Vega/Navi Rumors (Updated)

[AtenRa]

(The same goes for the 980 and how people moan how the 390X is better at DX12. The 390X is a little big larger and uses significantly more power - if it isn't stronger than the 980 then the architecture has problems).

Kepler GK110 (GTX 780) = 561mm2 launched in May 2013
Hawaii (R9 290X) = 438mm2 launched in October 2013

Hawai today is faster (438mm2 vs 561mm2)

Maxwell GM204 (GTX 980) = 398mm2 launched September 2014

Hawaii (390X) and Maswell GM204 (GTX 980) are almost equals today

Hawaii was an exceptional design back in 2013, it launched 5 months after big Kepler with a way smaller die and today is even faster than GTX 780Ti (Launched November 2013) especially in DX-12/Vulkan games.

GM204 Maxwell GTX 980 came almost one year after Hawaii, its not unheard of a new design to outperform an 11 months old design.

 

[zinfamous]

It is a shame that your post will not get the "like" action that it so desperately deserves. So you get a like and a "QFT."

QFT.

Well, I like his post and will add to the QFT (I haven't been much arguing about AMD doing well for themselves--I've long conceded that they still have major problems going forward with these gambles.)

I just don't "like button" posts on moral principal.

 

[railven]

I just don't "like button" posts on moral principal.

I'd agree with this. And frankly it makes it easier to see who are in what camp and who echos who. I'll stick to my "kudos" response when someone nails something (or makes a funny post! )

(That and I'm an anti-social fluff-n-nutter

 

[Arachnotronic]

Well, I like his post and will add to the QFT (I haven't been much arguing about AMD doing well for themselves--I've long conceded that they still have major problems going forward with these gambles.)

I just don't "like button" posts on moral principal.

I kind of figured, considering that you will cut those who like them! ;-)

Actually, crap. I think I actually liked a post of yours a little while back.

*hides*

 

[Arachnotronic]

Kepler GK110 (GTX 780) = 561mm2 launched in May 2013
Hawaii (R9 290X) = 438mm2 launched in October 2013

Hawai today is faster (438mm2 vs 561mm2)

Maxwell GM204 (GTX 980) = 398mm2 launched September 2014

Hawaii (390X) and Maswell GM204 (GTX 980) are almost equals today

Hawaii was an exceptional design back in 2013, it launched 5 months after big Kepler with a way smaller die and today is even faster than GTX 780Ti (Launched November 2013) especially in DX-12/Vulkan games.

GM204 Maxwell GTX 980 came almost one year after Hawaii, its not unheard of a new design to outperform an 11 months old design.

Keeping GK110 around for as long as NVIDIA did was less than ideal and risky. They needed a Maxwell GM104 out in late 2013/early 2014, but for whatever reason that just didn't happen. NVIDIA was fortunate that AMD shot itself in the foot with Hawaii's reference cooling solution and that it was able to charge what it did for the 780 Ti (since full GK110 offered leadership performance at the time) otherwise that cycle could have been a real financial issue for them.

 

[railven]

Kepler GK110 (GTX 780) = 561mm2 launched in May 2013
Hawaii (R9 290X) = 438mm2 launched in October 2013

Hawai today is faster (438mm2 vs 561mm2)

Maxwell GM204 (GTX 980) = 398mm2 launched September 2014

Hawaii (390X) and Maswell GM204 (GTX 980) are almost equals today

Hawaii was an exceptional design back in 2013, it launched 5 months after big Kepler with a way smaller die and today is even faster than GTX 780Ti (Launched November 2013) especially in DX-12/Vulkan games.

GM204 Maxwell GTX 980 came almost one year after Hawaii, its not unheard of a new design to outperform an 11 months old design.

I feel my problem with these kind of posts is they always seem to focus on now, when those chips being discussed are no longer relative.

When it mattered, ie AMD was selling 7970's for $550 and prone to make good return per each sold, this place admonished AMD for raising the MSRP. It's like AMD committed some unconscionable wrong to try to make money. Fast forward to Nano and this place bent overward backwards to justify the price. I just don't get it anymore but I think by then it was more a "everyone is a stock holder/miner" COI I personally think people adhere to now.

