors playbook. Do you think going around gloating because your "team" won is helping anything? Seriously, take a step back and look at your posts...

 

[golem]

Originally posted by: Ackmed
Extelleron is correct. Newer drivers have shown in reviews that there is a increase on performance. Its plainly obvious that better drivers for the 2900XT will increase performance. I doubt to catch up to the 8800GTX, but enough to be much closer. It doesnt make you a fanboy to want such drivers, or to think it will happen.

It doesn't make you a fanboy to want these drivers to come out. But you're giving REALLY bad advice to people to by telling them to buy this card on the hopes that they will, or to point to a bad port of an xbox 360 game as a barometer of how this card will do in UE3 based games.

Right now, today, 2900XT is just a bad buy compared to the alternatives for most people. If the drivers improve performance on a broad range of games, and if the price comes down then it could have some appeal to the broader market. But for now, not so much.

 

[dreddfunk]

I think I've finally grasped the allure of calling out, "fanboy!" at the top of your lungs. It's a lot like watching Jerry Springer: one gets to concoct a ridiculous, stereotyped image of a moron against which one's own--ever so humbly--reasoned and thoughtful insights stand in stark contrast. In an instant one can create an 'enemy' that is easily defeated with one fell swoop of one's rhetorical and analytical might.

In other words, it's like mainlining chocolate, only for your ego.

Meanwhile the HD2900xt is still somewhat of a disappointment, though not an utter disaster. Probably most folks who have some brand loyalty to ATI wouldn't hesitate to agree, if they could be heard over the gleeful cackling of those looking for a quick ego-fix.

 

[OneOfTheseDays]

I fully expect ATI to put out drivers that substantially improve performance of the R600. Take a look at their last series, the X1800, and how much it improved with sucessive driver releases. On paper the R600 should demolish the GTS, but it doesn't. It would seem more probable that they haven't quite figured out how to optimize for it than it being technically inferior to the GTS.

 

[cmdrdredd]

Originally posted by: Extelleron

Originally posted by: cmdrdredd

Originally posted by: Matt2

Originally posted by: Extelleron

Originally posted by: Matt2

Originally posted by: Genx87
Is the new buzzphrase ATI fanboi's going to use "Proper Review"?

I thought the war cry was, "G80 is not optimized for DX10!".

I think at the moment it's, "Wait a few more months for the driver bomb!".

To put it into perspective. How much performance did G80 gain from release drivers to current? I'm not talking stability, I am talking pure performance.

FEAR 2560x1600 4xAA/8xAF

GeForce 8800GTX Release: 32 FPS
GeForce 8800GTX @ Ultra launch: 49 FPS

Oblivion HDR + AA 2560x1600 4xAA/8xAF

GeForce 8800GTX Release: 17.0 FPS
GeForce 8800GTX @ Ultra launch: 22.6 FPS

Two titles, gains of 53% and 33%.

From Firingsquad.

What about something more like 1920x1200 or 1600x1200? I'm really not interested in 2560x1600 numbers because I dont have a 30" LCD.

well, there's no point in doing those tests. Reviewers test the cards at near MAX resolution and settings to see what they are capable of at the highest level of quality.

Faster at Oblivion HDR + AA 2560x1600 4xAA/8xAF means it's faster period.

Well yes, testing at 2560x1600 w/ AA and AF tests a cards true performance and which card is really stronger, but it doesn't always reflect real-world gaming scenarios. The vast majority of people don't have a 2560x1600 display and for them, those results don't mean a thing. It's nice to know how powerful a card is but if you're not going to see a benefit at your resolution who cares.

they wnat to test the power of the card. That's the entire point of the review testing.

Simple fact is most buyers say "hey I want a gaming PC" so they go to dell and have 2 choices... x2900 or 8800gts/gtx they will go on google and type in x2900 vs 8800 and come up with a review. That review will likely show that the 8800 is faster. They will buy that one. They will not be sitting around twiddling their thumbs waiting for "magical drivers". Obviously new drivers can and almost always do increase performance, but I'm looking at the performance you'd be getting if you had one in your box right now.

To clarify what I'm saying.

1) Most buyers of gaming PCs do not read forums and use beta drivers or build their own PC. Now I know almost every one of us here talking now do build our own PCs, but not everyone and that is the people this release will turn over to Nvidia.
2) most buyers will look at benchmark FPS numbers as the sole factor as to which card to purchase
3) most buyers (at least ones I know) are not content to sit on a card waiting for improvement. They buy a card because they want it to be fast now, the x2900 is fast...but you can save $100 and not worry about power requirements with a 8800gts, getting better performance right now...today.

 

[dreddfunk]

I don't know enough about how drivers work to optimize performance to say, but from what little I actually understood of the technical discussion in AT's review, the efficiency of the architecture seems to be determined by how well the compiler takes advantage of it. I honestly don't know if that can be addressed in the hardware's driver. It sure sounds like it depends on how the game itself is compiled.

 

[apoppin]

Originally posted by: OneOfTheseDays
I fully expect ATI to put out drivers that substantially improve performance of the R600. Take a look at their last series, the X1800, and how much it improved with sucessive driver releases. On paper the R600 should demolish the GTS, but it doesn't. It would seem more probable that they haven't quite figured out how to optimize for it than it being technically inferior to the GTS.

how is it that we are now agreeing?



if the HD2900xt improves *only* as much as the 8800GTS did - since release - we should have a 'final' nice performance increase over the GTS ... which is evidently not going to get much faster then it is now ... so the performance delta widens in favor of HDxt