AMD CEO Confirms 7nm Navi High-End Radeon RX Graphics Cards and 4thGen Ryzen CPUs For Notebooks

[Uvaman2]

There are no guarantees in tech. Large corporations can end up with crappy cultures that produce mediocre products for a long time.
By the same token a smaller one can create good things for a while too.
In research and development is not like you can have a baby in a month becuase you have nine moms, it doesn't work like that.

Navi just started. This is shit first gen Navi.
Nvidia does not have a great architecture, its fat, its not superior to previous. 2000 series is a to 1000 series as Vega is to prev gen... Nvidia done nothing but fatter chips, and typical process improvement.
Yes AMD can fumble, by my own account, no guarantees, but the prospects are positive. AMD just re started in GPUs, and first out the gate is a product with a lot of performance and potential.
When interviewed about the future the jebaited guy laughing told me all I needed to hear (he was REALLY laughing as in competition is fucked no matter what kind of laugh)
As far as CPUs go, who says they are done? Why would that be so?
Also why skewer AMD about HBM use, shit its the producer of that memory that screw them over, there is nothing wrong with HBM, maybe they should do some CPUs with 16gb of it.

 

[Dayaks]

There are no guarantees in tech. Large corporations can end up with crappy cultures that produce mediocre products for a long time.
By the same token a smaller one can create good things for a while too.
In research and development is not like you can have a baby in a month becuase you have nine moms, it doesn't work like that.

Navi just started. This is shit first gen Navi.
Nvidia does not have a great architecture, its fat, its not superior to previous. 2000 series is a to 1000 series as Vega is to prev gen... Nvidia done nothing but fatter chips, and typical process improvement.
Yes AMD can fumble, by my own account, no guarantees, but the prospects are positive. AMD just re started in GPUs, and first out the gate is a product with a lot of performance and potential.
When interviewed about the future the jebaited guy laughing told me all I needed to hear (he was REALLY laughing as in competition is fucked no matter what kind of laugh)
As far as CPUs go, who says they are done? Why would that be so?
Also why skewer AMD about HBM use, shit its the producer of that memory that screw them over, there is nothing wrong with HBM, maybe they should do some CPUs with 16gb of it.

I am surprised AMD didn’t do a APU with built in HBM yet. Maybe cost is the issue.

Turing actually has 15% better Cuda cores (IIRC, but can depend on the game) and I believe 50% higher effective memory bandwidth due to memory compression. That was from nVidia so grain of salt. Turing does perform better than expected in rasterized if you were to try and guage it off Cuda core count and Pascal performance per Cuda core.

It’s just the RTX crap takes up die space....

 

[Snowdog]

I am surprised AMD didn’t do a APU with built in HBM yet. Maybe cost is the issue.

Many of us would love to have that. At least until we see the price tag.

It definitely would increase cost a lot, and would it even fit in an AM4 socket?

Also it would almost certainly be a specific HBM part so volume comes into play. A higher priced lower volume part that requires it's own tape out probably doesn't have a viable business case.

 

[Aireoth]

People once thought the giant that was IBM would rule the computer world, being big and having cash is no guarantee of success.

IBM was slipping up for decades, we don’t have that environment in the CPU and GPU space at the moment, and AMD is responsible for that. AMD won’t route Intel and nVidia, they are simply to aware of the threat, and we are in a much different market both from a consumer standpoint and a government (we now have precedence that you can be too big to fail, and that the gov will step in and prevent that failure).

 

[threehundredsixftw]

IBM was slipping up for decades, we don’t have that environment in the CPU and GPU space at the moment, and AMD is responsible for that. AMD won’t route Intel and nVidia, they are simply to aware of the threat, and we are in a much different market both from a consumer standpoint and a government (we now have precedence that you can be too big to fail, and that the gov will step in and prevent that failure).

i could see it happening with intel because of 2 factors added together. 1) AMD has been on their A game with Ryzen. 2) Intel has been struggling (10nm, etc.) if both conditions continue like this, it wouldnt be inconceivable for intel to be kicked aside in the consumer market. IMO Intel and AMD will be pushing each other for a long time.

as far as AMD beating Nvidia... i dont see it. i believe the market is over saturated with 1080/1080ti's anyway so neither have much incentive to push the envelope. i wont be upgrading any time soon just to get ray tracing so ¯\_(ツ)_/¯

 

[Aireoth]

i could see it happening with intel because of 2 factors added together. 1) AMD has been on their A game with Ryzen. 2) Intel has been struggling (10nm, etc.) if both conditions continue like this, it wouldnt be inconceivable for intel to be kicked aside in the consumer market. IMO Intel and AMD will be pushing each other for a long time.

as far as AMD beating Nvidia... i dont see it. i believe the market is over saturated with 1080/1080ti's anyway so neither have much incentive to push the envelope. i wont be upgrading any time soon just to get ray tracing so ¯\_(ツ)_/¯

Intel does have a problem, and given its largely in the architecture it might take them 3 years to fix rather than 18 months, but they are active unlike IBM, they poached talent from AMD for example. This is just more of the same, Intel slips up, AMD kicks em in the balls, and Intel gets back on track.