Hawaii can be 100x faster than Kepler/Maxwell today, it doesn't matter when today's product stack don't favor AMD outside of select few titles. And in 2019 when Polaris 10 is destroying GP106 in everything it won't matter either if GV104 is destroying whatever AMD has on the platter.

Is AMD awesome? They create some awesome tech, but by god you'd figure a company leading the industry (sometimes by the hand) can make some money. And before anyone goes "derp are you a consumer or stock holder" I've always been a consumer, but I'm now buying NV because AMD has nothing to offer me. And by the time they do I have zero doubt NV will be ready to respond. Same with DX12. AMD can be as fast as they want in the 5-6 DX12 titles, when we're looking at 20-30 DX titles, something tells me NV will be more than ready to laugh again.

But, at least Tahiti is now faster than Kepler. That's all that matters.

EDIT: oops confused some GPU names, fixed em all - I hope.

 

[railven]

Keeping GK110 around for as long as NVIDIA did was a complete fail. They needed a Maxwell GM104 out in late 2013/early 2014, but for whatever reason that just didn't happen. NVIDIA was fortunate that AMD shot itself in the foot with Hawaii's reference cooling solution and that it was able to charge what it did for the 780 Ti, otherwise that cycle could have been a real financial issue for them.

Wasn't following NV back then as much, but that's what I saw from the AMD side. Here AMD brings out a damn good card, sticks that reference cooler from like HD 5k days - why? Okay, let's wait for custom cards, welps those won't be on the market for another 3-4 months. And while I can't entirely blame them for this, but then come the miners.

So a great product is now being gouged and when miners are like "LOL we just downvolt and who cares if I got that blower in the basement" stores are like "hey we can sell even the ref models for $200+, why not?"

/facepalm

So I got a GTX 780. And "LOL NV Fanboy!"

2x /facepalm

 

[zinfamous]

I kind of figured, considering that you will cut those who like them! ;-)

Actually, crap. I think I actually liked a post of yours a little while back.

*hides*

yeah, I'm just not a fan of the concept. Anything Facebooky is anti-social, ime.

 

[Bacon1]

Being competitive on the low end (and losing money doing so) does not imply that it is possible to be competitive on the high end.

They aren't losing money on Polaris. It's exactly where they said it was going to be priced back in January.

AMD is flat out noncompetitive in notebooks and has been since Maxwell launched because of the lack of technical merit in the architecture. Pasal in notebooks roflstomped everything on the market. This is a big problem, a serious problem for the bottom line of the company, and all this forum seems to want to do is go " power consumption doesn't matter".

http://www.pcworld.com/article/3115...-470-inside-for-amd-polaris-mobile-debut.html

Pretty sure they also got the Apple desktop and laptop as well for its refresh coming up.

This forum is simply not realizing that being "cheaper and faster than the competition" is not an end goal that a company strives towards. Companies strive towards making profits and AMD simply hasn't been making money (sufficiently in the black) for years. AMD selling you a "cheaper and faster card" is not something AMD wants to do, it is something they must do or they go out of business. In other words, the products are just NOT good enough to sell at anything resembling their desired margins.

Sure companies would always like to make more, but they've stated that they wanted to bring Polaris and newer cards for cheaper and that it cost them much less and they got more dies because of the node shrink. Thus even though they were selling for less, they cost a lot less which doesn't mean they are skimping themselves.

Nvidia making hand over fist in profit margins isn't a good thing. It just shows how over priced their goods are.

Obviously its good for companies to make profits, but its not anywhere near what your underlined statement would show. They set out to sell Polaris for much cheaper because their goal was "VR for the masses". Meaning you had to take the previous $400 price point down into the $200s. They've talked about this for months before release, no idea why people assumed they wanted to sell Polaris for more.

 

[Bacon1]

AMD was selling 7970's (Fiji)

Fiji is Fury. 7970 is tahiti

 

[railven]

Fiji is Fury. 7970 is tahiti

D'oh! What I get for trying to post while at work. Haha.

