2900XT Ain't That Bad...

Hi Proxy,
No, I wasn't out for a flame war. It's just that the sheer predictability of some guaranteed it. Why should I be asked to back down in the thread I initiated? Is the 8800GTX better based on the evaluation of other reviewers around the web? Hell yeah. But I'm not certain as to the 8800GTS. This thread's title says nothing about the 8800 series, though. If I had posted a thread entitled "My 2900XT makes your 8800 look shriveled and unwashed", then that would certainly be inciting a flame war. But the mere mention of 2900XT + 'not bad' raises hackles and brings the packs. If someone starts flinging the mud at the mere mention of 2900XT, then they were out trolling anyway. This post just gave these 'nomads of the forums' someplace to camp for a few days. I'll give you this, though: I should have been idiotically simple and specific with the thread's title, so some would have realised it was for ATI HD 2900XT owners to discuss the card and their experiences. Ah well, next time, if the forums allow for such a long title and if trolls can stay awake until periods. As for this topic being locked... whatever would others do with their free time then?

And with that said lets pass the peace pipe around. I think most level-headed people here expect the 2900xt's performance to pick up. Just like nVidia owners expect Vista SLI performance to pick up. Trust me, if people really wanted to troll, we'd be talking about ATI's release drivers (had to throw that in there ;) )
 
... Trust me, if people really wanted to troll, we'd be talking about ATI's release drivers (had to throw that in there ;) )

Oddly enough, that would be quite on topic! They really have people scratching their heads at every point on the compass and saying, "Why?"

Tweaktown just reviewed the MSI flavor of 2900XT and made just your point:

"...The new drivers have helped the card a lot - it really does go from a joke of a product to something that you could recommend. The biggest problem we see is that AMD may have already done the damage by not waiting another week or two for the new drivers to come out."

http://www.tweaktown.com/reviews/1115/1/page_1_introduction/index.html
 
It's very obvious you know nothing about video cards. If you did you would have provided evidence that 8800GTS gain/lost 9-10 FPS in that ss you complain about so much. HA! And, you also present a weak argument regarding which card is slower. If you examined the SS you complain so much about it's clear which card came out on top (just one example), HA! Your post is nothing more then regurgitated banter, ;) . The truth of it is, you are seeing performance improvements. That's why you are so upset. :D

LOL

(I don't mean to be a dick)

You keep saying the same thing and I will keep responding in the same way until you finally pick up what we're all trying to get you to understand.

Most benchmarks out today including the GTS and the XT show the XT getting destroyed. Both in price, power consumption, and speed.

Now look how I said ''Most''? Okay, so you have one screen shot where the GTS gets a lower FPS. Others have already said this, I have said this, the reviewer said this. These shots are inconsistent. They are not fair. And the GTS beats the XT in past reviews of the card including benches of Oblivion. (And benchmarks that aren't even that old) I guess I'll try one more time to get you to understand.

The fact that the GTS has more trees/canopies shadows ect. to produce makes the card work harder. If you have EVER played oblivion you'd know the game is like that, performance is very ranged. From inside to outside you could lose 20fps, or more. Looking from the ground to the air is a huge thing to consider as well. (more grass, shadows, rocks ect.)

That said, and with the fact that the reviewer even said the screen shots are inconsistent how can you continue to protect the XT? Not to mention that in all other benchmarks including the GTS and XT testing w/oblivion (that I've seen) have been like all other benchmarks for the XT, bad performance.

Failing once again, to keep up with a card that Nvidia released 4+ months ago.

You can continue to try and argue the point, but I'm done typing all this out.

Enjoy your card. Hope it continues to impress you. And I guess if you like the card thats all that should matter.
:)
 
... Enjoy your card. Hope it continues to impress you. And I guess if you like the card thats all that should matter.
:)

The Tweaktown review (are they on the respectable list?) has the 2900XT and the 8800GTS pretty much tying in most games. They did kind of a half-assed number of turning it up with 4xAA/16xAF benches, but only at 1280x1024 (not 1920x1200 or 2560x1600) where at the lower number the GTS managed a few frames better.

Anyway, thanks for the well wishes, Fiddyownz. Hope the new Forceware drivers are good for you.
 
Oddly enough, that would be quite on topic! They really have people scratching their heads at every point on the compass and saying, "Why?"

