HD 2900XTX Benches

[Regs]

Talk about worse timing in the world.

This reminds me when HP bought Compaq for 25 billion. AMD got way too far over their heads and I hope their stock holders rise up as one and slay Hector and the rest of their useless board of (no direction) directors.

You're manufacturing lines are behind, your products haven't been updated in years, and you can't feed your own supply lines. So...lets invest 4-6 billion (8-12 billion after interest) in another under dog company suffering the same problems. I was hoping AMD decided to merge with ATi because of their pipeline of products which was the only logical explanation. Today's grim outlook makes me puzzle how Hector even got his job.

If they end up losing the 2nd half of this year by another big margin then I'm pretty sure board members and stock holders will have Hector and crew looking for new jobs managing ice cream kiosks.

 

[Genx87]

Originally posted by: Regs
Talk about worse timing in the world.

This reminds me when HP bought Compaq for 25 billion. AMD got way too far over their heads and I hope their stock holders rise up as one and slay Hector and the rest of their useless board of (no direction) directors.

You're manufacturing lines are behind, your products haven't been updated in years, and you can't feed your own supply lines. So...lets invest 4-6 billion (8-12 billion after interest) in another under dog company suffering the same problems. I was hoping AMD decided to merge with ATi because of their pipeline of products which was the only logical explanation. Today's grim outlook makes me puzzle how Hector even got his job.

If they end up losing the 2nd half of this year by another big margin then I'm pretty sure board members and stock holders will have Hector and crew looking for new jobs managing ice cream kiosks.

I wasnt happy with the merger from day one and nearly 12 months after the announcement my worst fears are coming true.

I really hope these benchmarks are wrong in some way because if they arent, this is the 5800 debacle all over again and I dont think AMD\ATI has the pockets nor intelligent management to survive it like Nvidia.

 

[ShadowOfMyself]

Originally posted by: Genx87

Originally posted by: Regs
Talk about worse timing in the world.

This reminds me when HP bought Compaq for 25 billion. AMD got way too far over their heads and I hope their stock holders rise up as one and slay Hector and the rest of their useless board of (no direction) directors.

You're manufacturing lines are behind, your products haven't been updated in years, and you can't feed your own supply lines. So...lets invest 4-6 billion (8-12 billion after interest) in another under dog company suffering the same problems. I was hoping AMD decided to merge with ATi because of their pipeline of products which was the only logical explanation. Today's grim outlook makes me puzzle how Hector even got his job.

If they end up losing the 2nd half of this year by another big margin then I'm pretty sure board members and stock holders will have Hector and crew looking for new jobs managing ice cream kiosks.

I wasnt happy with the merger from day one and nearly 12 months after the announcement my worst fears are coming true.

I really hope these benchmarks are wrong in some way because if they arent, this is the 5800 debacle all over again and I dont think AMD\ATI has the pockets nor intelligent management to survive it like Nvidia.

They definitely miss the PR expertise of nvidia... Not only that, but Nvidia quickly came out with the 5900 ultra to fix things, and I doubt AMD has something waiting to replace this... whatever you wanna call it

 

[PingSpike]

I gotta agree with modular. All signs pointed to nothing special. When you have it you "accidentally leak" benchmarks to sites to steal your competitors thunder before you release. When you don't, you keep your lips tight. And we've had tight lips for a very long time. I was hoping they would pull it off, but I'm not all that surprised that they didn't. I'll await more benchmarks of course, but it doesn't look good now.

Here's what I think happened. The x1900 parts were way ahead of nvidia's 7 series on IQ, AA/AF performance and kicked its ass in shader heavy games which seemed to be the future. They tried to extend on that tech, thinking nvidia didn't have anything and would take at least another generation to stop the bleeding.

Out comes the 8 series...with all its terrible drivers and probably to early release. But with it nvidia corrected their poor IQ and switched to a much more shader friendly architecture. I think ATI just plain wasn't ready for it. I can't say I blame them, I didn't really expect nvidia to correct all of those shortcomings in one product generation either.