Intel is magnitudes larger than nVidia, it is more likely IMO that AMD would take down nVidia than Intel. It is just that Intel is making the current misteps and nVidia isn't, but they have in the past.

 

[FrgMstr]

Read the OP. Post on topic.

 

[Zarathustra[H]]

Many of us would love to have that. At least until we see the price tag.

It definitely would increase cost a lot, and would it even fit in an AM4 socket?

Also it would almost certainly be a specific HBM part so volume comes into play. A higher priced lower volume part that requires it's own tape out probably doesn't have a viable business case.

But why? Name even one consumer/gaming workload that noticibly benefits from HBM.

 

[Ready4Dis]

HBM was supposed to be cheaper, but you know, AMD and planning

Well, to be fair they were not the ones promising themselves the lower prices. They had a GPU with huge memory requirements... You thought they were power hungry and hot, lol, imagine they used ddr 5 with a 384 or 512 bit interface.... They would have ended up with 50-80 more watts to dissipate and no benefit to speed. The only benefit would have been a lower cost, but they would have had to design a new (wider) memory bus as well. They did what they thought was the correct choice, obviously hindsight and all. And let's be honest, their track record on best choices... Lol, yeah.

 

[Ready4Dis]

But why? Name even one consumer/gaming workload that noticibly benefits from HBM.

One work load? Hmmm... I have an hbm card, memory overclock gives more performance, GPU clocks... Not so much, meaning even with hbm it's still starved. If that was ddr5, they would have way higher Watts and had to increase the bus width to 384/512 to make it the same speed.

 

[Ready4Dis]

Read the OP. Post on topic.

Sorry, will get back on track.

 

[Snowdog]

But why? Name even one consumer/gaming workload that noticibly benefits from HBM.

I should have been more clear.

You aren't going to just stick HBM on an APU design that has the same amount of GPU cores. There really is no point doing using a GPU core designed for AM4 DDR4 bandwidth and just switching to HBM.

An HBM using APU, would have more GPU cores, to make use of that extra HBM bandwidth. You design/parts to make use of expensive resources, or you don't include expensive resources.

 

[funkydmunky]

I argue because fanbois think AMD will somehow miraculously release a 2080ti killer (or insert new 2080ti of tomorrow) and charge $700, that is way more of a illlogical leap than the assumption that AMD will only slightly undercut nVidia's pricing due to multiple factors:

1. R&D costs of making that 2080TI killer.
2. The high bar has already been set, and unless the market rejects nVidia's high bar, it is the new normal for pricing.
3. The stakeholders demand reasonable and prudent action from the C-Suite which includes pricing appropriately. This rolls back to item 1, which is the recovery of R&D cost and 2 which is the current market price. If you are going to undercut the market by almost 50% then you are going to need some magic numbers to prevent legal action against the board for throwing away profits of a publicly traded company.

It has been said that AMD is targeting a %45 profit margin on their cards. If they can make a 2080ti competitor that fits that $700 price point (I would guess $799) then they should launch such a market disrupting card. When you have a 5700XT that competes with a 2070 @ $400, shouldn't $800 buy you 2080ti performance? AMD would just be pricing accordingly and Nvidia who would have to adjust their hyper inflated cards, although it is a huge card and very expensive to make. Can they?

 

[5150Joker]

There are no guarantees in tech. Large corporations can end up with crappy cultures that produce mediocre products for a long time.
By the same token a smaller one can create good things for a while too.
In research and development is not like you can have a baby in a month becuase you have nine moms, it doesn't work like that.

