AMD May Not Be Trying To Compete With Nvidia

[Oshino Shinobu]

Didn't AMD say that they didn't want to be the value option? They've gained a reputation for being good price/performance in the low to mid tier GPU market, but they've said it's not what they want to do. 

With the improvements in iGPUs, AMD probably doesn't have too long until their GPU development becomes obsolete if they just focus on lower tier cards. From how I see it, they basically have to compete with Nvidia or go in a completely different direction with their GPU development if they want that part of the company to stay somewhat relevant. 

[Sakkura]

1 minute ago, FirstArmada said:

 

You could read the Ars Technica article, if you want a TL:DR read mine or skip to the paragraph in the Ars Technica article that starts with "That competition will come..." on the first page 

Well that's a bit nonsensical since it seems pretty sure that high-end Pascal isn't launching until next year. Which is when AMD is supposed to launch Vega. So it looks like a perfect repeat of the 28nm era to me. First medium-sized GPUs going up against each other, then later the big GPUs once yields are good enough (and in this case, once HBM2 production has ramped up). Remember when the 7970 cost $550?

[FirstArmada]

1 minute ago, Belgarathian said:

Very true... Besides, it's a whole lot easier for a company who invests heavily in R&D and plays in the top-performance market share to impinge on mid-low performance tiers than it is for mid-low to impinge on top-performance tiers. 

 

Intel vs AMD, Budget Airlines vs Premium Airlines. 

 

I advise you watch the Master Plan series as i advised aisle9 to Nvidia wont just give up that market share if that AMD does go through with what is specified in that video Nvidia will be essentialy pushed out of the mid to low end market and wont have any way of getting back into it with out lots and lots of cash well over what they are currently spending on R & D

[Sakkura]

1 minute ago, FirstArmada said:

I advise you watch the Master Plan series as i advised aisle9 to Nvidia wont just give up that market share if that AMD does go through with what is specified in that video Nvidia will be essentialy pushed out of the mid to low end market and wont have any way of getting back into it with out lots and lots of cash well over what they are currently spending on R & D

The Master Plan series looks like a bunch of rambling speculation to me.

[FirstArmada]

2 minutes ago, Sakkura said:

Well that's a bit nonsensical since it seems pretty sure that high-end Pascal isn't launching until next year. Which is when AMD is supposed to launch Vega. So it looks like a perfect repeat of the 28nm era to me. First medium-sized GPUs going up against each other, then later the big GPUs once yields are good enough (and in this case, once HBM2 production has ramped up). Remember when the 7970 cost $550?

 

I think Vega will be another card that doesn't compete with Nvidia but instead another card in between what Nvidia offers much like the Fury X , but i guess we wont know until we get closer to its launch 

[Sakkura]

Just now, FirstArmada said:

 

I think Vega will be another card that doesn't compete with Nvidia but instead another card in between what Nvidia offers much like the Fury X , but i guess we wont know until we get closer to its launch 

The Fury X used a big GPU meant to compete head to head with the GTX Titan X and GTX 980 Ti.

[FirstArmada]

4 minutes ago, Oshino Shinobu said:

Didn't AMD say that they didn't want to be the value option? They've gained a reputation for being good price/performance in the low to mid tier GPU market, but they've said it's not what they want to do. 

With the improvements in iGPUs, AMD probably doesn't have too long until their GPU development becomes obsolete if they just focus on lower tier cards. From how I see it, they basically have to compete with Nvidia or go in a completely different direction with their GPU development if they want that part of the company to stay somewhat relevant. 

 

I wouldn't say they are becoming the value option . much like in the world of cars a brand new 2017 Civic isn't considered value it would still be considered mid tier just like a 2017 Mercedes E Class would be considered high end 

[FirstArmada]

1 minute ago, Sakkura said:

The Fury X used a big GPU meant to compete head to head with the GTX Titan X and GTX 980 Ti.

 

The Fury x was priced lower then the 980 Ti and gave less performance whether it was intentional or not either way AMD created a new tier that didnt compete with a 980 Ti untill DX12 came out now the Fury x beats the 980 Ti in some benchmarks but it is still much cheaper  

[FirstArmada]

5 minutes ago, Sakkura said:

The Master Plan series looks like a bunch of rambling speculation to me.

 

It is all just speculation but it makes sense if it were to happen , we just play the waiting game now 

[Prysin]

27 minutes ago, Enderman said:

a random person's speculation is not "news" no matter how logical it is

 

its all just speculation and imagination

 

people cant predict the future, and he has no affiliation with AMD, so we can assume he knows absolutely 0 facts

 

 

he may still be right. After all, there is way more money in the 150-350 USD segment then the 350 USD+ segment. This is a known fact.

Us enthusiasts may only cater to 390Xs or 980s and better, but enthusiasts are within the smallest segment there is.