I hope you all internally changed it to Tahiti. Welps now to re-edit that post.

 

[nurturedhate]

Hmm, 7 pages in. 5 of those about how AMD is terrible and can't compete, all from the same people like every other thread. Great job guys! Mission accomplished. It is truly amazing how every thread needs to be turned into an Nividia is great and AMD sucks propaganda piece.

 

[Enigmoid]

They aren't losing money on Polaris. It's exactly where they said it was going to be priced back in January.

I wasn't really talking about Polaris.

Still doesn't mean that they are making significant money. The way the market was it would be very difficult for AMD to have priced it higher.

http://www.pcworld.com/article/3115...-470-inside-for-amd-polaris-mobile-debut.html

Pretty sure they also got the Apple desktop and laptop as well for its refresh coming up.

Apple sells a minuscule amount of macs with dGPUs. AMD's dgpu notebook marketshare is basically nonexistent, a fraction of what is was in the HD 5000/6000 era.

Sure companies would always like to make more, but they've stated that they wanted to bring Polaris and newer cards for cheaper and that it cost them much less and they got more dies because of the node shrink. Thus even though they were selling for less, they cost a lot less which doesn't mean they are skimping themselves.

This is marketing speak.

Nvidia making hand over fist in profit margins isn't a good thing. It just shows how over priced their goods are.

Nvidia will charge what the market will bear. You may personally feel that their products are overpriced but the market as a whole does not.

Obviously its good for companies to make profits, but its not anywhere near what your underlined statement would show. They set out to sell Polaris for much cheaper because their goal was "VR for the masses". Meaning you had to take the previous $400 price point down into the $200s. They've talked about this for months before release, no idea why people assumed they wanted to sell Polaris for more.

I never said Nvidia's ham fisted prices were good. They are not. Again, this is marketing speak.

 

[railven]

It is truly amazing how every thread needs to be turned into an Nividia is great and AMD sucks propaganda piece.

Odd, just casually browsing I get the complete opposite. Although it is getting tiring to see the in-fighting between a select few. It's like these posters should just admit they have feelings for each other and get a room. Haha.

 

[nurturedhate]

Odd, just casually browsing I get the complete opposite. Although it is getting tiring to see the in-fighting between a select few. It's like these posters should just admit they have feelings for each other and get a room. Haha.

But your own posts in this thread are clearly aimed to bring down pro-AMD posters based solely on the content of said posts. That's 100% anti-AMD and has absolutely nothing to do with Vega/Navi rumors. It is just more dragging the thread away from the original perceived topic of Vega/Navi rumors. Furthermore, I state as such that this thread has been purposely dragged in a specific direction and you come back with "Odd, just casually browsing I get the complete opposite.". We are obviously reading completely different threads.

 

[Bacon1]

This is marketing speak.

So them saying "We are going to target the mid range" and doing so is marketing speak? Ok sure, but again why are people expecting them to target anything but that when they've flat out said that's what their goal was?

Nvidia will charge what the market will bear. You may personally feel that their products are overpriced but the market as a whole does not.

That's because there are lots of people who will pay whatever, as long as it has Nvidia branding. They don't care that Nvidia is making more profits than ever, or are anti-consumer. Heck you have people boasting about how Nvidia is more profitable than ever while down 30% in sales.

How many posters here said that Fury X needed to be faster and cheaper than the 980 TI to consider buying it? Even when it came pre-water cooled.

Its impossible for AMD to compete and profit in the high end when they have to provide more features, better performance and at a lower cost before people will even consider buying them.

Not to mention all the gsync users who are now locked into Nvida's ecosystem. NVidia is withholding freesync support because that would give people an out and option to buy AMD in the future to keep their monitor performing as it should.

390x was a better buy than 980, yet how many recommended it over the 980?

Anyway I look forward to seeing what Vega and Navi provide.

