Radeon R9 290X Priced at $730 at Newegg

[Pariah]

Developers can be "incentivized" and AMD isn't shy about throwing money around.

Are you a developer with first hand knowledge of this?

The first red flag with Mantle is that it isn't available at launch. The "plan" is a patch two months after launch for BF4. How long has it been since AMD released a new flagship card? Almost 2 years? And they still aren't able to have one of the keynote features ready at launch? This would be less of a concern if AMD had a track record of timely excellent driver development. But even the most hardcore AMD fan has to acknowledge, that is certainly not the case.

 

[Vesku]

Programming frameworks aren't just assembled during a weekend coding binge. Takes quite a while for things to get rolling.

 

[Pariah]

Programming frameworks aren't just assembled during a weekend coding binge. Takes quite a while for things to get rolling.

I didn't know John Carmack posted on these forums. With such incredibly insightful posts like this, you have to be him.

 

[SiliconWars]

It's called common sense. There are no "red flags" unless you are trying desperately to find them. AMD has stated on plenty of occasions that Mantle is in it's very early stages.

 

[blackened23]

I still don't see why Mantle is being thrown around as a selling point. When BF4 Mantle is released, then it can be considered a selling point. When Mantle is in UE4 or CryEngine3, then it is a selling point. Right now? It's a point of interest and has potential but is not a reason for any sane person to buy a video card. People don't buy products based on vaporware. When the benchmarks are out and the results are quantifiable, then AMD can brag about it all they want.

That said, I think the concept behind Mantle is great and long overdue - but let's not kid ourselves, when 290X reviews hit on Oct. 15th they will look at one thing and one thing alone - the 290X is going to be judged by DX11 and that's IT. Not Mantle; Mantle is a non issue until December at the earliest. Mantle certainly does not warrant a 700$ price tag on the 290X, if that happens I guarantee that a large percentage of potential buyers will just laugh and buy a pre-overclocked GTX 780 (which already matches Titan) instead.

 

[raghu78]

It's all irrelevant if developers don't use it. If Mantle is used extensively for console development, and porting to PC is simple, it has a chance. Otherwise, it is totally dead in the water if it has to stand on its own in the PC market. Developers code to the largest customer pool, not the highest performing pool. Given the choice of developing for just DirectX which all PC's can use, or DirectX AND Mantle, which is AMD only right now even if it is significantly faster, which will developers choose?

Not every developer goes by numbers. Few are always striving to improve their work and also lead a path which others can follow. Mantle was spearheaded by DICE's Johan Andersson.

http://www.hardware.fr/focus/89/amd-mantle-interview-raja-koduri.html

"A game engine like Frostbite being able to deliver more performance, more draw calls, more things… their view is that it's a competitor edge for the game engine. The view that it's a competitor edge for AMD is a side effect. It's really about the competitor edge for the game engine that it can do more on this hardware."

In the gaming industry not everyone develops their own game engines as these are incredibly complex. there are few game engines which are widely used. The engine licensing revenue for these companies is a lucrative revenue stream.

DICE - Frostbite 3
EPIC - Unreal Engine 4
VALVE - Source
CRYTEK - Cryengine 3

Once a game engine developer integrates support for Mantle all games developed using that engine benefit. So all EA games benefit from Mantle as they all are using Frostbite 3. You can bet AMD has been courting the top game engine studios. expect to hear more at AMD Fusion developer summit in Nov.

Also you are missing the key factor here. Starting from Temash , Kabini and the upcoming Kaveri every AMD APU is GCN based. whats more even an ARM v8 based tablet SOC with GCN graphics will benefit from Mantle because its cross platform.

http://www.xbitlabs.com/news/cpu/di...on_Processor_for_Consumer_Devices_Report.html

AMD can even pay to get these key developers to support Mantle in their game engines because that would mean a wide range of games are covered and it benefits the entire AMD product stack starting from tablet APUs to desktop APUs to high end GPUs. an AMD Kaveri APU with Mantle in BF4 or an upcoming Unreal Engine 4 based game. now thats what we are talking

as far as DICE is concerned Mantle is being used as an game engine differentiation strategy. so other studios would also want to have a Mantle renderer to better compete with Frostbite 3 for engine licensing deals.