 

Going for the kill in the 150-350 USD segment will by all means have a higher impact on both companies bottom line.

[Sakkura]

8 minutes ago, FirstArmada said:

 

The Fury x was priced lower then the 980 Ti and gave less performance whether it was intentional or not either way AMD created a new tier that didnt compete with a 980 Ti untill DX12 came out now the Fury x beats the 980 Ti in some benchmarks but it is still much cheaper  

The Fury X and 980 Ti launched at exactly the same price, $650. The Fury X was generally on par with the 980 Ti at launch, it just didn't overclock as well. We can't draw final conclusions on DX12 performance because of the limited number of games supporting the API, but it does look like it boosts AMD performance.

6 minutes ago, FirstArmada said:

 

It is all just speculation but it makes sense if it were to happen , we just play the waiting game now 

At least it's only a few more months before we get some actual hardware to look at.

[PlayStation 2]

6 minutes ago, FirstArmada said:

 

The Fury x was priced lower then the 980 Ti and gave less performance whether it was intentional or not either way AMD created a new tier that didnt compete with a 980 Ti untill DX12 came out now the Fury x beats the 980 Ti in some benchmarks but it is still much cheaper  

Fury X's retail price was $649. 980 Ti's retail price is also $649. Both cards usually are on par if we're talking stock cards. The 980 Ti tends to pull ahead because it has the ability to overclock far better than the Fury X, and has more VRAM, whether it's negligable in many cases or not.

[FirstArmada]

1 minute ago, Sakkura said:

The Fury X and 980 Ti launched at exactly the same price, $650. The Fury X was generally on par with the 980 Ti at launch, it just didn't overclock as well. We can't draw final conclusions on DX12 performance because of the limited number of games supporting the API, but it does look like it boosts AMD performance.

At least it's only a few more months before we get some actual hardware to look at.

 

1 minute ago, Dan Castellaneta said:

Fury X's retail price was $649. 980 Ti's retail price is also $649. Both cards usually are on par if we're talking stock cards. The 980 Ti tends to pull ahead because it has the ability to overclock far better than the Fury X, and has more VRAM, whether it's negligable in many cases or not.

 

Wasnt aware since the Fury Xs release in Canada it has allways been 100$ cheaper then the 980 ti

[PlayStation 2]

Just now, FirstArmada said:

 

 

Wasnt aware since the Fury Xs release in Canada it has allways been 100$ cheaper then the 980 ti

ah

[Belgarathian]

22 minutes ago, FirstArmada said:

 

I advise you watch the Master Plan series as i advised aisle9 to Nvidia wont just give up that market share if that AMD does go through with what is specified in that video Nvidia will be essentialy pushed out of the mid to low end market and wont have any way of getting back into it with out lots and lots of cash well over what they are currently spending on R & D

That's BS. We both know it. 

 

To be the leading manufacturer in the mid range (or bulk of the market share for dGPUs) they would have to be the most power efficient, and cost effective an you don't get to both those positions without being the leader in R&D. If you're the leader in R&D, chances are you're also manufacturing the top-tier GPUs as well. 

 

Most mid-low range manufacturers in other industries are just improving the manufacturing processes for old technology so they can minimise their costs & overheads so they can sell their products at low margins to make them affordable, cost effective options for OEMs and consumers on a tight budget. This process doesn't work in the tech space as mobile technology requires that you stay on top of efficiency improvements, while reducing the cost per unit. All this requires huge spend on R&D, which invalidates your argument. 

 

AMD may focus on the med-low range dGPUs but they'd have to be capable of producing the best top-tier GPU to do so. Like wise, if Nvidia is only going to produce top-tier GPUs it would be very easy for them to re-hash an older GPU to compete effectively in the mid-low range. 

 

To only focus on the mid-low end of the market would be akin to nailing their own coffin shut. 

 

 

[Guest]

I think AMD can bring high end GPUs to the market.....they did that with the R9 390 and Fury X...I think that if they really focused on their GPU market, they could probably beat Nvidia

[FirstArmada]

17 minutes ago, Belgarathian said:

That's BS. We both know it. 

 

To be the leading manufacturer in the mid range (or bulk of the market share for dGPUs) they would have to be the most power efficient, and cost effective an you don't get to both those positions without being the leader in R&D. If you're the leader in R&D, chances are you're also manufacturing the top-tier GPUs as well. 

 

Most mid-low range manufacturers in other industries are just improving the manufacturing processes for old technology so they can minimise their costs & overheads so they can sell their products at low margins to make them affordable, cost effective options for OEMs and consumers on a tight budget. This process doesn't work in the tech space as mobile technology requires that you stay on top of efficiency improvements, while reducing the cost per unit. All this requires huge spend on R&D, which invalidates your argument. 