I bet they couldn't get the yields needed to beat either of the original 8800 parts. I think they were hoping with the wait they'd get the yields up to beat the GTX but after all this time it only scaled to beat the GTS speeds. I personally think they should have cut their losses earlier and just released them to compete in the low high end and then went right back to the drawing board for the next generation.

At this point, I think ATI should focus on releasing competitve parts in the $200 range to capitalize on nvidia's fvck up there.

 

[PingSpike]

Originally posted by: Genx87

Originally posted by: Regs
Talk about worse timing in the world.

This reminds me when HP bought Compaq for 25 billion. AMD got way too far over their heads and I hope their stock holders rise up as one and slay Hector and the rest of their useless board of (no direction) directors.

You're manufacturing lines are behind, your products haven't been updated in years, and you can't feed your own supply lines. So...lets invest 4-6 billion (8-12 billion after interest) in another under dog company suffering the same problems. I was hoping AMD decided to merge with ATi because of their pipeline of products which was the only logical explanation. Today's grim outlook makes me puzzle how Hector even got his job.

If they end up losing the 2nd half of this year by another big margin then I'm pretty sure board members and stock holders will have Hector and crew looking for new jobs managing ice cream kiosks.

I wasnt happy with the merger from day one and nearly 12 months after the announcement my worst fears are coming true.

I really hope these benchmarks are wrong in some way because if they arent, this is the 5800 debacle all over again and I dont think AMD\ATI has the pockets nor intelligent management to survive it like Nvidia.

Yeah, I never thought the merger was a good idea. AMD doesn't have the resources intel does, in order to fight them it has to remain nimble. I couldn't understand why AMD wanted to buy ATI and open up another front to fight on when they really should have thrown all cash into new CPU products. Furthermore, ATI's direct competitor was also AMDs strongest chipset maker so they were pissing on their ally. I suppose the expertise could allow them to field some innovative projects in the future...but right now AMD should be thinking about the present, because they don't have a competitive money making product and they won't survive until tomorrow without one.

 

[StopSign]

Originally posted by: Modular
Is anyone really surprised that this card is a dud? All signs have been pointing to a failure on launch of this card for quite a long time now: no pre-released benches, NDA, saying they have the thing ready to ship "today" but not doing so, delayed release dates (again by the way), etc, etc, etc. At least a few strongly allied persons will be buying these cards, but it's actually a disadvantage to do so.

Exactly! I can't believe everyone here is shocked by the lackluster performance of R600. This is exactly what I predicted when it was pushed back the first time. Do you people honestly believe that they were pushing it back in order to blow away G80?

 

[Ackmed]

These benches dont look good for people waiting for the R600 at all. While Id like some real reviews before I make a judgement, these dont bold well. If these numbers hold up, ATi needs lower prices, and more features to close the performance gap.

Originally posted by: StopSign

Originally posted by: Modular
Is anyone really surprised that this card is a dud? All signs have been pointing to a failure on launch of this card for quite a long time now: no pre-released benches, NDA, saying they have the thing ready to ship "today" but not doing so, delayed release dates (again by the way), etc, etc, etc. At least a few strongly allied persons will be buying these cards, but it's actually a disadvantage to do so.

Exactly! I can't believe everyone here is shocked by the lackluster performance of R600. This is exactly what I predicted when it was pushed back the first time. Do you people honestly believe that they were pushing it back in order to blow away G80?

They said they pushed it back to launch all cards, from highend to budget at once. Not because of any performance problems. Lots and lots of products have been delayed in the past, and have been very good products. Being delayed for one reason or another, doesnt mean its not going to be good. Having no pre-released benches is common, not the other way around. NDA's are common, again, not the other way around. Both NV and ATi do these two things.

 

[dreddfunk]

It certainly looks like the highest 2900 is going to be somewhere between underwhelming and down right horrific: late, slower and less power-efficient than its competitor. What does it say when the best case scenario is repeat of the x1800xt? Underwhelming at the time, hidden value later. The only thing that I can imagine saving the 2900 series is if DX10 titles put massive pressure on the memory subsystem. Still, clock for clock, AMD is way behind here.