Navi just started. This is shit first gen Navi.
Nvidia does not have a great architecture, its fat, its not superior to previous. 2000 series is a to 1000 series as Vega is to prev gen... Nvidia done nothing but fatter chips, and typical process improvement.
Yes AMD can fumble, by my own account, no guarantees, but the prospects are positive. AMD just re started in GPUs, and first out the gate is a product with a lot of performance and potential.
When interviewed about the future the jebaited guy laughing told me all I needed to hear (he was REALLY laughing as in competition is fucked no matter what kind of laugh)
As far as CPUs go, who says they are done? Why would that be so?
Also why skewer AMD about HBM use, shit its the producer of that memory that screw them over, there is nothing wrong with HBM, maybe they should do some CPUs with 16gb of it.

I don't know if I should take you seriously or if you're trolling with what you wrote. So let's deconstruct a little:

Navi just started. This is shit first gen Navi.
Nvidia does not have a great architecture, its fat, its not superior to previous. 2000 series is a to 1000 series as Vega is to prev gen... Nvidia done nothing but fatter chips, and typical process improvement.

Clock for clock, Turing is faster than Pascal, this has been proven. It also has much better async performance than Pascal, improvements in NVENC, color compression and has DXR + DLSS tossed in for good measure. Now what does Navi bring in comparison? Navi has RIS and 7nm working for it. If we're going to say anyone benefited from a smaller process, its AMD as their power usage would be through the roof as usual if they didn't have 7nm. So what else does Navi bring? Nothing, it only barely matches 3 yr old Pascal and the 1080 Ti is still superior to anything Navi has right now. Big Navi remains to be seen but it will be the same old, high power usage, 7nm dependence and it probably won't beat 2080 Ti in raw performance despite having less hardware features baked in. To me, Navi is not impressive at all, Turing is a far better architecture, it just suffers from being overpriced because NVIDIA loves its margins.

As for the rest of what you wrote, they're doing good in the CPU dept. but their GPUs are only reaching parity with 3 yr old Pascal and still falling short. Big Navi will be met with either price drops from NVIDIA or Ampere will eat its lunch.

 

[Gamer X]

Clock for clock Turing is 5% faster than their 3 year old GPU. This has been proven...

Turing finally gets async compute, Pascal didn't... This has been proven...

Turing has broken ray-tracing, as an ad-hoc attempt to sell/market "Ray Tracing" on stupid people. Most People don't buy, or upgrade for $800.. for 30% performance hit. This has been proven...

Turing's RTX is broken and the GeForce RTX 2000 Series will never be able to do Real Time Ray Tracing, without a severe hit in performance. Nvidia "advanced features" that are baked in are not for games... they are for Enterprise. People will try and use such features as an excuse as to why Turing isn't that much faster than the 3 year old Pascals. Turing will never get better in future games.

Turing uses DLSS as another gimmick to try and market their servers stuff to Gamers and to hide the actual shortcomings with Turing as a hand me down chip. This has been proven...




Compared to RDNA... Turing is obsolete.

 

[5150Joker]

Clock for clock Turing is 5% faster than their 3 year old GPU. This has been proven...

Turing finally gets async compute, Pascal didn't... This has been proven...

Turing has broken ray-tracing, as an ad-hoc attempt to sell/market "Ray Tracing" on stupid people. Most People don't buy, or upgrade for $800.. for 30% performance hit. This has been proven...

Turing's RTX is broken and the GeForce RTX 2000 Series will never be able to do Real Time Ray Tracing, without a severe hit in performance. Nvidia "advanced features" that are baked in are not for games... they are for Enterprise. People will try and use such features as an excuse as to why Turing isn't that much faster than the 3 year old Pascals. Turing will never get better in future games.

Turing uses DLSS as another gimmick to try and market their servers stuff to Gamers and to hide the actual shortcomings with Turing as a hand me down chip. This has been proven...




Compared to RDNA... Turing is obsolete.

No wonder AMD is dead in the gpu market, they scrape from the bottom of the barrel to hire shills from Craigslist as part of their Team Red marketing efforts while outsourcing gpu development to second tier gpu teams in China --no surprise why Raja jumped ship to Intel.

 

[Gamer X]

No wonder AMD is dead in the gpu market, they hire shills from Craigslist as part of their Team Red marketing efforts while outsourcing gpu development to second tier gpu teams in China --no wonder Raja jumped ship to Intel.

You should look in the mirror, and try to use REALITY and facts for the basis of your posts.

Instead, you keep attacking the messengers. It is because you are defending something no Gamer would defend. You are not a Gamer you are a mouthpiece for Jensen... since He can't make any more public appearances without getting egg all over his face.

So go cheerlead somewhere else, your team lost. The rest of us watch the Match & are now enjoying the AMD ticker-take parade down the main boulevard.




RDNA is killing it... while Turing is a FLOP. (Nothing you can say can change this.)

