AMD discrete GPU market share eroded to less than 20 per cent

[Guest Strangerbob]

I dunno. You can keep your discrete gpus....  

 

I want that ZEN with integrated graphics as good as the latest gen x60/x70gpu, with shared integrated HBM ram for both the CPU and GPU portions.

And then bring on the era of powerful gaming laptops and NUC sized gaming PC`s. The era of disgusting huge towers is coming to an end... good riddance.

[marldorthegreat]

I dunno. You can keep your discrete gpus....  

 

I want that ZEN with integrated graphics as good as the latest gen x60/x70gpu, with shared integrated HBM ram for both the CPU and GPU portions.

And then bring on the era of powerful gaming laptops and NUC sized gaming PC`s. The era of disgusting huge towers is coming to an end... good riddance.

Why? Who actually uses an apu? And the amouunt of heat and power would give off and draw would be huge. It would have to be a 200 watt chip

[patrickjp93]

 

One day people will stop claiming that a different cooler does anything at all to the amount of heat produced.

Why stop saying it when it's true? Resistance of any circuit is partially determined by temperature. The hotter you let it run, the more power it uses, diffused as heat by the function of current^2 * resistance of the circuit which = cross-sectional area of the wire * length * resistivity of the material * heat variance function (which is a huge mess).

 

Seriously people, I'm an American and I know this, and I'm not even an electrical engineer. Freshman year physics sequence, AP Physics C. Hell this is basic preparatory physics for the NJ high school system (at least back when I took physics in 2011 and 2012).

[Tomsen]

I dunno. You can keep your discrete gpus....  

 

I want that ZEN with integrated graphics as good as the latest gen x60/x70gpu, with shared integrated HBM ram for both the CPU and GPU portions.

And then bring on the era of powerful gaming laptops and NUC sized gaming PC`s. The era of disgusting huge towers is coming to an end... good riddance.

4 zen core + integrated GPU with DDR4 will be quite refreshing.. but zen APUs are first arriving in 2017 AFAIK..

Carrizo are however a beast with its power efficiency, and is quite suited for laptops. However the availability isn't really there..

[patrickjp93]

Why? Who actually uses an apu? And the amouunt of heat and power would give off and draw would be huge. It would have to be a 200 watt chip

We have AIOs and the NH-D15 for that. Hell IBM already gets away with the 225WW TDP Power 8 flagships (12 cores/96 threads @ 4.3GHz) on 1U cooling systems both on air and on compact water with up to 8 chips per board and have 2 boards directly linked for a 16-chip node in a 2U space.

[Tomsen]

AMD's problem is marketing. The 200 series was largely decent and competitive, yet people just flat out didn't buy it.

One day people will stop claiming that a different cooler does anything at all to the amount of heat produced.

It was quite popular when bitcoins hit the media. However you are quite right..

When did I make such a statement?

[That Norwegian Guy]

yes they would - were Nvidia to be the only discrete GPU manufacturer anti-monopoly laws would kick in forcing them to split the company and get a lot of law suits headed their way.

 

Sadly, there won't be any differentiation on the "discrete" part.

 

The FTC would consider Intel and Nvidia competition in the consumer GPU segment.

 

However, Intel would be forced to split as they'd have a monopoly on personal computer processors.

 

This is why Intel's strategy at the moment is to do nothing, and torture drip-feed the market with useless new chips that aren't upgrades, while Nvidia are actively trying to destroy AMD.

[Clyne]

Horrible business practices, misleading/blatant advertisement about "GAMING", larger R&D budget than AMD.

In no way i'll never buy Nvidia again for 9xx and up unless they brought something like Tesla, Fermi, maybe Kepler back. 

[CydiaDash]

If AMD go bust, imagine how much a 7950 or something like that will be worth in 5 years :0

[Davethegamer]

No they wouldn't be..?

US Monopoly Laws prevent them from having 100% market share they would have to split and that would be real messy.

[patrickjp93]

US Monopoly Laws prevent them from having 100% market share they would have to split and that would be real messy.

