[Digital Trends] The end of the battle between AMD and Nvidia won’t be great for PC

[master_shake_]

... well of course I'm talking about Nvidia; why would I be referencing AMD drivers in a post about already being in the "post AMD GPU market".

The point of my post is that Nvida, much like Intel, has already started to behave like AMD doesn't exist because they know AMD is in a bad state and no longer poses any real threat to them.

my apologies.

 

[Ramses]

Without AMD there will still be progress, but it will be a scheduled performance increase using economically optimized metrics and timelines. In other words, it will be X% of performance increase every year and Y price for each bracket. The pricing will be less volatile than it is now, to our detriment. There will still be innovation, because nvidia wants people to buy the next card, but it will be just enough innovation to move that needle instead of having to leap frog another company competing for those same sales.

It is kind of like where intel is now. Tick tock, boring performance increases.

Very. :thumbsup:


And GPU manufacturers have the ability, as we've seen, to not update or even cripple previous card drivers to enhance the desirability of newer cards even if they aren't huge increases in raw performance. There are tons of ways for no-competition markets to suck, people are mighty creative when it comes to extracting money.

 

[imaheadcase]

Just capitalism folks. This is how it is in every business, the biggest win.

It does not matter if one has better features, or if one is even %1 faster. It all comes down to marketing and money. I don't know what it is, but you can bet Nvidia marketing just blows away AMD marketing budget. They can also absorb losses easier than AMD.

I feel AMD will die if it fails on one launch..that is really all its taking right now. I don't care if they do, if people want drive to do better, do something better.

 

[Head1985]

Worst think they could ever do
I will agree for R9 290/X but HD7970 stock cooler was more than fine.

Temperatures was OK, but noise was bad
http://www.computerbase.de/2012-03/test-nvidia-geforce-gtx-680/16/
And 7970Ghz was even worse
http://www.computerbase.de/2012-06/test-amd-radeon-hd-7970-ghz-edition/8/

 

[Ketchup]

What upsets me is that AMD cried for years about Intel having all the CPU market even though they had a better chip. They get some headway and what do they do? Stop trying, making only hot, average-speed (at best) products.

Now we see the same thing happening with video cards: ATI was making good high-end cards until AMD bought them out, now they make cards that barely keep up, taking twice us much power do to so.

I hope Samsung is able to take them over before they further run the name into the ground (I still can't believe that Toshiba is keeping the OCZ name alive).

 

[AtenRa]

Temperatures was OK, but noise was bad
http://www.computerbase.de/2012-03/test-nvidia-geforce-gtx-680/16/
And 7970Ghz was even worse
http://www.computerbase.de/2012-06/test-amd-radeon-hd-7970-ghz-edition/8/

I dont remember anyone complaining for the HD7970 and GE default cooler.

HD7970 had lower temperatures with higher noise than GTX680, you could lower the noise and reach same temps as GTX680 if you wanted. Not to mention that the vast majority of users went for the after market cards from ASUS, MSI, Gigabyte etc etc. My ASUS HD7950 CU II even at 1100MHz has very low noise levels.

 

[redzo]

As many posters in this forum have commented, the 390x could be faster, cooler and more efficient than the Titan X and no one would buy it because it isn't Nvidia.

Those customers that are choosing products this way are moron customers. I'm optimistic and I think that idiots buying products this way do not have a major market influence. The 290x wasn't that bad. Even after AMD's screw-up with the cooling, reviews weren't that bad. It was by far to good of a GPU to be pushed back by a failed stock cooling. Those custom cooling designs that followed next are a proof of this. The 7970/Ghz and 290/x were highly competitive offerings.

The real battle takes place at 280x/290/290x-960/970/980 performance and price levels right now. Titan is a product that no educated buyer should care about in my opinion. And no educated buyer should care about 390x neither if its price/performance ratios is going to suck - no mater how fast or a little slower it is compared to titan: titan's price/performance ratio is out of this world like it has always been, it's just a niche market.

It is going to be interesting if AMD will manage to push 280x/290/290x performance a bit while reducing power consumption. That's the only thing that AMD needs to do in order to be highly competitive. If AMD will bet everything on 390x's success while not refreshing 280x/290/290x, even if 390/x is going to be competitive with nvidia's finest titan, that's a very big mistake in my opinion.

 

[AtenRa]

Now we see the same thing happening with video cards: ATI was making good high-end cards until AMD bought them out, now they make cards that barely keep up, taking twice us much power do to so.

Seams you dont remember thinks correctly,

HD5870 was faster and with lower power consumption than NVDIA for 6 months. HD5990 was the fastest Graphics Card for more than 18 months. That was under AMD.