 

[5150Joker]

You should look in the mirror, and try to use REALITY and facts for the basis of your posts.

Instead, you keep attacking the messengers. It is because you are defending something no Gamer would defend. You are not a Gamer you are a mouthpiece for Jensen... since He can't make any more public appearances without getting egg all over his face.

So go cheerlead somewhere else, your team lost. The rest of us watch the Match & are now enjoying the AMD ticker-take parade down the main boulevard.




RDNA is killing it... while Turing is a FLOP. (Nothing you can say can change this.)

 

[Gamer X]

Look at my first sentence in my second paragraph... (last post)

Do understand how YOU LOOK... to other people when all you can do in an discussion/debate/argument is toss memes, instead of facts.

 

[Gamer X]

More RDNA Incoming.....



Anyone know the score of the game?

 

[vegeta535]

It has been said that AMD is targeting a %45 profit margin on their cards. If they can make a 2080ti competitor that fits that $700 price point (I would guess $799) then they should launch such a market disrupting card. When you have a 5700XT that competes with a 2070 @ $400, shouldn't $800 buy you 2080ti performance? AMD would just be pricing accordingly and Nvidia who would have to adjust their hyper inflated cards, although it is a huge card and very expensive to make. Can they?

This isn't how business works. AMD is not some good guy that wants the best for consumers. They will charge w/e they can get away with. Sure they will under cut Nvidia for the same performance but not by $500. I can see then pricing a 2080ti killer at $1000 or maybe $900 but not $700. Even if they are making 50% profit at $700.

 

[Aireoth]

It has been said that AMD is targeting a %45 profit margin on their cards. If they can make a 2080ti competitor that fits that $700 price point (I would guess $799) then they should launch such a market disrupting card. When you have a 5700XT that competes with a 2070 @ $400, shouldn't $800 buy you 2080ti performance? AMD would just be pricing accordingly and Nvidia who would have to adjust their hyper inflated cards, although it is a huge card and very expensive to make. Can they?

First, we have no idea what a 45% profit margin looks like on a new card with new technology. While I have often told nVidia fans that despite the fact that the 2080ti Die is large and thus more costly, the end user doesn't care about die size. They don't care what the card cost to manufacture, or what profit the company wants, they care about the cost v. performance and their purchasing power.

Second, they could undercut the market, but you and many others are ignoring that they are a publicly traded company, which means they have stakeholders in the form of everyone and every institution that holds their stock. Individuals won't take any action against a company in the stock market, but institutional investors absolutely will. So if they undercut the market by 50% they need to be darn sure they can justify that move to investors, or they will find themselves in a lawsuit, particularly the C-suite.

Third, it just doesn't make long term business sense unless your product has so little market penetration (basically unknown player) that you are undercutting to aggressively take market share. AMD isn't an unknown, and both nVidia and Intel have show in the past that they are more than capable of quickly reacting to AMD, just look at the 980Ti. Long term, AMD needs to hold their profit % up and maintain a healthy war chest to fund their R&D costs while they compete with two larger entities in two separate markets. It makes a lot more sense for them to price inline competitive with nVidia.

The prime reason to uncut the market by that much from a business perspective (most people here only see it from the consumer side) is to acquire market share to drive people to your product. With nVidia willing to cut margins almost on the spot when threatened, AMD runs the risk of actually increasing nVidia's market share if they don't play their cards right. In that situation we (the consumer) would win in the short run, but lose in the long term as it would beat up AMD's ability to compete.

This is all hypothetical, but based in real world business and business models. The only way that AMD could really undercut and hurt nVidia, take market share, and appease investors is a miracle product. A miracle product is one that delivers 2080ti performance at a fraction of the cost, allowing them to take similar profits yet price at or just above nVidia's cost (that would be manufacturing cost, not shelf price). That only really happens when your competitor has completely lost their competitive edge, while nVidia is priced high, they are not noncompetitive in performance.

I just don't see 700$ halo cards happening again.

*Edit: some spelling and grammar.

 

[Nobu]

First, we have no idea what a 45% profit margin looks like on a new card with new technology. While I have often told nVidia fans that despite the fact that the 2080ti Die is large and thus more costly, the end user doesn't care about die size. They don't care what the card cost to manufacture, or what profit the company wants, they care about the cost v. performance and their purchasing power.

Second, they could undercut the market, but you and many others are ignoring that they are a publicly traded company, which means they have stakeholders in the form of everyone and every institution that holds their stock. Individuals won't take any action against a company in the stock market, but institutional investors absolutely will. So if they undercut the market by 50% they need to be darn sure they can justify that move to investors, or they will find themselves in a lawsuit, particularly the C-suite.