FFS they do not. They only prevent you from achieving that 100% market share by anticompetitive practices defined in the laws, from abusing your position as market leader via price gouging/fixing and/or bribing OEMs and/or retailers to not carry competitors' products, and buying up all your competitors to the point you poison the well of competition. Getting a monopoly by earning it through raw competition and winning is perfectly legal in the American system.

[patrickjp93]

Sadly, there won't be any differentiation on the "discrete" part.

 

The FTC would consider Intel and Nvidia competition in the consumer GPU segment.

 

However, Intel would be forced to split as they'd have a monopoly on personal computer processors.

 

This is why Intel's strategy at the moment is to do nothing, and torture drip-feed the market with useless new chips that aren't upgrades, while Nvidia are actively trying to destroy AMD.

No they wouldn't. Intel has no AIB solutions, and in fact it's AMD's and Nvidia's patents which prevent this. Further, integrated and discrete graphics do not compete with each other at all outside business/office computers. Even further, under DX 12 and Vulcan there is multiadaptor support if programmers choose to use it. Intel would not be considered competition to Nvidia in the consumer segment at all except perhaps cellphone and tablet SOCs. The other major area is HPC where you have the Xeon Phi vs. the Teslas.

[MrRavens]

Hold on AMD! D;

[totjup5]

Nvidia hasn't for years now.. This is nothing new.

At least they don't sit on their asses and not do shit quite as much as Intel has the last few years....

[Beskamir]

I like AMD so this is really sad and worrying news for me. Not only does Nvidia get closer to having a monopoly and thus they can pretty much do whatever they want but this also (obviously) won't help AMD with their financing.

[Bob Jim]

The 750Ti was released in early February of 2014.It's and old card now, and it's about to be refreshed.It's also part of the 1st maxwell generation, not 2nd like the 9xx series.

The 960 is worse than AMD's counterpart, that's true.

The 970 and 390 are about equal. You won't notice the difference. Maybe only at higher resolutions. It's all about the price : the cheapest one wins. Dunno about you guys in the US, but the 970 is cheaper than an R9 390 here.

I am in the UK and currently they are the exact same price, but obviously they fluctuate.

[Icecube26]

apparently people here on LTT want so see AMD die ...... so NVIDIA and INTEL can charge way more for their products... go ahead...

[Nineshadow]

FFS they do not. They only prevent you from achieving that 100% market share by anticompetitive practices defined in the laws and from abusing your position as market leader via price gouging/fixing and/or bribing OEMs and/or retailers to not carry competitors' products. Getting a monopoly by earning it through raw competition and winning is perfectly legal in the American system.

In Romania we have 2 big ISPs: UPC and RCS RDS. A few years ago RCS RDS wanted to buy UPC, giving them like 95% marketshare.However, our gov. didn't allow the purchase since it would create a monopoly.

[fatfinguh]

Looks like a comeback is looming.

[patrickjp93]

In Romania we have 2 big ISPs: UPC and RCS RDS. A few years ago RCS RDS wanted to buy UPC, giving them like 95% marketshare.However, our gov. didn't allow the purchase since it would create a monopoly.

Purchases to gain monopolies are also prevented based on a couple simple mathematical functions which determine societal benefit and loss/gain of competition. But the broad brush stroke used was entirely incorrect.

[HarmFull]

Let me just drop  this here:

 

http://www.eurogamer.net/articles/2013-01-16-former-amd-employees-accused-of-stealing-secrets-for-nvidia

 

This article is from Januari 2013, the point where AMD has been losing it big is around the start of 2014! Call me tinfoiled, but AMD has been losing their battle ever since both Nvidia and Intel started taking their employees.

 

Also and I keep mentioning this: AMD's marketing and naming schemes aren't clear to me! R7, R9, Fury / FX, A6, A8, A10, Athlon, Phenom.

 

Now compare that to Nvidia and Intel: GT(X) / Celeron, Pentium, Core M, i3, i5, i7.

 

Also, when you look on YouTube for example, we always see comparisons being made to the Nvidia cards as well as the Intel CPU's!

 

Even with new Nvidia cards, the comparisons are made between last gen cards instead of the AMD cards!