 

[96Firebird]

That said, I think the concept behind Mantle is great and long overdue - but let's not kid ourselves, when 290X reviews hit on Oct. 15th they will look at one thing and one thing alone - the 290X is going to be judged by DX11 and that's IT. Not Mantle; Mantle is a non issue until December at the earliest.

Exactly. But I foresee many posters will chime in with "Just wait until Mantle gets here!" when the reviews are posted. Just seems like more wait and see...

 

[AnandThenMan]

I agree with blackened23. We will (rightly so) judge the R9 xxxx cards by how they perform in games we can buy now, or when the cards launch. Not for Mantle enhanced games that hit down the road. BUT, given the potential of Mantle, at least for me it makes the cards much more intriguing.

Let's face it, DX gaming has become downright stagnant sans a few titles every year, at least with Mantle we might see a shake up. Plus the new consoles are most certainly going to bring much better games to the PC side, I have no doubt about that.

 

[thilanliyan]

The first red flag with Mantle is that it isn't available at launch. The "plan" is a patch two months after launch for BF4. How long has it been since AMD released a new flagship card? Almost 2 years? And they still aren't able to have one of the keynote features ready at launch? This would be less of a concern if AMD had a track record of timely excellent driver development. But even the most hardcore AMD fan has to acknowledge, that is certainly not the case.

This is not a feature that AMD ALONE can implement. The dev has to do a major part of it, and the game isn't even released yet. So how is Mantle not launching immediately down to AMD's 'incompetence'?

 

[Cloudfire777]

Are you a developer with first hand knowledge of this?

The first red flag with Mantle is that it isn't available at launch. The "plan" is a patch two months after launch for BF4. How long has it been since AMD released a new flagship card? Almost 2 years? And they still aren't able to have one of the keynote features ready at launch? This would be less of a concern if AMD had a track record of timely excellent driver development. But even the most hardcore AMD fan has to acknowledge, that is certainly not the case.

Yeah, I think the size of their software team is pretty small, looking back at drivers, crossfire issues and latency problems that took forever to improve.

DX was an universal API used by everyone and have been for a very long time now which I`m sure meant big improvements over a short period of time. Mantle is AMD exclusive, I wonder how this will work out for AMD.

 

[Fallen Kell]

Excellent. Looks like AMD is about to cornhole nvidia with a Titan killer.

You assume that it actually costs Nvidia >$700 to make a Titan. If the real costs are only $400, they can easily price drop to $600 and still be laughing all the way to the bank...

 

[AnandThenMan]

Exactly. But I foresee many posters will chime in with "Just wait until Mantle gets here!" when the reviews are posted. Just seems like more wait and see...

Not any different than people saying, "just wait until 20nm and Maxwell is here!". Wait and see is a perfectly reasonable thing to do, if you don't mind waiting.

 

[smackababy]

I still don't see why Mantle is being thrown around as a selling point. When BF4 Mantle is released, then it can be considered a selling point. When Mantle is in UE4 or CryEngine3, then it is a selling point. Right now? It's a point of interest and has potential but is not a reason for any sane person to buy a video card. People don't buy products based on vaporware. When the benchmarks are out and the results are quantifiable, then AMD can brag about it all they want.

That said, I think the concept behind Mantle is great and long overdue - but let's not kid ourselves, when 290X reviews hit on Oct. 15th they will look at one thing and one thing alone - the 290X is going to be judged by DX11 and that's IT. Not Mantle; Mantle is a non issue until December at the earliest. Mantle certainly does not warrant a 700$ price tag on the 290X, if that happens I guarantee that a large percentage of potential buyers will just laugh and buy a pre-overclocked GTX 780 (which already matches Titan) instead.

I think the problem with Mantle being a selling point is right after the Mantle slide, you had people in this section claiming Mantle was going to make 7870s get 120 FPS in games Titans are getting 60 FPS.

I, personally, am not sold Mantle will bring any meaningful performance increases and with no real display of the API, feel it will end up being very cumbersome for developers to use. The exact opposite is happening in software development than what people are saying developers in gaming want. Hibernate is a perfect example of this. I could write faster and more efficient code to map objects to my database, but at the expense of time and massive amounts of boilerplate code. Plus, it takes a highly skilled developer to do what it takes a novice to do with Hibernate. That is what the real downfall of Mantle I feel is.