Third, it just doesn't make long term business sense unless your product has so little market penetration (basically unknown player) that you are undercutting to aggressively take market share. AMD isn't an unknown, and both nVidia and Intel have show in the past that they are more than capable of quickly reacting to AMD, just look at the 980Ti. Long term, AMD needs to hold their profit % up and maintain a healthy war chest to fund their R&D costs while they compete with two larger entities in two separate markets. It makes a lot more sense for them to price inline competitive with nVidia.

The prime reason to uncut the market by that much from a business perspective (most people here only see it from the consumer side) is to acquire market share to drive people to your product. With nVidia willing to cut margins almost on the spot when threatened, AMD runs the risk of actually increasing nVidia's market share if they don't play their cards right. In that situation we (the consumer) would win in the short run, but lose in the long term as it would beat up AMD's ability to compete.

This is all hypothetical, but based in real world business and business models. The only way that AMD could really undercut and hurt nVidia, take market share, and appease investors is a miracle product. A miracle product is one that delivers 2080ti performance at a fraction of the cost, allowing them to take similar profits yet price at or just above nVidia's cost (that would be manufacturing cost, not shelf price). That only really happens when your competitor has completely lost their competitive edge, while nVidia is priced high, they are not noncompetitive in performance.

I just don't see 700$ halo cards happening again.

*Edit: some spelling and grammar.

I could see an $800-900 AMD card competing with the 2080/ti, if they didn't have to significantly increase cost to include rt hardware, or if they didn't bother at all. I don't expect an equivalent card to cost significantly less, at all.

 

[funkydmunky]

First, we have no idea what a 45% profit margin looks like on a new card with new technology. While I have often told nVidia fans that despite the fact that the 2080ti Die is large and thus more costly, the end user doesn't care about die size. They don't care what the card cost to manufacture, or what profit the company wants, they care about the cost v. performance and their purchasing power.

Second, they could undercut the market, but you and many others are ignoring that they are a publicly traded company, which means they have stakeholders in the form of everyone and every institution that holds their stock. Individuals won't take any action against a company in the stock market, but institutional investors absolutely will. So if they undercut the market by 50% they need to be darn sure they can justify that move to investors, or they will find themselves in a lawsuit, particularly the C-suite.

Third, it just doesn't make long term business sense unless your product has so little market penetration (basically unknown player) that you are undercutting to aggressively take market share. AMD isn't an unknown, and both nVidia and Intel have show in the past that they are more than capable of quickly reacting to AMD, just look at the 980Ti. Long term, AMD needs to hold their profit % up and maintain a healthy war chest to fund their R&D costs while they compete with two larger entities in two separate markets. It makes a lot more sense for them to price inline competitive with nVidia.

The prime reason to uncut the market by that much from a business perspective (most people here only see it from the consumer side) is to acquire market share to drive people to your product. With nVidia willing to cut margins almost on the spot when threatened, AMD runs the risk of actually increasing nVidia's market share if they don't play their cards right. In that situation we (the consumer) would win in the short run, but lose in the long term as it would beat up AMD's ability to compete.

This is all hypothetical, but based in real world business and business models. The only way that AMD could really undercut and hurt nVidia, take market share, and appease investors is a miracle product. A miracle product is one that delivers 2080ti performance at a fraction of the cost, allowing them to take similar profits yet price at or just above nVidia's cost (that would be manufacturing cost, not shelf price). That only really happens when your competitor has completely lost their competitive edge, while nVidia is priced high, they are not noncompetitive in performance.

I just don't see 700$ halo cards happening again

How is AMD pricing their 12 and soon 16 core CPU's compared to Intel? Stakeholders upset?
Same strategy applies to GPU's. No one is going to force them to overcharge as long they are hitting profit margins.

 

[Snowdog]

How is AMD pricing their 12 and soon 16 core CPU's compared to Intel? Stakeholders upset?
Same strategy applies to GPU's. No one is going to force them to overcharge as long they are hitting profit margins.

No, it really doesn't apply. AMDs chiplet design gives them some cost advantages at 12 and 16 cores, and Intel seldom (almost never) does price cuts to compete better.

Since the first Ryzen, Intel has just let AMD take market share without doing a price cut.

In GPUs AMD doesn't really have any kind of cost advantage, and NVidia is a much more fierce competitor that responds to each competitive challenge. See RTX "Super" cards which were prompted by Navi. Significantly undercutting NVidia will just draw a reaction from NVidia and lower AMDs profit.