HD7970 was first with 28nm three months earlier than NVIDIA GTX680. HD7970GE was faster than GTX680 until Titan was released. That was under AMD.

R9 290X was faster than GTX780 and today faster than GTX780Ti in many games. Under AMD.

And we wait for R9 390X and the rest of the 3xx series, all under AMD.

 

[Ketchup]

...HD5990 was the fastest Graphics Card for more than 18 months. That was under AMD....

Nope, ATI
http://forums.anandtech.com/showthread.php?t=2060963

I don't have time to check the rest.

 

[AtenRa]

Nope, ATI
http://forums.anandtech.com/showthread.php?t=2060963

I don't have time to check the rest.

AMD Bought ATI in the end of 2006. AMD Killed ATI logo in August 2010.

HD5xxx series was released in September of 2009 with the ATI Logo but the company was under the AMD. Same with HD5990 that was released a few months before AMD killed the ATI Logo.

 

[Headfoot]

What upsets me is that AMD cried for years about Intel having all the CPU market even though they had a better chip. They get some headway and what do they do? Stop trying, making only hot, average-speed (at best) products.

Now we see the same thing happening with video cards: ATI was making good high-end cards until AMD bought them out, now they make cards that barely keep up, taking twice us much power do to so.

Wow, talk about over exaggeration.

1) 290x wins at 4k vs 980, for $225 less.
2) 290x doesn't use twice the power. Post a review that shows that. It uses at most 80 watts more than a 980 and usually less than that (aftermarket, e.g. the only ones you can buy new that people actually buy). How many years will it take to break even from power usage 290x vs 980? Many, many times longer than the useful life of the card.
3) You can crossfire 290s for less than a single GTX 980 and it's a lot faster
4) A single 290 is +10 to -10% slower (reso dependent) than a 970 for $50-60 less.
5) A 295x2 costs $600 and is faster than the $1000 Titan X, sometimes substantially.
6) A 295x2 costs $600, and is over 50% (avg) faster than a GTX 980 which only costs $50 less.

And this is with last-gen cards when we're a month off from the new ones dropping...
If you think that's barely keeping up, you have a very inaccurate definition of "barely keeping up."

 

[PliotronX]

The notion that competition is the saviour of everything is wrong.

No innovation, no buy, too high prices, no buy. The industry is based on high cashflow and high R&D requiring high volume sales. So the companies would simply go bankrupt even as a monopoly if they changed.

However it would be bad as such for potential diversity. But that itself can be both good and bad.

That does not seem true at all to me. Just look at the sorry state of American broadband. Lack of competition means there is no sense of self regulation. I deal with a lot of specialized software that nobody has ever heard of because it normally does what no other software does and it usually has the poorest coding and support whereas areas of software that have a lot of competition have better pricing, quality, and support for end users because they can lose customers to competitors (even Microsoft has its competitors).

 

[ShintaiDK]

That does not seem true at all to me. Just look at the sorry state of American broadband. Lack of competition means there is no sense of self regulation. I deal with a lot of specialized software that nobody has ever heard of because it normally does what no other software does and it usually has the poorest coding and support whereas areas of software that have a lot of competition have better pricing, quality, and support for end users because they can lose customers to competitors (even Microsoft has its competitors).

Broadband doesnt fit into the criteria as described. US healthcare for example is another case with subpair performance and extremely high price. And in both cases regulation is the fix, not competition as such. In both cases you could just move to another place/country for better and cheaper service.

In terms of the PC segment. Volume is the limit. Right now PC volume is about stagnant. Design cost keeps spiralling up. So what would you like to see? Is multiple companies in "competition" delivering mediocore products the best solution? While each one of them is on the edge of giving up every day and unable to deliver any real progress. Or is it better to consolidate the revenue for the fixed volume into something that can bring you further.

And then we exclude all the issues with competition that work against one another and dont want to take any risks that could ruin their lifeblood.

 

[AnandThenMan]

Broadband doesnt fit into the criteria as described.

It most certainly does, it's a great example in fact.

 

[destrekor]

It most certainly does, it's a great example in fact.

Actually, broadband is kind of a special case.

America botched it from the start with regulation that essentially allowed defacto monopolies in any given location. Consumers can't really vote with their wallet all that effectively because they are limited based on where they live. Sure, you might be able to choose satellite over cable, but that's hardly an answer to most people who expect reliability and low latency.

If this net neutrality regulation holds, I think there is a chance that the non-typical broadband providers can get pole access just as fairly as standard carriers (think Google Fiber). I'm not sure if that regulation will help provide challenge to these local monopolies, but I sure wish something would.