Tweaktown just reviewed the MSI flavor of 2900XT and made just your point:

"...The new drivers have helped the card a lot - it really does go from a joke of a product to something that you could recommend. The biggest problem we see is that AMD may have already done the damage by not waiting another week or two for the new drivers to come out."

http://www.tweaktown.com/reviews/1115/1/page_1_introduction/index.html

Why did they test with no AA/AF first. Strangely enough, this is the same situation nVidia was at years ago. Specifically, the cards were fast but as soon as you started using AA/AF performance went... south. That review is suspect, at best...lol. It looks like [H] will be doing an overclocking re-review pretty soon so we shall see soon enough.

And yeah, why couldn't they wait a couple of weeks. If performance was right around the corner why not wait and avoid the negative pub?
 
Get this through your thick freakin' skulls--AA is the 2900XT's problem. If a review or benchmark isn't using atleast 4xaa, it is totally worthless. The "back end", the hardware that does the AA, is borked or insufficient--so make sure when you read the reviews you look for the settings you are actually going to use. I was forced to RMA this card because the AA performance was so poor. It couldn't even do 8xaa in Battlefield2.
 
Hi Blackstone,
Just out of curiosity, what motherboard were you using the HD 2900XT on? Sorry to ask, but I didn't see it in your sig.
 
Get this through your thick freakin' skulls--AA is the 2900XT's problem. If a review or benchmark isn't using atleast 4xaa, it is totally worthless. The "back end", the hardware that does the AA, is borked or insufficient--so make sure when you read the reviews you look for the settings you are actually going to use. I was forced to RMA this card because the AA performance was so poor. It couldn't even do 8xaa in Battlefield2.

http://forums.anandtech.com/messageview.cfm?catid=31&threadid=2056170
 

Nice find!

??? What is that post supposed to show? That a guy can't even run AAA with CROSSFIRED 2900XT in BF2?

heheh, hey i dont even run it on advanced in bf2. bf2 is the game ive seen it have the largest effect in but i like to keep well over 100fps at 1920x1200 so i just keep it on performance even with two HD2900's

[H] had two GTX's running 2142 in a much higher res w/16 AA and avged 90+ fps. (Yes I know its GTX vs GTS but these cards are also running much higher settings). Their single 8800gtx was able to run maxed settings, 16TRSS AA/16AF @ 1600x1200 with over 80fps avg.

my adaptive AA works fine though, i can tell its working just from the fps drop. just keep an eye on the fps while making changes, like adaptive aa, you should see a large fps gain with it disabled opposed to set on advanced.

So erm.. Was there something in those posts that was supposed to say something good about the 2900XT and AA? They mentioned a lot about how it wouldn't even turn on in games when you had it turned on in the control panel.
 
No, I got 8xaa to work, but the frames were so bad in BF2 that it was a joke. I use a 8800GTX and I'll never look back. THe performance is so good it is sick. 16xqaa 16x quality af 1680 x 1050 and I am locked at 60fps. Awesome!
That's not true, because I tried the similar thing and didn't run into the those problems.

Blackstone said:
Also note that in Battlefield 2, you would expect the card to handle such an old game with ease, but 8x antialiasing is also laggy when there are lots of trees with lots of leaves. On most maps, 8xaa is not really an option. As this is the only in game option higher than 4x available to me (and forcing at the driver level doesn’t work for me) this card does not yield any gameplay advantage over my old x1800XT except for better framerates.
Post
As you can see in my screen shots I am not getting the same results. Although he claimed that he had these problems he never took a screen shot. Anyone who plays BF2 know that the print screen key is used to take as many screen shots as you want. Yet, he didn't link 1 screen shot in his original post.

The only screen shots taken was for Oblivion and, RC7 beta drivers were used.
 
What is not true is that the HD 2900XT is a joke.

You're missing the point. His experience tells him otherwise. Just like the OP's experience tells him that the 2900xt is not that bad.

I can relate perfectly to how Blackstone feels. A few years ago i bought a launch BBA 9500PRO. Not mentioned in the rave reviews the 9000 series received, a minority of users had the infamous, "VPU failed to recover" error. I wish i could have returned it. Actually, i should have taken that card to the shooting range. Unfortunately, i had to wait for what seemed like forever for ATI to fix the driver issue. Blackstone chose not to deal with the AA issue. Not having something work like it's supposed to after plunking down a lot of cash is frustrating and ATI is to blame for that. I'm not going to excuse nVidia either.