 

[gamervivek]

AMD's problem on perf/W front isn't necessarily due to architecture. They have a huge clockspeed deficit to nvidia which is hurting them badly and process optimization has to play a big part in that. They are ahead on perf/mm2 if you normalize for clockspeed but that 10% advantage is small compared to the 50% advantage nvidia have in clocks.

This rumor while against a Big Vega is good news on clockspeed front which is the real problem AMD have.

 

[railven]

But your own posts in this thread are clearly aimed to bring down pro-AMD posters based solely on the content of said posts.

Well that's your interpretation of my post(s). I have no interest in bringing down AMD. I also have no interesting in bringing AMD up. The same applies for NV. I am, however, honest with my observations.

I've been a long time AMD/ATI supporter. People around here seem to have forgotten that probably because they can't seem to remember past a few weeks ago (or they choose not to). As someone who defended AMD, that was something I didn't like. I used the 7970 because a poster keeps going on about how it was the cat-meow but I found myself against him multiple times because I defended it, while he didn't.

It's just a common trend either institutalized by AMD themselves or buyers, but AMD has this "bargain brand" persona that continues to haunt it. To the point where AMD charging more is looked at negatively (7970) and then when AMD charges less it's cheered (290X). AMD can't survive on that kind of support.

That's 100% anti-AMD and has absolutely nothing to do with Vega/Navi rumors. It is just more dragging the thread away from the original perceived topic of Vega/Navi rumors.

By all means, report all the posts that are off-topic. This place has a habit of going off topic. Just about every thread devolves into AMD vs NV rhetoric.

Furthermore, I state as such that this thread has been purposely dragged in a specific direction and you come back with "Odd, just casually browsing I get the complete opposite.". We are obviously reading completely different threads.

No, we are interpreting things differently. If you wish for me to explain what I mean, it would yet again go off-topic.

 

[AtenRa]

Keeping GK110 around for as long as NVIDIA did was less than ideal and risky. They needed a Maxwell GM104 out in late 2013/early 2014, but for whatever reason that just didn't happen. NVIDIA was fortunate that AMD shot itself in the foot with Hawaii's reference cooling solution and that it was able to charge what it did for the 780 Ti (since full GK110 offered leadership performance at the time) otherwise that cycle could have been a real financial issue for them.

Agreed, AMD lost a huge opportunity with Hawaii back then.

 

[AtenRa]

I feel my problem with these kind of posts is they always seem to focus on now, when those chips being discussed are no longer relative.

R9 290X was faster than GTX 780 even on released date and thats on "quiet mode". On "uber" mode it was = to TITAN at almost half the price.

https://tpucdn.com/reviews/AMD/R9_290X/images/perfrel_1920.gif

ONE year later, R9 290X (launch price $549) was as fast as GTX 780Ti (launch price $699)

https://www.techpowerup.com/reviews/ASUS/GTX_980_Matrix/28.html

And if you think that R9 290 was selling for $250 at that time, it is mind blogging how people still say that Hawaii was not the best GPU of the last 3-5 years. If NVIDIA had Hawaii back in 2013, AMD would not even sell a single card at $400-550 segment in that time frame. But this is a good example that the best product dont always sell at higher volumes and marketing can make you sell anything at any price

 

[Piroko]

AMD is in it's position because of it's supporters. You can see it in Zen threads. AMD is suppose to launch a competitive product against Intel and still be cheaper. Oh-kay.

I don't consider those people supporters, but rather people who knowingly want to damage a brand to profit from it. You can see this everywhere, one of the more funny stories was with people that ran into US VW dealerships who were interested in Diesels, but only for a hefty discount "because the cars are faulty". Those are customers that are best ignored from your company's perspective because they are neither improving your situation nor are they brand loyal (ignoring that I think brand loyalty makes for bad decisions in the first place).

 

[Enigmoid]

So them saying "We are going to target the mid range" and doing so is marketing speak? Ok sure, but again why are people expecting them to target anything but that when they've flat out said that's what their goal was?

And AMD targeted those segments and that pricing because they had to, not because they wanted to. When Pitcarin launched >$300 launching a new card on a new node at a price that was essentially unlowerable was not a smart thing to do. The $200 480 was most definitely a bait and switch as those cards are nowhere to be found for instance -> $200 480 = marketing speak.