I could be wrong. The Mantle API could be amazing and easy to use. I would be more than happy to be wrong.

 

[SiliconWars]

I still don't see why Mantle is being thrown around as a selling point. When BF4 Mantle is released, then it can be considered a selling point. When Mantle is in UE4 or CryEngine3, then it is a selling point. Right now? It's a point of interest and has potential but is not a reason for any sane person to buy a video card. People don't buy products based on vaporware. When the benchmarks are out and the results are quantifiable, then AMD can brag about it all they want.

Anything that has the potential to give an edge - especially a big edge - is a selling point. If I were the staunchest Nvidia fanboy and looking at buying a >$500 graphics card right now, I'd be extremely concerned about it being obsoleted by Mantle. Nobody wants to buy a graphics card at that price and see it being taken apart by the competitors offering.

That said, I think the concept behind Mantle is great and long overdue - but let's not kid ourselves, when 290X reviews hit on Oct. 15th they will look at one thing and one thing alone - the 290X is going to be judged by DX11 and that's IT. Not Mantle; Mantle is a non issue until December at the earliest. Mantle certainly does not warrant a 700$ price tag on the 290X, if that happens I guarantee that a large percentage of potential buyers will just laugh and buy a pre-overclocked GTX 780 (which already matches Titan) instead.

Suffice it to say I disagree entirely. Mantle is big, it's big news and it has struck fear into the heart of Nvidia. They haven't even responded yet. Nobody who has heard of Mantle would buy a 780 at the same price right now - they would at least wait until the BF4 Mantle benchmark instead. Anyone who knows about Mantle and decides to buy a 780 right now only to find their card is "ridiculed" in BF4, and many future Frostbite titles gets what they deserve.

 

[raghu78]

This is not a feature that AMD ALONE can implement. The dev has to do a major part of it, and the game isn't even released yet. So how is Mantle not launching immediately down to AMD's 'incompetence'?

exactly. People are still thinking that Mantle is something like AMD's CUDA. If anything Mantle is Johan and DICE's brainchild. Mantle is a low level API for the PC developed by DICE in close collaboration with AMD. It has been in development for 2+ years. btw Mantle is a completely new API and not just a driver. The biggest proof of concept of such an API is a state of the art game engine like Frostbite 3 running Mantle and showing the performance benefits against DX11.1 . Now that too is in DICE's hands and not in AMD's. AMD are merely facilitating a process driven by Johan and DICE. Other developers can take DICE's lead and implement their own Mantle renderer.

 

[smackababy]

Suffice it to say I disagree entirely. Mantle is big, it's big news and it has struck fear into the heart of Nvidia. They haven't even responded yet..

They responded before Mantle was announced. They said DX was just fine and direct hardware access wouldn't offer any significant improvements. At the time, I thought they were talking about consoles, but a few days later I realized they meant Mantle.

 

[Elfear]

Really? We both know there is a history of sometimes lengthy waits for Crossfire profiles to be enabled on even big name titles. More common erratic performance behavior than SLI. In some cases they will even break crossfire in titles which previously worked fine. (<---This one was the hair that broke the camels back for me in fact, they managed to stuff up COD multiplayer....how do you mess up a COD title and not fix it immediately, months this went on).

SLI has a lot of the same issues though.

7970 vs 680 scaling almost exactly the same

6950 scales better than 570

7970 vs 680 - 7970 scales better overall but has worse issues in Batman: Arkham City


With regards to profiles, both companies could be a little quicker getting them out.

No Rome Total War II SLI profile

Deus Ex - Human Revolution missing SLI profile

Dirt Showdown

Dragon Age 2 took months for an SLI profile

AMD has their problems too but it isn't like their is a major difference between the two multi-GPU solutions except for the current frame latency issues on DX9 and multi-monitor resolutions.

Anyway, sorry for the OT. Can't wait to get real pricing info on AMD's new lineup. I really hope the $699 is for the BF4 edition cards and regular 290X versions will be $549-599.