 

AMD may focus on the med-low range dGPUs but they'd have to be capable of producing the best top-tier GPU to do so. Like wise, if Nvidia is only going to produce top-tier GPUs it would be very easy for them to re-hash an older GPU to compete effectively in the mid-low range. 

 

 

 

I just read that and is a tad to much for my 14 year old brain to comprehend all i got out of that is you cant own the majority of the mid range market with out being most power efficient cost effective spend the most in R&D and manufacturing the top tier GPU do explain in laymen's terms why this is not possible 

 

I dont see why its not possible or why if it was not possible before become possible when AMD pretty much reks Nvidia's performance in all console ported games once Navi releases if that video is to be believed which it very well could be wrong we just have to wait and see 

[FirstArmada]

4 minutes ago, jkeasley said:

I think AMD can bring high end GPUs to the market.....they did that with the R9 390 and Fury X...I think that if they really focused on their GPU market, they could probably beat Nvidia

Im 98% sure Polaris 10 flagship will out preform or preform like a 390x could be wrong once again we just have to wait and see hence the "may" in the title 

[Prysin]

5 minutes ago, Belgarathian said:

That's BS. We both know it. 

 

To be the leading manufacturer in the mid range (or bulk of the market share for dGPUs) they would have to be the most power efficient, and cost effective an you don't get to both those positions without being the leader in R&D. If you're the leader in R&D, chances are you're also manufacturing the top-tier GPUs as well. 

 

Most mid-low range manufacturers in other industries are just improving the manufacturing processes for old technology so they can minimise their costs & overheads so they can sell their products at low margins to make them affordable, cost effective options for OEMs and consumers on a tight budget. This process doesn't work in the tech space as mobile technology requires that you stay on top of efficiency improvements, while reducing the cost per unit. All this requires huge spend on R&D, which invalidates your argument. 

 

AMD may focus on the med-low range dGPUs but they'd have to be capable of producing the best top-tier GPU to do so. Like wise, if Nvidia is only going to produce top-tier GPUs it would be very easy for them to re-hash an older GPU to compete effectively in the mid-low range. 

 

 

you're way too stuck up with your inaccurate ideals.

It is obvious which vendor you are rooting for, unfortunatly for you, in the silicone industry, it isnt unlikely for a manufacturer to suddenly change the game overnight. Because there is so many factors, both in hardware and software.

 

To rule the low/mid range in GPUs you need ONE THING. BETTER PRICE TO PERFORMANCE. All you need is to undercut the competition with equal or better products at equal or lower prices.

 

Look at the current GPU game.

Nvidia has two "wins", the GTX 950 and GTX 980Ti (TitanX is just a show-off)

In every other price category AMD is the "winner" by offering better price to performance.

The 360 => 750TI

The 370 < 950

The 380 > 960 2gb

The 380X >> 960 4gb OC

The 390 > 970

The 390X < 980, but undercuts it by being much cheaper

The Fury/Nano > 980

The FuryX < 980Ti

 

ATM, Nvidia is still selling 900 series cards simply because they got the market share and are "known" for their gaming products. All AMD needs to do is simply overcome the idea that "Nvidia is only option for gaming".

 

If AMD can simply make a name for themselves, they will be MUCH MUCH MUCH more likely to beat Nvidia.

[Belgarathian]

1 minute ago, FirstArmada said:

 

I just read that and is a tad to much for my 14 year old brain to comprehend all i got out of that is you cant own the majority of the mid range market with out being most power efficient cost effective spend the most in R&D and manufacturing the top tier GPU .

 

I dont see why its not possible or why if it was not possible before become possible when AMD pretty much reks Nvidia's performance in all console ported games once Navi releases if that video is to be believed which it very well could be wrong we just have to wait and see 

Hopefully this will help - 

 

The lions share of the market is mobile devices - to play here, you have to be power efficient and cost effective (so you have to be on the newest manufacturing node, possess the most power efficient architecture, and the most efficient manufacturing process with the best yields. 

 

Desktop GPUs aren't quite as bad, as we don't really care about power consumption or TDP so long as we can dissipate it without throttling. This market usually determined by marketing effort, but has a lot to do with price point and the same issue above come into play. If you can manufacture more cost effectively (higher yields, lower cost per unit) then you can sell your cards at a price point your competitors can't afford to. - But to be in that position, you have to be a market leader (which typically means investment in R&D). 

 

Lets say company X is the market leader and they have a card that's a couple of generations ahead of company Y who is manufacturing graphics cards for the average gamer at $300. 

 

company X decides they want to be part of the $300 graphics card segment, and realises that they could rebrand a graphics card they developed 2 years ago to compete with company Y's graphics card. Company X has the advantage, because they're 2 years ahead in the development of the manufacturing process and can likely get better yields, and benefit from reduced costs as the R&D has already been accounted for in the sales from the previous 2 years. 

