Poll: Do you think AMD will be able to compete with Nvidia again during the next few years?

[Darujhistan]

They're competitive at the price points they want to be, no-one can deny that. I think in general AMD are better value for money.

Will they be competitive at the very high end?, no I don't think so, and I don't think they're interested in being so, or if they are it's a very low priority for them. They're leading the way in price/performance with Ryzen, they're better VFM in the parts of the GPU market they wish to be. They're probably happy with that, I think they believe trying to have the best GPU isn't worth the effort and there's money lower down and in forging ahead with Ryzen. NV have a larger fanbase among enthusiasts and I think they believe that if they released an identically performing halo card priced indentically NV would still sell more cards.

There's no need for AMD to win the top end GPU war.

 

[Exsurgo]

In the high end market, I am doubtful AMD can compete in the next year/2. They have not been so far behind Nvidia since the R600, and even then they sort of bounced back with the mid-range HD38XX series.

They can make a fast semi-high end card in the Vega, but it's costs far too much to produce. They can't ditch super expensive HBM2 or their power-efficiency drops even further. The last few generations they have been forced to massively clock their cards to keep up, like with Polaris, super efficient at 1Ghz core but became a power hungry beast in it's battle with the GTX1060. Same with Vega. Their GCN big async-compute/DX12 advantage has been eroded with the RTX cards that patch up Nvidia's weakness in that regard.

In the mid-range, I do think Navi will be a good card. They should be able to better leverage their console advantage and process shrink advantage to hopefully compete with the RTX 2070, but it's probably going to be a fairly long wait for Navi.

I don't have much hope for a 7nm Vega, they will still barely make any profit on a consumer version and it won't reach 1080TI/RTX2080 performance, so what's the point.

 

[humbug]

They're competitive at the price points they want to be, no-one can deny that. I think in general AMD are better value for money.

Will they be competitive at the very high end?, no I don't think so, and I don't think they're interested in being so, or if they are it's a very low priority for them. They're leading the way in price/performance with Ryzen, they're better VFM in the parts of the GPU market they wish to be. They're probably happy with that, I think they believe trying to have the best GPU isn't worth the effort and there's money lower down and in forging ahead with Ryzen. NV have a larger fanbase among enthusiasts and I think they believe that if they released an identically performing halo card priced indentically NV would still sell more cards.

There's no need for AMD to win the top end GPU war.

Yeah pretty much, nVidia's fan base are all about what they perceive to be the best at any cost, AMD just sell good cards at a reasonable price in the mid range, AMD are not interested at the highest end because no matter how competitive AMD are there people buy nVidia anyway, its been like that for a decade.

 

[4K8KW10]

Yeah pretty much, nVidia's fan base are all about what they perceive to be the best at any cost, AMD just sell good cards at a reasonable price in the mid range, AMD are not interested at the highest end because no matter how competitive AMD are there people buy nVidia anyway, its been like that for a decade.

Well, let me quote once again the predictions about Navi.
I'd like big Navi 20 possible 7680 shaders to be my next graphics card, to replace the R9 380.

Yes: https://www.techpowerup.com/249819/...aystation-5-says-the-rumor-mill?cp=3#comments

 

[Darujhistan]

Yeah pretty much, nVidia's fan base are all about what they perceive to be the best at any cost, AMD just sell good cards at a reasonable price in the mid range, AMD are not interested at the highest end because no matter how competitive AMD are there people buy nVidia anyway, its been like that for a decade.

Yeah, I can't think of a time when this was an equal fight either, it's David and Goliath. And if you think about it, NV and AMD are a reflection of two kinds of consumer, 'must have the best' and 'can I get the same thing but cheaper?'.

 

[humbug]

Well, let me quote once again the predictions about Navi.
I'd like big Navi 20 possible 7680 shaders to be my next graphics card, to replace the R9 380.

Well, 7680 shaders is about 85% more than Vega 64, that would probably put it in the middle between the 2080 and 2080TI, which is a bit higher than where Vega 64 was last generation, it was just a little faster than the 1080.

That's a bit more competitive than Vega 64 was.

I think its Navi 10 that's the RX 580 replacement, it will be about 15% faster than Vega 64, thats the card i'm looking forward to as a replacement to my 1070, if the price is right.

 

[LePhuronn]

In the mid-range, I do think Navi will be a good card. They should be able to better leverage their console advantage and process shrink advantage to hopefully compete with the RTX 2070, but it's probably going to be a fairly long wait for Navi.

I don't have much hope for a 7nm Vega, they will still barely make any profit on a consumer version and it won't reach 1080TI/RTX2080 performance, so what's the point.