AMD is not some kind of white knight. They are like every other profit motivated company. They want the fattest margin they can get, and they will do that by targeting prices just a little below NVidas. Like they have done with Polaris, Vega, Vega 7nm and Navi.

 

[Gamer X]

I could see an $800-900 AMD card competing with the 2080/ti, if they didn't have to significantly increase cost to include rt hardware, or if they didn't bother at all. I don't expect an equivalent card to cost significantly less, at all.

If you can see AMD doing that (selling big navi @ $800-900), then why can't you see them selling it at $499 -$599, instead...?

From a Business standpoint, they would sell every chip they could make... and AMD plans on making a lot of them, because understand, that "big navi" is still going to be small. Perhaps only 335mm^2 on 7nm. If so, then AMD can readjust their whole product tier once RDNA(1.5) lands, & sell big-navi-12 (RX 5800 Series) starting at $499, while little-navi-10, takes another $50 price reduction starting at $299. And given economy of scale will be able to sell every GPU AMD makes. "Big-Navi" (navi12) will be a pure pixel pumping GPU that crunches 1080p/1440p/2160p resolutions for mainstream Gamers.



Late next year in 2020, comes Nvidia's Ampere and AMD's RDNA2 (next Gen.

 

[Nobu]

If you can see AMD doing that (selling big navi @ $800-900), then why can't you see them selling it at $499 -$599, instead...?

From a Business standpoint, they would sell every chip they could make... and AMD plans on making a lot of them, because understand, that "big navi" is still going to be small. Perhaps only 335mm^2 on 7nm. If so, then AMD can readjust their whole product tier once RDNA(1.5) lands, & sell big-navi-12 (RX 5800 Series) starting at $499, while little-navi-10, takes another $50 price reduction starting at $299. And given economy of scale will be able to sell every GPU AMD makes. "Big-Navi" (navi12) will be a pure pixel pumping GPU that crunches 1080p/1440p/2160p resolutions for mainstream Gamers.



Late next year in 2020, comes Nvidia's Ampere and AMD's RDNA2 (next Gen.

I just don't expect them to, simple as that. I'm not saying they can't or won't, but it doesn't seem likely.

 

[ccityinstaller]

If you can see AMD doing that (selling big navi @ $800-900), then why can't you see them selling it at $499 -$599, instead...?

From a Business standpoint, they would sell every chip they could make... and AMD plans on making a lot of them, because understand, that "big navi" is still going to be small. Perhaps only 335mm^2 on 7nm. If so, then AMD can readjust their whole product tier once RDNA(1.5) lands, & sell big-navi-12 (RX 5800 Series) starting at $499, while little-navi-10, takes another $50 price reduction starting at $299. And given economy of scale will be able to sell every GPU AMD makes. "Big-Navi" (navi12) will be a pure pixel pumping GPU that crunches 1080p/1440p/2160p resolutions for mainstream Gamers.



Late next year in 2020, comes Nvidia's Ampere and AMD's RDNA2 (next Gen.

AMD isn't going to sell a 2080ti class GPU for anything less then $800+. There is no reason for them to price a GPU that meets or exceeds the competitors offerings for half the price. It is terrible business, and it's also illegal for AMD to undervalue their product when there is a clear case that they would sell every card they make for $800+.

You need to chill with the insane level of irrational trolling you are doing. You are just as bad as the Nvidia crowd here.

Knock it off.

 

[Gamer X]

AMD isn't going to sell a 2080ti class GPU for anything less then $800+. There is no reason for them to price a GPU that meets or exceeds the competitors offerings for half the price. It is terrible business, and it's also illegal for AMD to undervalue their product when there is a clear case that they would sell every card they make for $800+.

You need to chill with the insane level of irrational trolling you are doing. You are just as bad as the Nvidia crowd here.

Knock it off.

Yes, there is a reason for AMD to sell their 335mm^2'ish "bigger-navi" sized GPU for Vega prices.

What is AMD going to do with all those small 7nm chips falling off the wafers? You do realize that Navi10 is about the size of Polaris and if AMD wanted, could sell the 5700 Series much cheaper and still rake in profits and more mindshare and market space. Nvidia isn't in the position to do that.

