TR Forums

Besides disregarding and or alienating home users and enthusiats i mean. After years and billions spent, what do they have to show for it? Was it worth it?

It may not be a direct result but Intel's focus on cost, efficiency and integrated features seems to be working rather well for their desktop CPUs too. It's also looking like Nvidia's Maxwell will have some mobile-themed efficiency enhancements, which equals lower monthly electric bills. Can't think of a good excuse for the other two; AMD is floundering because they made a bad bet on parallelization and Microsoft is struggling with the reality that theyr'e not Master of the Universe anymore.

Microsoft is the only one of the four that could be said to have pursued mobile at the expense of other technologies, though. For the others it was just a new growth opportunity.

NovusBogus wrote:

Microsoft is the only one of the four that could be said to have pursued mobile at the expense of other technologies, though. For the others it was just a new growth opportunity.

I agree. The 3 hardware companies have delivered a benefit of lower power consumption, particularly at idle, which is also beneficial on the desktop in enterprises (e.g. machines left on overnight). Microsoft's tablet strategy on the desktop (i.e. Metro) was too aggressive and Windows RT too neutered. Microsoft's mobile phone strategy is good, IMHO.

The main problem they're all going to run into is the tablet/smartphone market saturation, and thus lower profit margin.

Apple, Google, Samsung, Qualcomm, and other ARM-based manufacturers already got first dibs on the marketshare and major profit margings.

Improving performance/watt is where most of the growth opportunity is. Not just for mobile, but datacenters as well -- all those nifty mobile apps rely on hefty investments in back-end infrastructure, and managing the power budgets of all those datacenters is a pretty big deal. I can certainly understand why that's where the focus has been lately.

I think single-core IPC is hitting a wall (or at least a serious knee in the curve, leading to diminishing returns). Going forward, advances in desktop CPU performance will come mostly from software taking better advantage of multiple cores.

GPU performance still has some room to scale, since the workload is already highly parallel. But with consoles being the lowest common denominator these days, there probably isn't the sort of pressure that there was back in the day to push the state-of-the-art.

UnfriendlyFire wrote:

The main problem they're all going to run into is the tablet/smartphone market saturation, and thus lower profit margin.

Apple, Google, Samsung, Qualcomm, and other ARM-based manufacturers already got first dibs on the marketshare and major profit margings.

Yeah, but they all have to pay ARM a slice of their profits as licensing fees. It may end up being close to a wash.

AMD and nVidia both gained more power management expertise. With high end GPU's being power limited, this has paid off in either lower power consumption for a given performance level or more performance at a given power level. This is arguably something they'd have to do anyway as process shrinks are going to be increasingly rare going forward.

nVidia has some brought forth the Tegra line of SoC's and unlike Intel, isn't afraid to produce an end user device if there aren't enough partner's biting. Shield and their new tablet are examples of this. They did manage to get some SoC contracts with an unusual niche: automobile makers. The disadvantages of Tegra compared to other mobile SoC's don't have the same impact as part of a car's internal system. They may not be a home run success, but the company does have at least something to show for their efforts.

AMD's mobile efforts on the CPU side won them console contracts with MS and Sony. Definitely a win there for the company not a very visible one. They've wisely sat out the smart phone race as they'll gladly jump in if Intel can make a foot hold for x86 n that eco system. However, AMD is putting effort into ARM SoC's now so they may not even attempt to move into the smart phone area with an x86 based chip. Moving ARM into the server market is also something that they can show for their mobile efforts, though the verdict is out if this will be successful.

Intel has the least to show for their mobile efforts. Certain their tech demos are impressive but Intel needs to push them into actual products consumers purchase. Otherwise Intel is going to continue to burn cash. This is Intel's ultimate problem: they're not the ones designing the end product but rather a component that goes into the end product. This problem has been extended by Intel designing SoC's to what they think the market wants in an SoC and trying to sell it to OEM's where as other SoC makers are more will to design a custom chip tailored for a handset maker. The only other thing Intel has to show is that they've driven down average power consumption on PC's. That isn't a bad thing but only plays a real role in the ultra book category (which again suffers from poor OEM designs). Intel needs to do some saber rattling to get the likes of HP, Dell, Lenovo and Toshiba to produce noticeably better laptops by fielding reference designs to reporters. Nothing like holding OEM's to the fire and letting the world know what they could be doing better.

