** The AMD Navi Thread **

[Marine-RX179]

All true but your forgetting the failure rates of Turing too .
But on the other side of the fence, at VII and Rx590 launch;

Oh you're currently on a V56/V64 or below and nvidia is a no no for you, and you're so loyal and you've been waiting patiently you can have our VII which is 1080ti performance for the same price as 2 years ago.

Oh not in your budget, well how about a £230-£270 Rx590, whilst our eol v56 is about £30 more, or our competitor can offer you an older eol Gtx1070 that whops it for £270.

I don't think most people that buy AMD is because they are "loyal" though, but more of the case the alternative from Nvidia's offering in the same price bracket generally are worse in value and performance, with generally with less vram and without comparable free games promotion

Only the 2080 is possibly a better option Radeon VII, with it despite having less vram, it is generally faster by a reasonably margin and price more or less the same.

Efficiency don't mean much to people that don't really care about it or to people that care more about performance to cost.

 

[LePhuronn]

Heard today Lisa Su said she wanted 45% profit on Navi cards. I reckon that makes pricing all the more unlikely.

So for every $200 Navi card, $138 of that is to cover R&D, fabrication, assembly and shipping. For every $500 Navi card, $345 covers costs.

Of course it's much more granular than that; it'll cost the same to fabricate, assemble and ship the 3 different Navi 10 SKUs which makes the profit margin on the higher SKUs larger.

I imagine what Lisa probably meant was she wants to make 45% on Navi as an entity. So if it costs, say, $10M for all the R&D, fabrication, assembly and shipping, she wants final sales to total $14.5M. And then factor in the rumour that Sony co-funded Navi R&D, then AMD's total Navi cost is a lot lower, allowing them to recoup and earn profit through lower margins.

Doesn't sound too unreasonable to me, if I'm honest.

 

[CuriousTomCat]

So for every $200 Navi card, $138 of that is to cover R&D, fabrication, assembly and shipping. For every $500 Navi card, $345 covers costs.

Surely AMD will sell the Navi chipset to board partners at a 45% profit. Board partners then assemble the graphics card for a percentage.

So AMD don't make 45% on a £200 card. AMD only make 45% on the price they sell it to Asus and Gigabyte etc (unless they make their own founders edition).

 

[LePhuronn]

Surely AMD will sell the Navi chipset to board partners at a 45% profit. Board partners assemble the graphics card for a percentage.

So AMD don't get 45% of a £200 card. They only get 45% added on to their cost value.

True true, but I was ish-kinda factoring in AIBs given that the leaked price points probably reflect AMD's MSRP. The point was 45% margin isn't actually that big so I don't think we'll see stratospheric pricing. At least not from AMD anyway. Asus Tax is a thing, capitalism is a thing and if Navi proves to be competitive in the market segments it's intended to operate, retailers will gouge the bejeezus out of it. Because "sales expections" and other tripe.

I said before when those leaked prices came out they were a little on the low side but not crazily so, and represent AMD's goal of disrupting the market with price/performance. If Navi cards appear at a much higher cost, it'll be AIBs and retailers gouging, not AMD overcharging.

 

[Phixsator]

I don't think most people that buy AMD is because they are "loyal" though, but more of the case the alternative from Nvidia's offering in the same price bracket generally are worse in value and performance, with generally with less vram and without comparable free games promotion

Only the 2080 is possibly a better option Radeon VII, with it despite having less vram, it is generally faster by a reasonably margin and price more or less the same.

Efficiency don't mean much to people that don't really care about it or to people that care more about performance to cost.

There is a few reasons for me to go the Vega 64 card I ended up with instead of the GTX 1080ti I had before. The biggest reason was more of a gamble that ended up being worth it for me and that was for better frame intervals. I of course lost performance when comparing the average fps between the two but the frame intervals is better with my current vega card. I'm not saying this is going to be the case for everyone but it has been for me. Overall i don't miss my gtx 1080ti one bit and considering I got exactly the same for it 1½ years after I bought it i didn't loose a dime and it made it possible for me to upgrade my aging haswell platform.

 

[ANDREW GREEN]

So for every $200 Navi card, $138 of that is to cover R&D, fabrication, assembly and shipping. For every $500 Navi card, $345 covers costs.

Of course it's much more granular than that; it'll cost the same to fabricate, assemble and ship the 3 different Navi 10 SKUs which makes the profit margin on the higher SKUs larger.