 

Hope this helps. 

[FruitBasketSilex]

1 hour ago, Sakkura said:

Load of nonsense from a random youtuber. Not even worthy of the news section IMO.

This all fucking day.

[Belgarathian]

12 minutes ago, Prysin said:

you're way too stuck up with your inaccurate ideals.

It is obvious which vendor you are rooting for, unfortunatly for you, in the silicone industry, it isnt unlikely for a manufacturer to suddenly change the game overnight. Because there is so many factors, both in hardware and software.

 

To rule the low/mid range in GPUs you need ONE THING. BETTER PRICE TO PERFORMANCE. All you need is to undercut the competition with equal or better products at equal or lower prices.

 

Look at the current GPU game.

Nvidia has two "wins", the GTX 950 and GTX 980Ti (TitanX is just a show-off)

In every other price category AMD is the "winner" by offering better price to performance.

The 360 => 750TI

The 370 < 950

The 380 > 960 2gb

The 380X >> 960 4gb OC

The 390 > 970

The 390X < 980, but undercuts it by being much cheaper

The Fury/Nano > 980

The FuryX < 980Ti

 

ATM, Nvidia is still selling 900 series cards simply because they got the market share and are "known" for their gaming products. All AMD needs to do is simply overcome the idea that "Nvidia is only option for gaming".

 

If AMD can simply make a name for themselves, they will be MUCH MUCH MUCH more likely to beat Nvidia.

Why would Nvidia drop their price when they're still outselling AMD 2:1 (I don't know the actual figures, but they still hold the majority of new card sales). 

 

You wait until it's 1:1 and AMD starts gaining market share. - $50 says that Nvidia's pricing game steps up a notch. 

 

There is no point giving up margin when you're still meeting targets & KPIs.

[VagabondWraith]

33 minutes ago, FirstArmada said:

 

The Fury x was priced lower then the 980 Ti and gave less performance whether it was intentional or not either way AMD created a new tier that didnt compete with a 980 Ti untill DX12 came out now the Fury x beats the 980 Ti in some benchmarks but it is still much cheaper  

You can find 980 Ti cards for $599. Fury X being around $629. Not only is it cheaper, but more powerful to boot.

[Sakkura]

6 minutes ago, Prysin said:

 

To rule the low/mid range in GPUs you need ONE THING. BETTER PRICE TO PERFORMANCE. All you need is to undercut the competition with equal or better products at equal or lower prices.

 

Look at the current GPU game.

Nvidia has two "wins", the GTX 950 and GTX 980Ti (TitanX is just a show-off)

In every other price category AMD is the "winner" by offering better price to performance.

The 360 => 750TI

Price to performance should be important, but in reality other factors matter more. People who buy value or budget cards are typically not enthusiasts and don't check reviews that much. They just hear that "GTX 980 Ti is awesome" and jump on a 960 because that must be a bit awesome too, but cheaper. 

 

Also, the 750 Ti should be well ahead of the R7 360, since it's generally a little better than the 260X, which is the big brother of the 260 and 360. The latter compete with the basic 750.

[FirstArmada]

1 minute ago, Belgarathian said:

Hopefully this will help - 

 

The lions share of the market is mobile devices - to play here, you have to be power efficient and cost effective (so you have to be on the newest manufacturing node, possess the most power efficient architecture, and the most efficient manufacturing process with the best yields. 

 

Desktop GPUs aren't quite as bad, as we don't really care about power consumption or TDP so long as we can dissipate it without throttling. This market usually determined by marketing effort, but has a lot to do with price point and the same issue above come into play. If you can manufacture more cost effectively (higher yields, lower cost per unit) then you can sell your cards at a price point your competitors can't afford to. - But to be in that position, you have to be a market leader (which typically means investment in R&D). 

 

Lets say company X is the market leader and they have a card that's a couple of generations ahead of company Y who is manufacturing graphics cards for the average gamer at $300. 

 

company X decides they want to be part of the $300 graphics card segment, and realises that they could rebrand a graphics card they developed 2 years ago to compete with company Y's graphics card. Company X has the advantage, because they're 2 years ahead in the development of the manufacturing process and can likely get better yields, and benefit from reduced costs as the R&D has already been accounted for in the sales from the previous 2 years. 

 

Hope this helps. 

 

Ahh i see it now but that doesn't quite relate to the AMD master plan video and what he talks about in their it doesn't matter if Nvidia can make a competitive GPU at that price point they will also be forking a bunch of money on optimization it would make much more sense for them to just stick to the market sector they have be excelling in and only fork out the cash to optimize those cards 

 

secondly it would also cost a bunch of money to refine 28nm lithography to the point were it preforms to a 1st gen 14nm product once again even if they did or could sooner or later they would have to advance to the next fabrication node