AMD is winning here with RNDA, because Dr Lisa Su took a long hard look and went directly after Gamers with RDNA. It is not a all-around card, RDNA is 100% Gaming card. (not a hand-me-down-architecture like Turing is)

And with RDNA's 7nm advantage, AMD can use economy of scale to push Nvidia out of the mainstream market. AMD has all the time they need to lay this all out, before Nvidia can respond to RDNA. RDNA is really good architecture and extremely well accepted.. which means that AMD is also grabbing mainstream mindshare & about to claim more. (I am sure everyone who turns on their xbox/playstation will see the RDNA logo screen too..) ​

Have you ever heard of "ThreadRipper"..? (Offer twice what Intel offered -&- at half the price..!)




Dr Lisa Su said that Gamers were going to get their Threadripper moment in the GPU space. I don't think she was lying...!

More, for less...!

 

[Ready4Dis]

I really hope AMD doesn't put out a 2080ti performance card for $499... It's just not logical. $699-$799 makes more sense for them to pull in some much needed cash. Market share is good, but it doesn't pay the bills. If they can sell out at $800 a card, there is no point in selling lower, they are just leaving money on the table for nothing. It will take a little while for prices to come down, it won't (and shouldn't) happen in a single release.

 

[Aireoth]

AMD isn't going to sell a 2080ti class GPU for anything less then $800+. There is no reason for them to price a GPU that meets or exceeds the competitors offerings for half the price. It is terrible business, and it's also illegal for AMD to undervalue their product when there is a clear case that they would sell every card they make for $800+.

You need to chill with the insane level of irrational trolling you are doing. You are just as bad as the Nvidia crowd here.

Knock it off.

Just ignore the crazies, they will always be with us.

Personally I’d be happy with a $800 2080Ti red team model.

 

[IdiotInCharge]

I really hope AMD doesn't put out a 2080ti performance card for $499... It's just not logical. $699-$799 makes more sense for them to pull in some much needed cash. Market share is good, but it doesn't pay the bills. If they can sell out at $800 a card, there is no point in selling lower, they are just leaving money on the table for nothing. It will take a little while for prices to come down, it won't (and shouldn't) happen in a single release.

Let's say that magic happens and AMD can still hit both their revenue and margin targets with a US$500 2080Ti competitor- hypothetically, so DXR or no- and let's say that they also wanted to claw back some marketshare, then they'd absolutely be justified in selling low.

The challenge is, based on the current Navi release, Big Navi just isn't likely to get anywhere close to that- and while I certainly hope that they do the release differently with drivers / firmware / coolers, their past actions don't make for a favorable outlook for a solid release and that's what they'd need to really grab some marketshare back even if they can get close in terms of performance.

 

[Ready4Dis]

Let's say that magic happens and AMD can still hit both their revenue and margin targets with a US$500 2080Ti competitor- hypothetically, so DXR or no- and let's say that they also wanted to claw back some marketshare, then they'd absolutely be justified in selling low.

The challenge is, based on the current Navi release, Big Navi just isn't likely to get anywhere close to that- and while I certainly hope that they do the release differently with drivers / firmware / coolers, their past actions don't make for a favorable outlook for a solid release and that's what they'd need to really grab some marketshare back even if they can get close in terms of performance.

My point is if they sell @ $500 and can't keep them in stock, or sell @ $700 and can't keep them in stock, the lower price doesn't give them more market share, they are limited on production. If they price $1000 and they aren't selling, then it's not a good price point. They *should* sell them for as much as they can to be able to be on that border of selling out. Remember how pissy people get when they can't get something on release immediately. There where threads on there claiming the 3900x was a paper launch because they were selling.out immediately. Just imagine if a 2080ti equivalent came out at $500... (Even without dxr), it'd be sold out every day causing frustration, and not gaining any more market share while making AMD less money. You have to try to keep supply and demand balanced.

 

[Ready4Dis]

Oh, and I have no clue how close it will get in performance, this is a still hypothetical of course

 

[IdiotInCharge]

My point is if they sell @ $500 and can't keep them in stock, or sell @ $700 and can't keep them in stock, the lower price doesn't give them more market share, they are limited on production.

I don't disagree- but since it's hypothetical, what if they can? What if they can produce and distribute enough US$500 2080TI-killers to actually satisfy market demand?

Yeah, they'd literally be paying out of pocket for marketshare, but as you see with Nvidia and Intel (and other market-dominating companies), that has its benefits down the road- expanding marketshare when you can is literally a form of investment!

 

[Ready4Dis]

I don't disagree- but since it's hypothetical, what if they can? What if they can produce and distribute enough US$500 2080TI-killers to actually satisfy market demand?