MS of course has Windows 8, Surface and Windows Phone to show for their mobile efforts. None of these have been widely successful for various reasons. Yet they've kept MS in the game which at times has simply been MS goal for their products (Windows Phone 7 for example). They also have the most to lose in the game as the company is dependent on Windows and Office licensing. This could also have a domino effect by effecting sales of their server software. They also have the self inflected wound of Office for iPad. Sure, it protects their Office revenue by going to a subscription model on a new popular platform but it sacrifices their Windows licensing. I have a very strong feeling that MS will regret releasing Office for the iPad long term if they really want their vision of Windows everyone to take hold. MS recently purchased Nokia's phone division and got a new CEO so things will definitely be changing for MS going forward.

the wrote:

They also have the self inflected wound of Office for iPad. Sure, it protects their Office revenue by going to a subscription model on a new popular platform but it sacrifices their Windows licensing. I have a very strong feeling that MS will regret releasing Office for the iPad long term if they really want their vision of Windows everyone to take hold. MS recently purchased Nokia's phone division and got a new CEO so things will definitely be changing for MS going forward.

It's a bit of a risk, but I think more importantly we're in a BYOD world now. Microsoft knows that a lot of people have iPads, and a lot of people are going to continue to have iPads (same with Android). Giving them Office as an option means they're more likely to use it, rather than look for something else (and don't forget, Office is the biggest money-maker at MS). I don't think you can argue it's going to hurt their Windows revenue much; most people who are going to go iPad already have, they're not waiting around for Office on iPad to switch over.

Microsoft haven't acheived very much at all. AMD, Intel and Nvidia on the other hand are making massive progress:

Fast-enough processors that can run for a whole day on a single, lightweight battery.
IGP's that actually run stuff reliably and capably.
dGPUs that provide gaming power without needing a 5lbs power brick and three-inch laptop chassis to keep cool.

The progress has been slow and incremental but mobile computing is waaaaaaay better than it used to be.

the wrote:

This problem has been extended by Intel designing SoC's to what they think the market wants in an SoC and trying to sell it to OEM's where as other SoC makers are more will to design a custom chip tailored for a handset maker. The only other thing Intel has to show is that they've driven down average power consumption on PC's. That isn't a bad thing but only plays a real role in the ultra book category (which again suffers from poor OEM designs). Intel needs to do some saber rattling to get the likes of HP, Dell, Lenovo and Toshiba to produce noticeably better laptops by fielding reference designs to reporters. Nothing like holding OEM's to the fire and letting the world know what they could be doing better.

AMD has even worse problems with the OEMs. Have you seen the lineup of Richland and Kaveri laptops?

Intel is able to get away with it because the OEMs crank out so many laptop models that at least several of them are worth considering. AMD can't because they're stuck with a handful of laptop models.

Arclight wrote:

Besides disregarding and or alienating home users and enthusiats i mean. After years and billions spent, what do they have to show for it? Was it worth it?

Ask Apple and Google if it was worth it. That is the motivation. Chasing the money. Its an incontrovertible law of business: copy the other guys who are succeeding. And it isn't over yet. Because the first person there, isn't always the most successful. See: Palm, RIM, MySpace, etc.

All of them gained valuable experience in certain do's and don'ts. Who will easily shake off the "don'ts" they have committed is yet to be seen.

absurdity wrote:

the wrote:

They also have the self inflected wound of Office for iPad. Sure, it protects their Office revenue by going to a subscription model on a new popular platform but it sacrifices their Windows licensing. I have a very strong feeling that MS will regret releasing Office for the iPad long term if they really want their vision of Windows everyone to take hold. MS recently purchased Nokia's phone division and got a new CEO so things will definitely be changing for MS going forward.

It's a bit of a risk, but I think more importantly we're in a BYOD world now. Microsoft knows that a lot of people have iPads, and a lot of people are going to continue to have iPads (same with Android). Giving them Office as an option means they're more likely to use it, rather than look for something else (and don't forget, Office is the biggest money-maker at MS). I don't think you can argue it's going to hurt their Windows revenue much; most people who are going to go iPad already have, they're not waiting around for Office on iPad to switch over.

Short term it is harder to see if Office for iPad has any impact on Windows sales. The problem is that Windows itself is going through a transition with XP support finally ending and Windows 8 having difficulties in the market by itself. These two thing do give tablet markers an oppertunity to move in and take market share away from traditional PC's. This is exactly what has happened as the PC market is shrinking but slowly. When major corporations start deploying iPads as standard devices that is when the big shift will happen. MS should have only released Office for iPad after it was clear that there would be such a change in direction. In other words, Office for iPad should have been a plan B scenario with Windows + Office as plan A. I just don't see the data that'd merit this change in plans (yet).