[bogus]

Let me just drop  this here:

 

http://www.eurogamer.net/articles/2013-01-16-former-amd-employees-accused-of-stealing-secrets-for-nvidia

 

This article is from Januari 2013, the point where AMD has been losing it big is around the start of 2014! Call me tinfoiled, but AMD has been losing their battle ever since both Nvidia and Intel started taking their employees.

 

Also and I keep mentioning this: AMD's marketing and naming schemes aren't clear to me! R7, R9, Fury / FX, A6, A8, A10, Athlon, Phenom.

 

Now compare that to Nvidia and Intel: GT(X) / Celeron, Pentium, Core M, i3, i5, i7.

 

Also, when you look on YouTube for example, we always see comparisons being made to the Nvidia cards as well as the Intel CPU's!

 

Even with new Nvidia cards, the comparisons are made between last gen cards instead of the AMD cards!

 

AMD problem is a marketing problem. That's all.

 

AMD with a way lower R&D budget is on par with NVIDIA, and it's bringing new tech to the industry and pushing industry forward in several fronts.

 

Like I said - show us each marketing budget and then we can talk about market share. You see alot of comments here talking literally shit, thinking that the market share = tech superiority lol... and they actually belive in that

 

Ofc such news snowball into kids thinking that one rules over the other because the high end card of NVIDIA has 5 more frames per second on average then AMD high end card.

 

And remember - you can be a grown man and still be a kid

[othertomperson]

 

Why stop saying it when it's true? Resistance of any circuit is partially determined by temperature. The hotter you let it run, the more power it uses, diffused as heat by the function of current^2 * resistance of the circuit which = cross-sectional area of the wire * length * resistivity of the material * heat variance function (which is a huge mess).

 

Wow ok this hinges on a bunch of assumptions. Firstly you're assuming that a CPU, which is a bunch of semi-conductors, behaves identically to a length of copper wire. Huge citation needed here.

 

Secondly you are overestimating the impact of the heat on the resistance by a huge amount. This equation is useful when determining the resistance curve of a lightbulb, ie something getting so hot that it is glowing. It is hundreds of degrees Celsius. For a given clock speed what difference can an aftermarket vs stock cooler actually make? 40 degrees? That is nothing like enough to add any appreciable resistance.

 

No need to point out that you are not an electrical engineer, that's quite apparent.

 

Horrible business practices, misleading/blatant advertisement about "GAMING", larger R&D budget than AMD.

In no way i'll never buy Nvidia again for 9xx and up unless they brought something like Tesla, Fermi, maybe Kepler back. 

 

...you're pissed off that the 900 series was advertised as gaming cards? o.0

[patrickjp93]

Wow ok this hinges on a bunch of assumptions. Firstly you're assuming that a CPU, which is a bunch of semi-conductors, behaves identically to a length of copper wire. Huge citation needed here.

Secondly you are overestimating the impact of the heat on the resistance by a huge amount. This equation is useful when determining the resistance curve of a lightbulb, ie something getting so hot that it is glowing. It is hundreds of degrees Celsius. For a given clock speed what difference can an aftermarket vs stock cooler actually make? 40 degrees? That is nothing like enough to add any appreciable resistance.

No need to point out that you are not an electrical engineer, that's quite apparent.

...you're pissed off that the 900 series was advertised as gaming cards? o.0

This law applies for all resistors, regardless of extra qualities (Halliday, Resnick Fundamentals of Physics 6th Edition). All conductors are resistors unless they are superconductors.

The temperature factor is a quadratic order function. You're underestimating.

[Clyne]

 

 

 

Wow ok this hinges on a bunch of assumptions. Firstly you're assuming that a CPU, which is a bunch of semi-conductors, behaves identically to a length of copper wire. Huge citation needed here.

 

Secondly you are overestimating the impact of the heat on the resistance by a huge amount. This equation is useful when determining the resistance curve of a lightbulb, ie something getting so hot that it is glowing. It is hundreds of degrees Celsius. For a given clock speed what difference can an aftermarket vs stock cooler actually make? 40 degrees? That is nothing like enough to add any appreciable resistance.

 

No need to point out that you are not an electrical engineer, that's quite apparent.

 

 

...you're pissed off that the 900 series was advertised as gaming cards? o.0

 

I'm just pissed that it's straight up blatant stupid advertising.