It's still early but what a potential train wreck in the making.

The jury is out for me on the merger, with regard to its impact on AMD. The merger got AMD into the platform business, and we'll see how that works over the next several years. From a neutral standpoint, I'd like to see them survive and thrive. Intel needs a long-term competitor, as does nVidia.

I just hope something, somewhere begins to put downward pricing pressure on lower high-end parts. With this revelation, I can't imagine any new, lower-high end part like an 8800gs falling in price very quickly, which bums me out. Not to mention that this revelation bodes poorly for the true midrange. The 8600 series disappointed me and I can't imagine that the 2600 will fare any better after seeing the 2900XT/XTX benchmarks.

DX10, as always, is still the wildcard.

 

[LittleNemoNES]

sad ass day...


Oh well, looks like I'll be upgrading sooner than I thought
Selling my x1900 cards! Who wants 'em?

 

[n7]

This is like a nightmare come true.

I didn't expect the HD 2900 XTX to be much faster than the 8800 GTX, but i thought it'd be comparable at least.

It's very odd how more vRAM & higher bandwidth does essentially nothing for the card's performance.
I suspect the poor performance is due to the stream processors layout for the AMD cards vs. nV's, but i can't find anything stating just how it's set up.

Even though i'm extreme disappointed, i still really want to see these cards hits the stores, preferably with nice low prices.
We need even this crappy competition to get the pricing on the 8800s lower, & future 8900s.

Hey when do the 8900s come out anyway?

 

[Creig]

Originally posted by: dreddfunk
It certainly looks like the highest 2900 is going to be somewhere between underwhelming and down right horrific: late, slower and less power-efficient than its competitor. What does it say when the best case scenario is repeat of the x1800xt? Underwhelming at the time, hidden value later. The only thing that I can imagine saving the 2900 series is if DX10 titles put massive pressure on the memory subsystem. Still, clock for clock, AMD is way behind here.

It's still early but what a potential train wreck in the making.

The jury is out for me on the merger, with regard to its impact on AMD. The merger got AMD into the platform business, and we'll see how that works over the next several years. From a neutral standpoint, I'd like to see them survive and thrive. Intel needs a long-term competitor, as does nVidia.

I just hope something, somewhere begins to put downward pricing pressure on lower high-end parts. With this revelation, I can't imagine any new, lower-high end part like an 8800gs falling in price very quickly, which bums me out. Not to mention that this revelation bodes poorly for the true midrange. The 8600 series disappointed me and I can't imagine that the 2600 will fare any better after seeing the 2900XT/XTX benchmarks.

DX10, as always, is still the wildcard.

Ditto

 

[allies]

Wow. I'm quite glad I picked up my 8800GTX for $500 in December and didn't wait for the R600.

R600 = NV30?

AMD needs to rebound... ever since the 9700Pro ATi has been the performance leader when their newest card shipped. This is uncharacteristic. Nvidia threw them a HUGE curveball with G80.

I just hope that AMD has something special for R700... I just don't understand how how a foundation like Xenos couldn't be perfected for PC in a couple years.

Wow

 

[fierydemise]

I think in this case we really have to wait for more complete benchmarks, the problem is the XTX has ultra high memory bandwidth and it'll probably take higher res and higher AA (4xAA isn't really that high) for it to really show its power. If you look at the numbers its obvious the XTX hasn't been putting that bandwidth to any use because its performing near identically to the XT, I expect at higher resolutions and/or higher AA the XTX's curve (from AT's review) will look relatively flat while the XT will start to drop off as will the GTX. I think the XTX will really be the card for the 30" monitors and thats about it, any lower and the card really won't be able to show the power of its high memory bandwidth. The only real reason for the XTX is to have a flagship card, I think its completely unnecessary because the XT is a faster card then the GT and with the XT apparently have a lot of OCing headroom. I expect the R600 generation to be one of the best generations for AMD, nVidia left the door wide open on the low end by only selling to the OEMs and the midrange cards aren't anything special.