 

[Elfear]

I still don't see why Mantle is being thrown around as a selling point. When BF4 Mantle is released, then it can be considered a selling point. When Mantle is in UE4 or CryEngine3, then it is a selling point. Right now? It's a point of interest and has potential but is not a reason for any sane person to buy a video card. People don't buy products based on vaporware. When the benchmarks are out and the results are quantifiable, then AMD can brag about it all they want.

That said, I think the concept behind Mantle is great and long overdue - but let's not kid ourselves, when 290X reviews hit on Oct. 15th they will look at one thing and one thing alone - the 290X is going to be judged by DX11 and that's IT. Not Mantle; Mantle is a non issue until December at the earliest. Mantle certainly does not warrant a 700$ price tag on the 290X, if that happens I guarantee that a large percentage of potential buyers will just laugh and buy a pre-overclocked GTX 780 (which already matches Titan) instead.

+1

I'm excited about Mantle's potential but it's not much of a selling point until we see some actual results.

 

[RussianSensation]

Final specs?

 

[Gikaseixas]

Final specs?

Noticed that memory bandwidth is all over the place, i've seen 280 and 300 being thrown around

 

[Erenhardt]

+1

I'm excited about Mantle's potential but it's not much of a selling point until we see some actual results.

True that, but to be fair one should partially ignore PhysX as a selling point. You never know of there will be more games with it in the feature.

 

[AnandThenMan]

Interesting to see under 3D Technology: DirectX 11.2, OpenGL 4.3, Mantle. I mean I know it's there, but to see a new supported API listed after all this time, jarring.

 

[Face2Face]

True that, but to be fair one should partially ignore PhysX as a selling point. You never know of there will be more games with it in the feature.

But PhysX won't add performance, just more particles and stuff falling on the ground and such... ()

 

[raghu78]

I still don't see why Mantle is being thrown around as a selling point. When BF4 Mantle is released, then it can be considered a selling point. When Mantle is in UE4 or CryEngine3, then it is a selling point. Right now? It's a point of interest and has potential but is not a reason for any sane person to buy a video card. People don't buy products based on vaporware. When the benchmarks are out and the results are quantifiable, then AMD can brag about it all they want.

Yeah you are right. Johan spent 2 years leading the Mantle project for an insignificant performance improvement. keep believing what you want to :biggrin:

but let's not kid ourselves, when 290X reviews hit on Oct. 15th they will look at one thing and one thing alone - the 290X is going to be judged by DX11 and that's IT. Not Mantle; Mantle is a non issue until December at the earliest. Mantle certainly does not warrant a 700$ price tag on the 290X, if that happens I guarantee that a large percentage of potential buyers will just laugh and buy a pre-overclocked GTX 780 (which already matches Titan) instead.

R9 290X vs GTX 780 OC at stock and voltage overclocked. You say this is the key metric. I say this one is going to be close with no decisive winner. This is not like GTX 580 vs HD 6970 where there was one decisive winner let see how it turns out. now if thats a tie then it comes to the other factors.

1. 4GB VRAM on R9 290X useful in multi monitor and 4K displays in the latest games.
2. Mantle with BF4 being the first showcase in Dec. all future EA titles are guaranteed to run Mantle because Frostbite 3 supports it.
3. A very good possibility that other top game engines also are likely to follow soon with their Mantle implementations.

its a matter of being able to see Mantle in action and then the floodgates will open. You can bet AMD is courting Epic, Crytek and Valve to ready Mantle versions of their engines. Also with DICE leading the way these studios would not want to miss out on an important comparison feature for their engines on the PC.

 

[imaheadcase]

They responded before Mantle was announced. They said DX was just fine and direct hardware access wouldn't offer any significant improvements. At the time, I thought they were talking about consoles, but a few days later I realized they meant Mantle.

Not only that, but Nvidia has many things up its sleeves. Nvidia does not worry about mantle at all.

1. It can easily lower prices on existing cards ore release OC versions of existing one.
2. Devs have to support it to begin with. Keep in mind a dev can say they will support it, but how easy that is for them is another story. Remember Frostbite was built with it in mind, existing ones might have a harder time before games come out using it.
3. Nvidia has its own API already.
4. The advantage is still not known, and could vary by game.