/being rant
Why haven't more reviewers taken ATI to task for releasing a new card with poor drivers? What the hell happened to QA? A customer's first impressions are hard to overcome and just about everyone is thinking that the 2900xt is not that good of a card (justified or not).
/end rant
 
You're missing the point. His experience tells him otherwise. Just like the OP's experience tells him that the 2900xt is not that bad.

I can relate perfectly to how Blackstone feels. A few years ago i bought a launch BBA 9500PRO. Not mentioned in the rave reviews the 9000 series received, a minority of users had the infamous, "VPU failed to recover" error. I wish i could have returned it. Actually, i should have taken that card to the shooting range. Unfortunately, i had to wait for what seemed like forever for ATI to fix the driver issue. Blackstone chose not to deal with the AA issue. Not having something work like it's supposed to after plunking down a lot of cash is frustrating and ATI is to blame for that. I'm not going to excuse nVidia either.

/being rant
Why haven't more reviewers taken ATI to task for releasing a new card with poor drivers? What the hell happened to QA? A customer's first impressions are hard to overcome and just about everyone is thinking that the 2900xt is not that good of a card (justified or not).
/end rant

The same applies to the OP does it not? No need to be one sided.
 
Oddly enough, that would be quite on topic! They really have people scratching their heads at every point on the compass and saying, "Why?"

Tweaktown just reviewed the MSI flavor of 2900XT and made just your point:

"...The new drivers have helped the card a lot - it really does go from a joke of a product to something that you could recommend. The biggest problem we see is that AMD may have already done the damage by not waiting another week or two for the new drivers to come out."

http://www.tweaktown.com/reviews/1115/1/page_1_introduction/index.html
The card gets its ass beat when they turned AA and AF on. The only time it got good results was when aa and af were off, which no one buying a 2900XT would do.
 
I cant believe people on this forum are actually buying the 2900xt. What part of your rational human psyche tells you to do that after viewing the benchmarks were the 2900 gets its ass handed to it by the 8800 gtx and gts, dont even get me started on the ultra. We buy things for the best performance from viewing the gaming benchmarks from reputable sites, never mind the synthetic ones such as 3dmark, running games is where it counts, screw the brand type be it ati or nvidia, right now I want best performance for my money, don't you? This time around 8800ultra>8800gtx>8800gts>2900xt. So after seeing that youre gonna go ahead and buy the 2900xt?:eek:
 
The card gets its ass beat when they turned AA and AF on. The only time it got good results was when aa and af were off, which no one buying a 2900XT would do.

They only tested at a low resolution (1280x1024), and in fact the HD 2900XT scales pretty well as resolutions rise. If they had continued on up the scale, it's more probable the 2900XT would have equaled / bested the GTS in many of those games (please take note I didn't say all, and I'm referring to the games used in the Tweaktown article). Now that is speculation based on what reviews at reputable sites using Cat 7.5 have been showing all around.
 
http://www.hardforum.com/showthread.php?t=1196019

http://www.tweaktown.com/reviews/1115/7/page_7_benchmarks_half_life_2_lost_coast_hdr/index.html
if, after reading the above tweak town review, you decide to go on a little hunt looking for other reviews, pay close attention to the date, and driver version.

No, imho, this cards performance squarly lands in its price range.

God the Pro-Nvidia fanboyism is just staggaring....

Anyways, ATI has made a schrewd move here. They could have released the 1Gb GDDR4 version along with the 512Mb version, but chose not to. This gives the drivers time to mature between releases, and when the 1Gb version is reviewed, its likly that the 512Mb will be re-reviewed in the process (as one of the cards they're comparing it too).

When this happens, a large portion of [H]ard|forum (including its staff) may be eating their words.
 
Get this through your thick freakin' skulls--AA is the 2900XT's problem. If a review or benchmark isn't using atleast 4xaa, it is totally worthless. The "back end", the hardware that does the AA, is borked or insufficient--so make sure when you read the reviews you look for the settings you are actually going to use. I was forced to RMA this card because the AA performance was so poor. It couldn't even do 8xaa in Battlefield2.

No, get this through your thick skull, the problem is AF not AA. The 2900XT is very, very deficient on texture filtering capability. Around HALF what the 8800GTX has. So you wonder why it's a bottleneck when AF is turned up?

ATI equipped this card with very deficient texture capabilities, just like the R580 before it. It's a shader monster and crushes G80 in that area though.