 

[n7]

I personally will need to see Vista benches before i buy anything as well, since i'm not running XP these days.

It does look like an 8800 GTX/8900 series card is on my horizon though.

Now get these out @ dirt cheap pricing AMD!

 

[apoppin]

Originally posted by: dreddfunk
It certainly looks like the highest 2900 is going to be somewhere between underwhelming and down right horrific: late, slower and less power-efficient than its competitor. What does it say when the best case scenario is repeat of the x1800xt? Underwhelming at the time, hidden value later. The only thing that I can imagine saving the 2900 series is if DX10 titles put massive pressure on the memory subsystem. Still, clock for clock, AMD is way behind here.

It's still early but what a potential train wreck in the making.

The jury is out for me on the merger, with regard to its impact on AMD. The merger got AMD into the platform business, and we'll see how that works over the next several years. From a neutral standpoint, I'd like to see them survive and thrive. Intel needs a long-term competitor, as does nVidia.

I just hope something, somewhere begins to put downward pricing pressure on lower high-end parts. With this revelation, I can't imagine any new, lower-high end part like an 8800gs falling in price very quickly, which bums me out. Not to mention that this revelation bodes poorly for the true midrange. The 8600 series disappointed me and I can't imagine that the 2600 will fare any better after seeing the 2900XT/XTX benchmarks.

DX10, as always, is still the wildcard.

maybe AMD should bring Jerry Sanders back

http://www.eetimes.com/blog/news/archiv...QSNDLPCKH0CJUNN2JVN?loc=semiconductors

Maybe Jerry Sanders should return to AMD

1. AMD is floating some $2.2 billion in notes. My assessment: AMD is highly leveraged and may become a house of cards. Notes are a good short-term solution, but AMD really needs to develop a new and innovative processor line--again. Opteron was a winner. But the company has been resting on its laurels since the intro of Opteron. AMD's next-gen processor, dubbed Barcelona, just doesn't excite me. A real snoozer. Come on AMD! You can do better than that!

2. To keep up with Intel, AMD has got major fab projects in Germany, India and New York. My assessment: AMD still acts like an IDM. Let's get real and stop pretending. How is AMD going to fund these fab projects? They should drop fab plans in India and New York. Focus on the fab in Germany. Outsource more chips to Chartered, IBM and Samsung.

3. Some say AMD should go the private-equity route. My assessment: That's the company's only route right now. I hate private equity, but AMD won't last the year unless Silver Lake or Francisco Partners jumps in to save them.

4. Perhaps AMD should get acquired by IBM or a member of IBM's ''fab club.'' My assessment: IBM is a bad choice. IBM is a great IP firm, but it never really understood the semi business. Here's a crazy idea: Samsung wants to go toe-to-toe against Intel. How about Samsung buying AMD? However, I'm afraid Intel would kick Samsung's ass in processors. Samsung is a mere one-hit commodity wonder.

5. One more crazy idea: Maybe Jerry Sanders could return to AMD. My assessment: He couldn't do any worse than AMD's current management, which squandered its technology lead over Intel. What a waste.

i am truly SAD to say it ... it also looks like my "rants" were right

... over the past 6 long months i have been saying the r600 is a *flop* and AMD is *doomed*

the next *disaster* i expect will be Barcelona
-they they can stick a fork in it

toldjaso

 

[Aikouka]

Paying $650 when the 8800GTX came out doesn't sound so horrible anymore .

It's kind of like buying a car now and then seeing the new models that will come out.. then being relieved when they're really not better than what you have.