That's because there are lots of people who will pay whatever, as long as it has Nvidia branding. They don't care that Nvidia is making more profits than ever, or are anti-consumer. Heck you have people boasting about how Nvidia is more profitable than ever while down 30% in sales.

People will not pay whatever, they will pay up to what they think the card is worth. The problem is that Nvidia has been working hard to cause the consumer to overvalue their GPUs.

How many posters here said that Fury X needed to be faster and cheaper than the 980 TI to consider buying it? Even when it came pre-water cooled.

Its impossible for AMD to compete and profit in the high end when they have to provide more features, better performance and at a lower cost before people will even consider buying them.

This is a major problem that AMD has to resolve. The problem is that Nvidia has all the brand recognition. If two products are more or less identical in terms of features and price the average consumer will choose the Nvidia option. Many of the people on this forum will do the same.

 

[Piroko]

My point is this. AMD is not winning on technical merit right now (similar to how nvidia's Fermi did not win on technical merit) despite what is being said about DX12. They are simply getting by with larger dies and higher power consumption than the competition and selling at lower prices.

I don't think this is entirely true. AMD has very competitive theroetical shader throughput per mm², even with Polaris.
Polaris:
5.1 Tflops @1.12 GHz on 232mm², around 6.2 TFlops with OC (1.35 GHz)
Pascal:
3.8 Tflops @1.50 GHz on 200mm², around 5.4 TFlops with OC (2.10 GHz)

That's actually really close. And it's not a completely worthless stat, over time new game releases had the tendency to prefer cards with higher shader throughput. It's imho one of the largest reasons why Hawaii held up so well, it had more raw power than its direct competitors. Eventually games tapped into those shader units to run more and more complex effects while -for example- geometry didn't scale nearly as fast, giving Hawaii (5.6 to 5.9 Tflops, 438mm²) an edge over GK110 (5.0 Tflops, 561 mm²) and letting it catch up to GM104 (oc versions up to 5.7 Tflops, 398mm²).
It's not a perfect metric obviously (HD4970 had higher throughput than Fermi yet still lost due to lack of vram and architecture quirks) but it's a data point to keep in mind.

 

[zinfamous]

And if you think that R9 290 was selling for $250 at that time, it is mind blogging how people still say that Hawaii was not the best GPU of the last 3-5 years. If NVIDIA had Hawaii back in 2013, AMD would not even sell a single card at $400-550 segment in that time frame. But this is a good example that the best product dont always sell at higher volumes and marketing can make you sell anything at any price

BetaMax, anyone?

 

[Enigmoid]

I don't think this is entirely true. AMD has very competitive theroetical shader throughput per mm², even with Polaris.
Polaris:
5.1 Tflops @1.12 GHz on 232mm², around 6.2 TFlops with OC (1.35 GHz)
Pascal:
3.8 Tflops @1.50 GHz on 200mm², around 5.4 TFlops with OC (2.10 GHz)

That's actually really close. And it's not a completely worthless stat, over time new game releases had the tendency to prefer cards with higher shader throughput. It's imho one of the largest reasons why Hawaii held up so well, it had more raw power than its direct competitors. Eventually games tapped into those shader units to run more and more complex effects while -for example- geometry didn't scale nearly as fast, giving Hawaii (5.6 to 5.9 Tflops, 438mm²) an edge over GK110 (5.0 Tflops, 561 mm²) and letting it catch up to GM104 (oc versions up to 5.7 Tflops, 398mm²).
It's not a perfect metric obviously (HD4970 had higher throughput than Fermi yet still lost due to lack of vram and architecture quirks) but it's a data point to keep in mind.

Theoretical FLOPS generally mean very little between architectures as design tradeoffs greatly determine effective throughput. AMD has always been shader heavy, even to the point what that kind of throughput is unutilized (Tahiti, Fiji) due to bottlenecks elsewhere.

I agree that as time has gone on games have become more and more shader heavy (probably due to consoles).