8-10 months for Navi, I reckon. Lisa Su will be announcing their next gaming cards at CES in January; no reason to think it won't be Navi. The smaller Navi 10 will probably roll out around July and smack the GTX 1080 (and RTX 2070 by extension) in the chops. If the rumoured "big" Navi 20 is real then that'll probably be end of 2019, early 2020, to keep things ticking over until Arcturus.

And 7nm Vega won't happen as a consumer gaming card; it's a datacenter compute card, always was.

 

[chrcoluk]

I dont know why AMD never built a big GPU based on what they use in the ps4 pro and xbox one X, using gddr5/6.

 

[LePhuronn]

I dont know why AMD never built a big GPU based on what they use in the ps4 pro and xbox one X, using gddr5/6.

They do, it's called the RX480. Albeit downclocked with a few Vega bits tacked on.

 

[4K8KW10]

Well, 7680 shaders is about 85% more than Vega 64, that would probably put it in the middle between the 2080 and 2080TI, which is a bit higher than where Vega 64 was last generation, it was just a little faster than the 1080.

That's a bit more competitive than Vega 64 was.

I think its Navi 10 that's the RX 580 replacement, it will be about 15% faster than Vega 64, thats the card i'm looking forward to as a replacement to my 1070, if the price is right.

If the architecture efficiency stays the same? I am asking for more work to be done from the Navi shaders.

 

[chrcoluk]

They do, it's called the RX480. Albeit downclocked with a few Vega bits tacked on.

Thats a medium GPU not big.

 

[FoxEye]

8-10 months for Navi, I reckon. Lisa Su will be announcing their next gaming cards at CES in January; no reason to think it won't be Navi. The smaller Navi 10 will probably roll out around July and smack the GTX 1080 (and RTX 2070 by extension) in the chops. If the rumoured "big" Navi 20 is real then that'll probably be end of 2019, early 2020, to keep things ticking over until Arcturus.

And 7nm Vega won't happen as a consumer gaming card; it's a datacenter compute card, always was.

We won't be seeing any Navi before Q3 2019, in all likelihood. And big Navi is probably a full year behind.

 

[LePhuronn]

We won't be seeing any Navi before Q3 2019, in all likelihood. And big Navi is probably a full year behind.

July is Q3 but it'll be a paper launch as always with actual availability around end of September. But a full year after this is supposedly Arcturus, so a Big Navi at this point would be redundant. Unless Arcturus will be intended as the super top-end, Big Navi covers the upper mainstream, Navi covers the mid to entry range and Polaris finally gets put out to pasture.

 

[LePhuronn]

Thats a medium GPU not big.

You want an even bigger Polaris 20 die? And people complained that Vega was hot and hungry...

 

[4K8KW10]

You want an even bigger Polaris 20 die? And people complained that Vega was hot and hungry...

Well, the bigger dies are physically easier to cool down. Also, if the frequencies are in check, there will be no worries about the power consumption.
It all depends on the architecture efficiency and the process sweet region of frequencies.

 

[AMDPower]

everyone keeps talking about navi and big navi, and how it will sit in between 2080 and 2080ti with late 2019 release but that is not good enough.
Everyone forgets nvidia will have an answer for navi and probably on 7nm too. Navi might be good enough to compete with rtx series IF and only if it releases now not 1 year from now.

 

[Participant]

everyone keeps talking about navi and big navi, and how it will sit in between 2080 and 2080ti with late 2019 release but that is not good enough.

Whether it's 'good enough' or not depends on the price.

 

[jigger]

everyone keeps talking about navi and big navi, and how it will sit in between 2080 and 2080ti with late 2019 release but that is not good enough.
Everyone forgets nvidia will have an answer for navi and probably on 7nm too. Navi might be good enough to compete with rtx series IF and only if it releases now not 1 year from now.

Really confusing post. Why will Navi not be able to compete with current cards?

 

[4K8KW10]

everyone keeps talking about navi and big navi, and how it will sit in between 2080 and 2080ti with late 2019 release but that is not good enough.
Everyone forgets nvidia will have an answer for navi and probably on 7nm too. Navi might be good enough to compete with rtx series IF and only if it releases now not 1 year from now.

Absolutely correct.
GTX 1080 Ti now has 3584 shaders, RTX 2080 Ti now has 4352 shaders.
nvidia has nothing else to do but to double this shaders count up to 7168 shaders and 8704 shaders.

 

[Distracted]

If Navi isn't due until 2019 I don't want it to compete with current cards. I would like to see AMD release something that either competes with the top end of Nvidias next release or outright beats it.

AMD releasing something next year to compete with what Nvidia is offering this year would be incredibly short sighted.