 

[terentenet]

Originally posted by: fierydemise
I think in this case we really have to wait for more complete benchmarks, the problem is the XTX has ultra high memory bandwidth and it'll probably take higher res and higher AA (4xAA isn't really that high) for it to really show its power. If you look at the numbers its obvious the XTX hasn't been putting that bandwidth to any use because its performing near identically to the XT, I expect at higher resolutions and/or higher AA the XTX's curve (from AT's review) will look relatively flat while the XT will start to drop off as will the GTX. I think the XTX will really be the card for the 30" monitors and thats about it, any lower and the card really won't be able to show the power of its high memory bandwidth. The only real reason for the XTX is to have a flagship card, I think its completely unnecessary because the XT is a faster card then the GT and with the XT apparently have a lot of OCing headroom. I expect the R600 generation to be one of the best generations for AMD, nVidia left the door wide open on the low end by only selling to the OEMs and the midrange cards aren't anything special.

My oppinion stands *ONLY* if the DT benches are the real thing and really reflect the HD2900XTX performance. Higher res and AA/AF will not help.
Look at the performance drop going from 1280x1024 to 1920x1200. If the XTX indeed had the power and memory bandwidth helped it, the drop wouldn't be that big and the FPS drop line should be flat..
But it drops.
Do you really think something different is going to happen at 2560x1600 16AA/16AF?
All of a sudden 8800GTX will drop from 98FPS at 1920x1200 to 21FPS at 2560x1600 and the XTX will drop only at 35FPS? I highly doubt it.

If the benches are true, AMD/ATI's problem is not the memory bandwidth, but core architecture. I remember ATI posted some slides showing how technologically advanced their shaders architecture will be in R600 compared to G80's.
But it's 128 shaders vs. 320 and we get this? Trully dissapointing.

I really love my 8800GTX SLI and I was Nvidia all the way. Though, I was hoping for a better ATI to force Nvidia release the 8900GTX faster.

 

[allies]

Originally posted by: josh6079
God, 2007 has really sucked for us gamers so far.

QFMFT.

 

[5150Joker]

AMD needs to get rid of Hector and his useless lackeys and instead bring back Atiq Raza if he's remotely interested. AMD flourished under his control with the purchase of NexGen, maybe he could turn AMD/ATi around.

 

[nZone]

The review make sense because the only thing differentiate between XT and XTX is XT = GDR3; XTX=GDR4. Both uses the exact same core.

In contrast; the 8800GTS/GTX the difference is:

Yes, since GTS has some units disabled.

Compared to GTS, GTX has:
- 50% more shader power ((1350x128)/(1200x96)=1.5)
- 38% more ROP power ((575x6)/(500x5) = 1.38)
- 53% more TMU power ((575x128)/(500x96) = 1.53)
- 35% more memory bandwidth (86.4/64 = 1.35)

 

[SexyK]

Man, ever since the failed launch 6 weeks back I've been saying that there must have been major performance problems or a showstopping bug, and I was shouted down every time. "They're just waiting for a top-to-bottom release" was thrown around so many times it's unbelievable. Now not only was I correct about the performance issues, but it appears that there still WON'T be a top-to-bottom hard launch anytime soon. Don't see how this launch can be any more of a failure. Even if they make a lot of XT sales with low-ball pricing, that will still be murder on their bottom line. It's a disappointing day, I truly loved my 9500 Pro back in the day. On the bright side I've been playing every game maxed out on my 8800 for over 5 months now, so it's not all that bad!

 

[apoppin]

Originally posted by: 5150Joker
AMD needs to get rid of Hector and his useless lackeys and instead bring back Atiq Raza if he's remotely interested. AMD flourished under his control with the purchase of NexGen, maybe he could turn AMD/ATi around.

AMD need someone with a plan and lots of money to buy them out ... now


before barcelona flops too

Jerry Sanders would do much better then their current arrogant clown show


http://www.eetimes.com/blog/news/archiv...QSNDLPCKH0CJUNN2JVN?loc=semiconductors

 

[otispunkmeyer]

Originally posted by: jim1976

Originally posted by: sisq0kidd
I don't believe Dailytech is lying at all or being disingenuous with their benchmarks in the slightest.

