It doesn't matter how much Intel want to push fab tech when there is simply no consumer killer app to justify new PCs or Intel CPUs for the vast majority of people. We are now in a situation where even ARM has too much excess general CPU power for this market, computation-intensive H264/HEVC decoding can be handled by the GPU, and game devs on Android/iOS/x86 are not pushing the graphical envelope further due to diminishing returns and increasing costs.
It doesn't matter how much Intel want to push fab tech when there is simply no consumer killer app to justify new PCs or Intel CPUs for the vast majority of people. We are now in a situation where even ARM has too much excess general CPU power for this market, computation-intensive H264/HEVC decoding can be handled by the GPU, and game devs on Android/iOS/x86 are not pushing the graphical envelope further due to diminishing returns and increasing costs.
No. That's been "the future" for 40 years. The reality is that fat clients keep getting better, and servers are inefficient for small-scale tasks (too much reliance on a high-latency network infrastructure, which just shifts the load).
Yes. It uses excessive bandwidth to do that, and the farther you go, and the more devices share the network, the worse actual performance will get. With thin client type usage, as soon as deadlines start to be missed, it's unusable.When I have streamed Steam games over my wired home network, I haven't seen any appreciable lag.
So It must the size and/or type of the network matters.
Sandy Bridge is still a very viable CPU, and that is well over 4 years old is it not? I think it will still be viable for another 3 or so years the way CPU's are moving. Part of the problem is the lack of competition in the desktop space. AMD just aren't competitive so Intel don't have any motivation to release more powerful CPUs.
GPU's aren't really that amazing either
This is a very boring time for desktop computing and I don't see it getting any better.
Performance/watt is the key. And yes the desktop is dead.
Servers love it. Mobile love it. Massive improvements in both.
The only thing that can stop the show is if they hit a barrier in terms of cost/shrinking/changing electrical properties.
I cant see any problem for Intel the next 10 years.
hell, this is the future I want! this and self driving cars are sooo worth waiting for. just 5 more years!Once Atom-like and ARM CPUs exceed Core 2 IPC (in real world use), which is probably going to be in another 5 years or so, SFF PCs will be financially viable for any desktop user, and convertible tablets that are worth using for everything will be cheap.
You can build an ARM desktop for $70-100 that you can do software development on (RPi 2, fully encased and powered, with a sizable and fast SD card). For $150, you can get some that leave the RPi in the dust.
10 years? Do you have any idea of how much the market changes in that time frame? In 2004 Intel had just fallen flat on its face with the launch of Prescott and plenty of people were chiming in that AMD had slain the giant. 90nm process nodes were cutting edge, ATI was an independent company and Nvidia still was wearing the bruises from the FX5000 series. There's no way someone at that point would have guessed that a decade later Nvdia would have been holding the performance crown (and majority of the market), that AMD would be an also-ran in the desktop or everyone would have smartphones in their pockets more potent than the desktops of that era with ample access to 20~30mbit internet connections over the air. If you could accurately predict the future 10 years out you'd be a very rich man, but in reality you're just blowing hot air.
Very few people need desktop?This. Very few people actually NEED a desktop computer. The vast majority only use it for social media and web browsing anyway. The only real difference for most people is that historically, it was a pain in the ass to hook a phone up to a monitor. But with now functionality like Miracast, you can have a movie on your phone or tablet and watch it on a compatible HDTV. Same with web surfing. Add bluetooth keyboard support and there's no real difference - for most people that is.
I think people will still have laptops for a handful of special applications but there won't be many reasons to upgrade unless you're using a machine for work. The only real exception is the gaming market. That seems to be the only area where applications can take full advantage of a desktop with a dedicated GPU.
Very few people need desktop?
PC Gaming industry is growing at an exponential rate and i cannot imagine playing a game on a laptop.
Always hooked to the power outlet, uncomfortable, getting very hot, small screen, overpriced, etc make laptops terrible for PC gaming.
If you are a PC gamer, you have to have a desktop.
PC gamers are in the minority compared to the pc market which include home/casual users and business/office machines.
And that's just the "extreme" gamers, whatever that means. But it is safe to assume they are enthusiasts with really high performance hardware. Then there's a lot of casual / non-extreme gamers too in addition to that."Intel claims something like 10% of the PC market is extreme gamers [...]
I think you missed the context of the thread.