Arclight wrote:

Besides disregarding and or alienating home users and enthusiats i mean.

I don't know how they've alienated anyone.

Arclight wrote:

After years and billions spent, what do they have to show for it? Was it worth it?

Absolutely. Instead of holding total power usage constant and increasing performance, they've been holding performance relatively constant and lowering power usage. And that performance might as well statistically be good enough for just about anything.

Intel and Nvidia have the most to show for it, obviously.

Haswell, while not light years beyond Ivy/Sandy in IPC, enjoys a brutal performance per watt and performance per die area versus the competing AMD chips. AMD cannot compete here and has nothing to show for it.

Nvidia has probably gained the most. Kepler was their Conroe, Maxwell is looking to be their Nehalem or Sandy Bridge. The efficiency from GF110 to GK104 was nothing short of a miracle, and now the efficiency of GK110 to GM204 will repeat that. AMD cannot compete here either. Hawaii and Tahiti are too power hungry, too hot and just plain don't make good use of the resources they do have. Pitcairn and Bonaire are much more competitive with their Kepler counter-parts, but get annihilated by Maxwell. The upcoming Tonga may improve efficiency somewhat, but I doubt it will be competitive with GM204/GM206, either in terms of efficiency or performance. Again, AMD has nothing to show for it.

As for Microsoft. They are the only game in town. OS X, iOS, Linux, UNIX and Android have all attempted to dethrone Windows but have utterly failed. Active Directory, MS Office, Exchange, SQL Server, Sharepoint, those are industry standards. They are the only players. No *NIX or Crapple alternative has a chance of competing.

AMD has learned a lot about power efficiency over the last 3 years, and it is slowly starting to show up in their chips. Intel and Nvidia do have the lead in this area. But AMD will get very close to them over the next 2 years.

As far as money goes, so far for Intel, tablets and phones have been a multi-Billion dollar bottomless sink hole. The design wins they have managed to get are so few and far between, that its going to pretty much take a miracle for them to reach profitability in the tablet and/or phone markets. But since Intel has incredibly deep pockets, you can almost never count them out. They also have some of the best chip designers on the planet. 5 more years, and Intel could be a huge player in that market. Their x86/x64 chips will keep them afloat for decades to come though.

AMD probably saved itself by getting the gaming console contracts with MS and Sony. That gives AMD some breathing room to work on getting the power reductions into new chips and start working on more wins for those products. And by AMD not trying to lead the world into 20nm and smaller tech this time, they will save huge amounts of money. They still were not profitable in this last quarter, but the losses seem to be shrinking. If they keep executing on a timely manner over the next 2 years, they may well become a money making entity. Compute is an area where I think AMD is going to make fortunes over the next 5 years. Their NextGen compute technology is very efficient, and they are putting it into as many products as they can, as fast as they can, ranging from sub-$100 consumer chips all the way up into their high end boards.

Nvidia is a cash cow. They have been smart, and they have been lucky. They decided to focus on power a good 2 to 3 years before AMD did. They were pretty much forced to by producing GPU's that you could grill steaks on, but sometimes thats what it takes to wake a company up. They've built chips specifically aimed at the table and phone markets, but with mixed responses from OEMs. But their high end computing products have worked out very well for them.

Microsoft has been the biggest loser in the tablet and phone business. 3 full versions of Windows Phone, and they have less than 3% market share. Metro was and will continue to be a complete disaster for them. People have had years now to look at metro, and they have soundly rejected it on every platform. And yet Microsoft keeps pushing it. Once the horse is dead, bury it. Painting the back end and trotting it back out again is tantamount to putting lipstick on a pig. No matter how much lipstick you put on metro, it will still be a dead pig. Windows 9 will go over better on the desktop, but phones and tablets are not going to change enough to make a difference. Heads have already rolled at Microsoft over Windows 8, and I predict that even more heads will roll before someone has the finally has the balls to kill off Metro. Once that happens, everything about tablets and phones for Microsoft becomes a humongous question mark.

Chip makers all over the work are working on new chips for the mobile and handheld markets. I think we can assume that for the most part, ARM chips will continue to rule the roost. But new chips will constantly be trying to chip away at the ARM lead, and it is inevitable that some of those will take some market share away from ARM. But this will happen slowly and I don't see any huge changes here over the next 5 years.

Its a wide open world out there. With optical chips apparently nearing release from a British company, and quantum computing seemingly on the horizon, there are going to be a lot of challenges for the companies that have been leading the industry. And tons of opportunity for new companies with fresh ideas and technologies.

Ballmer got the Clippers out of it.