They benched what they received. It's a driver issue on ATI's part... I hope

I used to respect them but not anymore..But that's just me I suppose..
Also I find weird that in 3dmark scores the X2900XT was dominating and in games falls so much behind.. No matter how much I don't trust 3DMock as an indication of real gaming performance it just doesn't add up.. Nm I'll wait for some more previews/reviews b4 I make my opinion about X2900..

well the 8600GTS did the same....shone in 3D 06, tanked in real life, i thought that was because it had decent shader grunt, but no good memory bandwidth for todays games which use heavy textures. but the XT hasnt got that excuse really has it. 3d 06 really is just for braggers to get the high score.

 

[Mem]

Originally posted by: apoppin

Originally posted by: dreddfunk
It certainly looks like the highest 2900 is going to be somewhere between underwhelming and down right horrific: late, slower and less power-efficient than its competitor. What does it say when the best case scenario is repeat of the x1800xt? Underwhelming at the time, hidden value later. The only thing that I can imagine saving the 2900 series is if DX10 titles put massive pressure on the memory subsystem. Still, clock for clock, AMD is way behind here.

It's still early but what a potential train wreck in the making.

The jury is out for me on the merger, with regard to its impact on AMD. The merger got AMD into the platform business, and we'll see how that works over the next several years. From a neutral standpoint, I'd like to see them survive and thrive. Intel needs a long-term competitor, as does nVidia.

I just hope something, somewhere begins to put downward pricing pressure on lower high-end parts. With this revelation, I can't imagine any new, lower-high end part like an 8800gs falling in price very quickly, which bums me out. Not to mention that this revelation bodes poorly for the true midrange. The 8600 series disappointed me and I can't imagine that the 2600 will fare any better after seeing the 2900XT/XTX benchmarks.

DX10, as always, is still the wildcard.

maybe AMD should bring Jerry Sanders back

http://www.eetimes.com/blog/news/archiv...QSNDLPCKH0CJUNN2JVN?loc=semiconductors

Maybe Jerry Sanders should return to AMD

1. AMD is floating some $2.2 billion in notes. My assessment: AMD is highly leveraged and may become a house of cards. Notes are a good short-term solution, but AMD really needs to develop a new and innovative processor line--again. Opteron was a winner. But the company has been resting on its laurels since the intro of Opteron. AMD's next-gen processor, dubbed Barcelona, just doesn't excite me. A real snoozer. Come on AMD! You can do better than that!

2. To keep up with Intel, AMD has got major fab projects in Germany, India and New York. My assessment: AMD still acts like an IDM. Let's get real and stop pretending. How is AMD going to fund these fab projects? They should drop fab plans in India and New York. Focus on the fab in Germany. Outsource more chips to Chartered, IBM and Samsung.

3. Some say AMD should go the private-equity route. My assessment: That's the company's only route right now. I hate private equity, but AMD won't last the year unless Silver Lake or Francisco Partners jumps in to save them.

4. Perhaps AMD should get acquired by IBM or a member of IBM's ''fab club.'' My assessment: IBM is a bad choice. IBM is a great IP firm, but it never really understood the semi business. Here's a crazy idea: Samsung wants to go toe-to-toe against Intel. How about Samsung buying AMD? However, I'm afraid Intel would kick Samsung's ass in processors. Samsung is a mere one-hit commodity wonder.

5. One more crazy idea: Maybe Jerry Sanders could return to AMD. My assessment: He couldn't do any worse than AMD's current management, which squandered its technology lead over Intel. What a waste.

i am truly SAD to say it ... it also looks like my "rants" were right

... over the past 6 long months i have been saying the r600 is a *flop* and AMD is *doomed*

the next *disaster* i expect will be Barcelona
-they they can stick a fork in it

toldjaso

I hope not,AMD is the only real competitor to Intel and Nvidia.I just hope they get the pricing right(as well as performance) on the rest of their new range next month,guess only time will tell,as to Barcelona, think its too early to speculate on that one.

 

[happy medium]

If Nvidia can get that 8800gs out and fill the gap between the 8600gts and 8800gts 320mb there lineup will be solid. I really think ATI got caught with there pants down. They are ready to launch the 8800 ultra which completes the domination. Sad day for competition.