Go read Xbit labs HD2900 review, it will tell you the scoop on the 2900's lack of texturing ability: http://www.xbitlabs.com/articles/video/display/radeon-hd-2900-games.html
 
http://www.hardforum.com/showthread.php?t=1196019

http://www.tweaktown.com/reviews/1115/7/page_7_benchmarks_half_life_2_lost_coast_hdr/index.html
if, after reading the above tweak town review, you decide to go on a little hunt looking for other reviews, pay close attention to the date, and driver version.

No, imho, this cards performance squarly lands in its price range.

God the Pro-Nvidia fanboyism is just staggaring....

Anyways, ATI has made a schrewd move here. They could have released the 1Gb GDDR4 version along with the 512Mb version, but chose not to. This gives the drivers time to mature between releases, and when the 1Gb version is reviewed, its likly that the 512Mb will be re-reviewed in the process (as one of the cards they're comparing it too).

When this happens, a large portion of [H]ard|forum (including its staff) may be eating their words.

sorry, i smell fanboi, failed to read after page 11? please tell me, so that i can read it for you.
 
ard|forum (including its staff) may be eating their words.


While i do agree that the 2900XT does need a re-review with newer drivers, it is no fault of the reviewers that the card performed the way that it did. The reviewers gave and honest opinion of the cards current performance. If you look back at the 8800 series it always hasnt been roses and sugery sweets either, but the 8800 series has had several times to be reviewed with newer drivers each time and the performance has increased.

All i ask is please do a re-review when there are a couple DX10 games that you can include in the review. Yes, I know it will still be to hot and use to much power, but humor me.
 
http://www.hardforum.com/showthread.php?t=1196019

http://www.tweaktown.com/reviews/1115/7/page_7_benchmarks_half_life_2_lost_coast_hdr/index.html
if, after reading the above tweak town review, you decide to go on a little hunt looking for other reviews, pay close attention to the date, and driver version.

No, imho, this cards performance squarly lands in its price range.

God the Pro-Nvidia fanboyism is just staggaring....

Anyways, ATI has made a schrewd move here. They could have released the 1Gb GDDR4 version along with the 512Mb version, but chose not to. This gives the drivers time to mature between releases, and when the 1Gb version is reviewed, its likly that the 512Mb will be re-reviewed in the process (as one of the cards they're comparing it too).

When this happens, a large portion of [H]ard|forum (including its staff) may be eating their words.

Hi MrWizard6600,
You've made a lot of very good points there, and the observation about AMD's motives in releasing a 1Gb version does give us something to think about: sites will be falling all over each other to declare 8800GTX/Ultra the winner, and in the process might very well show the 2900XT (512Mb) in a more favorable light. Might. If that were the case, AMD might do well to hold off until one more driver release.

^^^^^^^^^^^^ HD29000XT in your profile.. must be new.. LAWL!

Brother! How old are you?
 
I agree that the HD X2900XT isn't that bad.
I also agree that the HD X2900XT isn't that good, seeing how there is in fact better out there.
 
^^^^^^^^^^^^ HD29000XT in your profile.. must be new.. LAWL!

What are you, 13? This is a message board for mature discussion. If you don't have anything mature or intelligent to add, please do everyone (including yourself) a favor and don't post.

And in case you're wondering, no, I'm not a fanboi on either side. I tend to jump brands every generation actually. It just bugs me to see people who are otherwise intelligent degrade into a screaming, laughing child with their fingers in their ears...all over a computer hardware brand.
 
The same applies to the OP does it not? No need to be one sided.

Are you actually reading each post or are you just arguing for the sake of arguing?

From my last post: Just like the OP's experience tells him that the 2900xt is not that bad.

Reading comprehension: learn it, use it.
 
They only tested at a low resolution (1280x1024), and in fact the HD 2900XT scales pretty well as resolutions rise. If they had continued on up the scale, it's more probable the 2900XT would have equaled / bested the GTS in many of those games (please take note I didn't say all, and I'm referring to the games used in the Tweaktown article). Now that is speculation based on what reviews at reputable sites using Cat 7.5 have been showing all around.

Show me one reputable benchmark where the 2900XT did well with high AA and AF. The only sites praising it are the ones that are either biased or didn't test them with AA and AF on.
 
God the Pro-Nvidia fanboyism is just staggaring....

nösferatu;1031174109 said:
sorry, i smell fanboi, failed to read after page 11? please tell me, so that i can read it for you.