I imagine what Lisa probably meant was she wants to make 45% on Navi as an entity. So if it costs, say, $10M for all the R&D, fabrication, assembly and shipping, she wants final sales to total $14.5M. And then factor in the rumour that Sony co-funded Navi R&D, then AMD's total Navi cost is a lot lower, allowing them to recoup and earn profit through lower margins.

Doesn't sound too unreasonable to me, if I'm honest.

There's retailer and partner margins to take into account too. They do need sub-£200 card to be competitive though, it just depends how fast they'll be. 1660 or 1660ti at sub £200?

It might well be that we don't get an improvement in value proposition over the current Vega 56 deals.

 

[LePhuronn]

It might well be that we don't get an improvement in value proposition over the current Vega 56 deals.

Bear in mind that Vega 10 is likely EOL now, so although the deals at this minute are tasty, you're probably not going to see much improvement price/performance as Navi replaces RX Vega lock, stock and barrel.

 

[ANDREW GREEN]

Bear in mind that Vega 10 is likely EOL now, so although the deals at this minute are tasty, you're probably not going to see much improvement price/performance as Navi replaces RX Vega lock, stock and barrel.

It makes me question if I should go Vega or if my 520W PSU wouldn't handle running Vega off a single rail when undervolted.

 

[Poneros]

http://ir.amd.com/shareholder-services/annual-meeting

You know, I just realised AMD won't have anything to surpass a 2080 ti until 2021-2022. At best we'll get a 2080 ti equivalent next year. That's quite sad. Will have to swap to Nvidia but I have this nagging feeling they'll only barely refresh their line-up again, too. Which means £1500 2080ti succesor, but again only a 20%ish boost in performance.

Hope it won't turn out that way, else I'll just do crossfire on the cheap.

 

[4K8KW10]

You know, I just realised AMD won't have anything to surpass a 2080 ti until 2021-2022. At best we'll get a 2080 ti equivalent next year. That's quite sad. Will have to swap to Nvidia but I have this nagging feeling they'll only barely refresh their line-up again, too. Which means £1500 2080ti succesor, but again only a 20%ish boost in performance.

Hope it won't turn out that way, else I'll just do crossfire on the cheap.

AMD needs to find 40% performance out of somewhere to even equalise an RTX 2080 Ti. In theory it can happen as soon as they put in production a 500 sq.mm 7nm chip with a new fixed architecture.
It might happen in 2020.

Also, if it's difficult for AMD to surpass the RTX 2080 Ti performance, so will be for nvidia itself, too, as it needs a relatively large and expensive 7nm chip.

 

[CuriousTomCat]

You know, I just realised AMD won't have anything to surpass a 2080 ti until 2021-2022. At best we'll get a 2080 ti equivalent next year. That's quite sad. Will have to swap to Nvidia but I have this nagging feeling they'll only barely refresh their line-up again, too. Which means £1500 2080ti succesor, but again only a 20%ish boost in performance.

Hope it won't turn out that way, else I'll just do crossfire on the cheap.

Nvidia are 100% guaranteed not to lower their prices or improve their performance for one reason - you and the other 80% of their customers are happy to buy one. You're encouraging them to keep doing it. Why should they waste money developing anything more than a 20% boost if 80% of the market will buy one anyway? They're a business and they're trying to make as much money as possible and who can blame them. They have a winning formula.

However, if only 50% of the market bought a 9 series card and onwards, I believe that Nvidia's current 2080ti will be close to £700. AMD would have had more money for R&D so will be able to compete.

The reason I'm buying a Navi card is to do my bit to end the monopoly. Obviously it wont happen but if AMD's market share ever grew larger than Nvidia, then I'd switch GPU manufactuer immediately to avoid another monopoly. We all gain from an even market - faster and cheaper GPU's from both company's. Everyones a winner.

The issue is that people only think of the here and now. They want a good framerate in 2019 so buy the best GPU for that purpose. Then they loose out in 2 years because Nvidia have a bigger monopoly. It's human nature I guess.

 

[flea.rider]

but as has been said the money is not in the top card .. so all she needs to do over the next yr is get rid of the debt they owe ..
then use that next lot of money to surpass NVidia .. shes playing the long game .. kudos to her

 

[CuriousTomCat]

but as has been said the money is not in the top card .. so all she needs to do over the next yr is get rid of the debt they owe ..
then use that next lot of money to surpass NVidia .. shes playing the long game .. kudos to her

All of AMD's current GPU's up to the Vega 56 are faster than an Nvidia card for the same money. However, people seem to buy the slower Nvidia card without checking the benchmarks because of the mindshare and the brand. I've heard people say it "I like Nvidia cards". They just like them, regardless of the price/performance. How can AMD compete with a belief that's inside peoples heads? I dont know if it's marketing or what it is.