Yeah, they'd literally be paying out of pocket for marketshare, but as you see with Nvidia and Intel (and other market-dominating companies), that has its benefits down the road- expanding marketshare when you can is literally a form of investment!

If they could get that much capacity from TSMC, then sure, again supply and demand. If they could produce that many and still make their shareholders happy with margins, then yes. But, let's be honest, this is all highly unlikely. More likely is supply for their top tier binned GPU won't be high enough to sell cheaply for many reasons. Which means they should try to keep supply and demand equalized, this is regardless of whether it's slower or faster than the 2080 super or 2080ti.

 

[Uvaman2]

I gotta say i do get it about pricing a notch below Nvidia and calling it a day... But also princing for disruption has a lot of value. This is exactly what they are doing in server... Truth be told, they have to. Even with the security holes intel has, the over price, the lack of performance... Intel could and probably will remain dominant on market intertia alone.
AMD has NEVER grown dominant enough to count on intertia to keep them shoveling billions in even when offering inferior products like Intel is right now.
Nvidia grew, even if they lose all the performance metrics, unertia alone will keep them in the billions.
Perhaps AMD ain't ready to deliver market disruption, I think they should try if able to, and that means undercutting Nvidia AGGRESSIVELY, not a little bit, but like a savage.
I think they are perhaps looking to release a definitive top end card, and then butcher nvidia everywhere else, fuck it all style.
I still say, AMD has invisible mGPU, I'm telling you, this is the nvidia killer, the Navi 2-3 .
It makes sense that they would be trying to go full chiplet on gpus too, and a kind of evolved io die could be one that divides and balances work among multiple gpu dies, and finishes the video output... Yes yes im smoking...
Not about Nvidia getting deeply undercut, if they could, I think they NEED to, same as server...

 

[Snowdog]

Perhaps AMD ain't ready to deliver market disruption, I think they should try if able to, and that means undercutting Nvidia AGGRESSIVELY, not a little bit, but like a savage.

I always wonder why people think NVidia won't respond in kind, and just leave their prices high and let AMD take the market share?

AFAICT, Unlike Intel, NVidia always responds to significant pricing challenges.

What exactly is achieved if AMD releases a $600 2080Ti equivalent, and NVidia then drops their price into the same ballpark? AMD just ends up losing a pile of money needlessly.

 

[nEo717]

I always wonder why people think NVidia won't respond in kind, and just leave their prices high and let AMD take the market share?

AFAICT, Unlike Intel, NVidia always responds to significant pricing challenges.

What exactly is achieved if AMD releases a $600 2080Ti equivalent, and NVidia then drops their price into the same ballpark? AMD just ends up losing a pile of money needlessly.

-------

Where does the idea of AMD putting out a $600 (or less) Ti or Titan killer come from? Isn't the rumor mill out of AMD suggesting $999 and $799 cards on the way with a possible 3rd card near 5700XT price (pushing 5700XT down another $50 or so)?

When we look at the current batch of Navi cards, yes they are a bit ($50 or so) less than nVidia counter parts - However not half price nor hundreds less - Personally to me there seems to be unrealistic expectations being put on AMD (as the norm I guess) for price and performance.

I'd suspect if AMD really launches like the rumor mills suggest we will see both a RTX Ti Black and Titan RTX Black, perhaps the last cards from nVidia until die shrink later next year - and if we are being silly with the rumors, I'm expecting the Ti Black to launch at $649 and Titan Black at $849 (adds perspective to AMD pricing rumor).

 

[Snowdog]

-------

Where does the idea of AMD putting out a $600 (or less) Ti or Titan killer come from? Isn't the rumor mill out of AMD suggesting $999 and $799 cards on the way with a possible 3rd card near 5700XT price (pushing 5700XT down another $50 or so)?

When we look at the current batch of Navi cards, yes they are a bit ($50 or so) less than nVidia counter parts - However not half price nor hundreds less - Personally to me there seems to be unrealistic expectations being put on AMD (as the norm I guess) for price and performance.

I'd suspect if AMD really launches like the rumor mills suggest we will see both a RTX Ti Black and Titan RTX Black, perhaps the last cards from nVidia until die shrink later next year - and if we are being silly with the rumors, I'm expecting the Ti Black to launch at $649 and Titan Black at $849 (adds perspective to AMD pricing rumor).

Almost every time before a new AMD reveal, there is a group of unrealistic, wishful thinkers, expecting AMD to completely change the price/performance model, of the GPU landscape, and every time AMD prices slightly less than NVidia prices.

You would think after many doses of reality, they would get the message, but wishful thinking is a powerful form of self deception.