Sigh... Obviously there is some strong arguments on either side, but why does it have to come down to calling the other side fanboys? Anyone with a strong opinion can be CALLED a fanboy, that's easy... but how about the word fanboy, fanboi, fanbuoy, etc. is struck from everyones vocabulary from here on out in reference to the 2900XT, it doesn't get anyone anywhere except to make the other side upset.

Maybe I'll put in my two cents. As I've said before, the 2900XT is a decent card, but STILL, if I'm paying $400+ dollars for a video card, I would except to be able to turn things up on nearly every game. That INCLUDES AA/AF to reasonable levels. And at a decent res, ie. 1680x1050 preferably AT LEAST. The 2900XT does well at this screen res WITHOUT AA/AF, and that's why I wouldn't consider it much. That, and I would need to upgrade my power supply, which would be another $150+.

As it stands right now, I bought my card for $250, I play at 1280x960 or 1600x1200 if the card can take it, and I'm mostly playing newer games that aren't being able to enable AA w/effects, ie. Call of Juarez. But it plays that game amazingly with everything turned up, Stalker it does ok in (vram troubles I think, but still have everything turned on @ 1280x960), SupCom plays awesome. I've always used ATI, but I'm happy I joined the green team this round :D

Edit: In regards to the tweaktown review... I really don't understand how they pulled that overclocking section out their ass. It makes next to no sense. They overclock, show a .17 increase in min fps as what looks like 10fps increase, and not only that, used a game that is heavily CPU dependant... wtf was the point of that? I'm not saying it's a worthless review, but I think there is something wrong with those guys if they think that is showing a good overclocking summary....
 
Show me one reputable benchmark where the 2900XT did well with high AA and AF. The only sites praising it are the ones that are either biased or didn't test them with AA and AF on.

Hi Eagle!
Well, here you go. I'm sure there are more, but these are sites I often view. I trust not all of them are disreputable in your sight. Each one benched using AA & AF, and I hope you might notice how the 2900XT scales as resolution increases, and I'd also like to cast attention on the competition between the 2900XT and the 8800GTS 640. In more games than not, the 2900XT scores higher.

1. http://www.bit-tech.net/hardware/2007/05/16/r600_ati_radeon_hd_2900_xt/1
2. http://www.bjorn3d.com/read.php?cID=1099&pageID=3433
3. http://www.hardwarezone.com/articles/view.php?id=2260&cid=3&pg=1
4. http://www.hexus.net/content/item.php?item=8687&page=1
5. http://www.legionhardware.com/document.php?id=650
6. http://www.legitreviews.com/article/503/1/ * (OK, not the best review, but it is a head to head with the GTS 640)
7. http://www.pcper.com/article.php?aid=406&type=expert&pid=2
8. http://www.techreport.com/reviews/2007q2/radeon-hd-2900xt/index.x?pg=1
9. http://www.tweaktown.com/reviews/1115/msi_radeon_hd_2900_xt_512mb_graphics_card/index.html
10. http://www.xbitlabs.com/articles/video/display/radeon-hd-2900-games.html
 
The TechReport review always makes me laugh because they say they couldn't enable full dynamic lighting, so they used static lighting, making it a DX8 crap looking game o.0

I like to go to the STALKER benchs because I know the 2900XT has had problems with that when it first came out. Apparently it's gotten a lot better with the Legion Hardware review, as it has it now beating the GTX in most resolutions, and by a lot! Not really sure how that happened... but I guess so. Not sure why they had Shadow Detail down so low though. Although I do like how they show all the settings they used.
 
The TechReport review always makes me laugh because they say they couldn't enable full dynamic lighting, so they used static lighting, making it a DX8 crap looking game o.0

I like to go to the STALKER benchs because I know the 2900XT has had problems with that when it first came out. Apparently it's gotten a lot better with the Legion Hardware review, as it has it now beating the GTX in most resolutions, and by a lot! Not really sure how that happened... but I guess so. Not sure why they had Shadow Detail down so low though. Although I do like how they show all the settings they used.

Yeah, the equivalent of DX8 really isn't something any of us are gunning for.
 
I think this sums it up the best, and remember they are comparing the GTS @ $379 and the 2900 @ $399

Looking @ newegg right now, cheapest 2900 is $419 ($409 but OOS), cheapest 8800gts 640mb is $379 ($349 after $30 MIR from Evga, very fast rebate btw), or $329AR (OOS).

valuebigbq5.png


So that means that the graph is actually more in favor of the GTS since its cheaper than their #s, while the 2900 is more expensive.
 
Back
Top