I think one of the reasons people choose Nvidia when buying a mid end GPU is because they know Nvidia have the best high end card. It brand awareness. In a similar way, people may choose a particular brand of family car if that brand does well in formular one. So having the fastest helps to sell more of the cheapest.

Maybe Ryzen 3000 will help with the GPU mindshare? I dont know.

A big issue is Nvidia g sync monitors. Some people are sticking to Nvidia because they dont want to loose g sync. That's a genius move by Nvidia. The power to have a closed standard that keeps people buying your products.

 

[flea.rider]

All of AMD's current GPU's up to the Vega 56 are faster than an Nvidia card for the same money. However, people seem to buy the slower Nvidia card without checking the benchmarks because of the mindshare and the brand. I've heard people say it "I like Nvidia cards". They just like them, regardless of the price/performance. How can AMD compete a belief that's inside peoples heads? I dont know if it's marketing or what it is.

Maybe Ryzen 3000 will help with the GPU mindshare? I dont know.

once the 1.3b debt is gone she can go for the throat .. ie do something NVidia cant really do reduce selling point .. yeah NVidia may be able to knock 10-15% off there cards but after that the stockholders will start screaming and selling shares .
but with the ryzen being so strong .. she could easily keep gpu selling point down and still make a profit

 

[CuriousTomCat]

once the 1.3b debt is gone she can go for the throat .. ie do something NVidia cant really do reduce selling point .. yeah NVidia may be able to knock 10-15% off there cards but after that the stockholders will start screaming and selling shares .
but with the ryzen being so strong .. she could easily keep gpu selling point down and still make a profit

OMG I didn't realise their debt was so big

 

[Grim5]

AMD needs to find 40% performance out of somewhere to even equalise an RTX 2080 Ti. In theory it can happen as soon as they put in production a 500 sq.mm 7nm chip with a new fixed architecture.
It might happen in 2020.

Also, if it's difficult for AMD to surpass the RTX 2080 Ti performance, so will be for nvidia itself, too, as it needs a relatively large and expensive 7nm chip.

So you're saying the 2080ti is going to hold it's second hand value for a long time like the 1080ti, I like where this is going. When I do sell the 2080ti, let's see if I can 80% of my money back

 

[Besty]

AMD needs to find 40% performance out of somewhere to even equalise an RTX 2080 Ti. In theory it can happen as soon as they put in production a 500 sq.mm 7nm chip with a new fixed architecture.
It might happen in 2020.

Also, if it's difficult for AMD to surpass the RTX 2080 Ti performance, so will be for nvidia itself, too, as it needs a relatively large and expensive 7nm chip.

This is not true. The market does not need a 2080ti rival, it needs a 1080 desktop rival that will work at desktop performance levels inside a laptop without melting it.

The market is shifting faster than many people here realise, laptops are the new battleground and it is no coincidence that intel are prioritising newer laptop chips over desktop.

Desktop is a busted flush, the age of massive CPUs and GPUs is about to become a niche within a niche.

The golden age of mobile gaming is about to arrive and it is the new battleground for growth.

 

[Grim5]

Nvidia to launch refresh Turing line-up after Navi using upgraded faster GDDR6 modules

https://www.pcgamesn.com/nvidia/turing-gpu-16gbps-gddr6

 

[4K8KW10]

This is not true. The market does not need a 2080ti rival, it needs a 1080 desktop rival that will work at desktop performance levels inside a laptop without melting it.

The market is shifting faster than many people here realise, laptops are the new battleground and it is no coincidence that intel are prioritising newer laptop chips over desktop.

Desktop is a busted flush, the age of massive CPUs and GPUs is about to become a niche within a niche.

The golden age of mobile gaming is about to arrive and it is the new battleground for growth.

The mobile market needing a shrunk, power efficient RX Vega 64 is just a temporary goal. I hope you don't expect that the RX Vega 64 level of performance will be the last and final.
Next will be an RTX 2080 Ti level of performance shrunk and power efficient in the mobiles. But that will probably happen with the help of 3nm process (after the current 7nm and the future 5nm).

But, I wouldn't rely on the mobiles for gaming, except if the gaming doesn't involve some light e-sports titles.

4K and 8K gaming on mobiles..... Good luck with that!

 

[Grim5]

I don't understand why a phone would need a 4k screen anyway, 1080p is perfect for the size of those screens - and mobile gaming will always suck without a standard control input device - pressing fake button on a screen is not a input device, it's garbage