Sure they can fail for whatever reason. But It wont be due to the market itself or limitations of production.
Yup, but that are not that few. Not the 1% that some claim.
Intel for example says:
And that's just the "extreme" gamers, whatever that means. But it is safe to assume they are enthusiasts with really high performance hardware. Then there's a lot of casual / non-extreme gamers too in addition to that.
Freescale has quoted 13.5 SPEC int 2006 for a 2 GHz Cortex-A72. That is more than a Core 2 T7400 running at 2167 Mhz so they already have reached that IPC ;-)Once Atom-like and ARM CPUs exceed Core 2 IPC (in real world use), which is probably going to be in another 5 years or so, SFF PCs will be financially viable for any desktop user, and convertible tablets that are worth using for everything will be cheap.
Yup, but that are not that few. Not the 1% that some claim.
Intel for example says:
And that's just the "extreme" gamers, whatever that means. But it is safe to assume they are enthusiasts with really high performance hardware. Then there's a lot of casual / non-extreme gamers too in addition to that.
For the first time it is supporting the "overclocking" of some of its mobile device chips.
The practice involves pushing processors to perform calculations at a faster rate than they were set to, causing them to give off more heat. In the past, Intel had locked its mobile chips to prevent this.
"Intel claims something like 10% of the PC market is extreme gamers, and that's who it's for," said Jack Gold.
"For those guys and girls who want to cool their laptops with liquid nitrogen [and other substances] and get the maximum they can out of them."
Forget SandyBridge. Even Nehalem/Westmere are still adequate.Sandy Bridge is still a very viable CPU, and that is well over 4 years old is it not? I think it will still be viable for another 3 or so years the way CPU's are moving. Part of the problem is the lack of competition in the desktop space. AMD just aren't competitive so Intel don't have any motivation to release more powerful CPUs.
And that is a good step (just the fast they are good enough to be even be willing to perform SPEC benches on them!), but I'll believe it's just as fast, or faster, when they can come up with less cache-friendly benchmark sets, and show comparisons that are on even planes (Debian or Ubuntu kernels and packages, FI, and identical storage).Freescale has quoted 13.5 SPEC int 2006 for a 2 GHz Cortex-A72. That is more than a Core 2 T7400 running at 2167 Mhz so they already have reached that IPC ;-)
https://www.spec.org/cpu2006/results/res2007q3/cpu2006-20070723-01562.html
http://www.realworldtech.com/forum/?threadid=152735&curpostid=152741
Note the Intel result is old enough that Intel didn't have yet time to implement SPEC specific tuning in their compiler (e.g. for libquantum...).
For someone who's stretching their money, like very small businesses, and non-profits, that's happening. Businesses that really worry about losses from IT-related downtime are upgrading more for newer parts and software than anything else, or to reduce costs (that's generally going on in the server room, though). I still see Core 2s, Athlon64 X2s, Athlon IIs, etc., doing just fine (I use a C2D regularly, too, and it's fine, honestly, as primarily a web browsing box).I think providing 5-8% yearly performance increased may have worked decently for Intel for some years, since the people with 5+ year old computers still got a decent upgrade when upgrading to the latest CPU.
But soon we'll be having a situation where all CPUs that are 5-8 years old are still good enough. Who will upgrade then?
If Intel doesn't find an answer to that question, I think the downwards PC trend will escalate, and Intel will be in big trouble...
As I see it, we're already in the middle of that, and that is a large aprt of the downward trend in PC markets. With high-end/business notebooks, the same long lives can be had in portable units, too.
The typical chiclet keyboards, and lack of RJ45s, are all I really dislike about some of the better newer Ultrabooks, too, TBH. I hated some of the earlier ones, especially any under $1500 or so, but they are pretty durable and serviceable, now, if you get a decent one, which can often be had for $800 or so. You lose storage, keyboard space, and IO compared to a regular notebook, but that's about it (and in terms of keyboard quality, everybody is going to chiclet for their big models, too D.Yup, but Ultrabooks have seen decent improvement in recent years though, mostly on weight and battery life. But they too are stagnating now. So they are just lagging behind desktop PCs by 2-3 years or so with regards to the "no need to upgrade" issue...
But soon we'll be having a situation where all CPUs that are 5-8 years old are still good enough. Who will upgrade then?
If Intel doesn't find an answer to that question, I think the downwards PC trend will escalate, and Intel will be in big trouble...