 

[ShintaiDK]

It most certainly does, it's a great example in fact.

Not at all. It doesnt require massive R&D or cashflow. And its location determined. Plus I can show you monopoly locations with faster internet than the ones with competition.

 

[Ramses]

http://www.hardocp.com/news/2015/03/23/att_drops_gigapower_prices_in_raleigh#.VRXK0PnF9zo

 

[ShintaiDK]

Actually, broadband is kind of a special case.

America botched it from the start with regulation that essentially allowed defacto monopolies in any given location. Consumers can't really vote with their wallet all that effectively because they are limited based on where they live. Sure, you might be able to choose satellite over cable, but that's hardly an answer to most people who expect reliability and low latency.

If this net neutrality regulation holds, I think there is a chance that the non-typical broadband providers can get pole access just as fairly as standard carriers (think Google Fiber). I'm not sure if that regulation will help provide challenge to these local monopolies, but I sure wish something would.

Exactly. They simply abused a monopoly position, that is normally illegal. On the concept that its hard for you to move location.

 

[ShintaiDK]

http://www.hardocp.com/news/2015/03/23/att_drops_gigapower_prices_in_raleigh#.VRXK0PnF9zo

Still very expensive

And both Google and AT&T spies on you, unless you pay more. Something only AT&T offers.

If competition worked, New York City etc would have 1gbit everywhere for 20-30$. In South Korea you get 1gbit for 20$ for example. In Japan you pay 50$ for 2Gbit.

 

[Ramses]

Just throwing it out there. Competition within reason helps more than hinders ime.

 

[Rezist]

I really don't like nVidia's games works and the state of games that release with it. With AMD leaving maybe it would be a little less toxic and more focused. Still I think PC gaming is getting a little too political, as long as the PS5/Xwhatever can goto multiple vendors come next go around I'll be ok.

 

[ShintaiDK]

Just throwing it out there. Competition within reason helps more than hinders ime.

It depends, but competition is usually the best yes.

However some cases the endgame doesnt allow it. Or other cases competition is a race to the bottom. Or the cost of change is too high. And then there is simply cases where the companies dont want to change on their own or invest in a market because they already got money in the "old". Thats why so much tax money is spend on promoting this until it gets big enough for companies to want to work with.

 

[Rezist]

Still very expensive

And both Google and AT&T spies on you, unless you pay more. Something only AT&T offers.

If competition worked, New York City etc would have 1gbit everywhere for 20-30$. In South Korea you get 1gbit for 20$ for example. In Japan you pay 50$ for 2Gbit.

Internet should be a utility, honestly there's no reason for what we have now, you wouldn't let a private corporation run your water.

 

[swilli89]

What upsets me is that AMD cried for years about Intel having all the CPU market even though they had a better chip. They get some headway and what do they do? Stop trying, making only hot, average-speed (at best) products.

Now we see the same thing happening with video cards: ATI was making good high-end cards until AMD bought them out, now they make cards that barely keep up, taking twice us much power do to so.

I hope Samsung is able to take them over before they further run the name into the ground (I still can't believe that Toshiba is keeping the OCZ name alive).

Dude.. "AMD cried for years...". A Federal court found Intel GUILTY of anti trust business practices. So beyond your questionable terminology in saying a publicly traded company was "crying", their complaints were VALIDATED IN COURT.

Intel was the one that was miss stepping by choosing netburst as its strategy. Intel solved this and started performing much better. AMD continued to keep its R&D budget the same years later. You are overly simplifying the business environment and greatly exaggerating power use between AMD and nvidia cards.

AMD owned ATI when it was the efficient 5870 vs GTX 480, but thats not very convenient for you as it doesn't fit your overly simplified model.

 

[ShintaiDK]

Internet should be a utility, honestly there's no reason for what we have now, you wouldn't let a private corporation run your water.

Oh I agree.

Things you essentially cant bargain with for different reasons and always need should never be in private hands. Internet, infrastructure, medical, district heating, power, water etc. Its also very ineffective, specially today. Companies tend to have a very short investment sight. Its all about making money now. The semiconductor industry is better than the average market here.

The ROI on laying down fiber is 20-30 years. Maybe as low as 10 years if you can get it done while someone else is digging (Power, water, district heating etc.). Again, look on the semiconductor business. If any of the CEOs doubled the R&D budgets today, the result would first materialize around the year 2020. And thats where modern business fails. Imagine going out to shareholders today. Saying you gonna spend billions on something they have to wait countless years on. This is also why result based bonuses is the worst thing to exist.