Do you guys think that companies like Samsung just aren't interested in making x86 cpu's or that Intel won't sell them a license to do so?
AMD went so long with out releasing a new architecture that Intel's current position is damn close to a monopoly.
Other urls found in this thread:
en.wikipedia.org
twitter.com
>intel won't sell samsung x86 license
What the fuck? Just how retarded are you?
Samsung doesn't have the AMD64 tech to offer competition to Intel and AMD.
Clearly I'm extremely retarded because I have no idea what you're trying to say.
I asked why there are only 2 companies making x86. Is it because no one else wants to or because they can't get the license to
Nobody can get "x86 license" from intel you absolute shit mongler faggot nigger.
VIA also makes x86 chips.
x86-16 and x86-32 are both probably free at this point. The problem is the instruction set extensions aren't all free and those can be a huge.
didn't they quit?
I fucking swear every chink produces 8086
Samsung primarily makes mobile devices. Why would they want to make x86 chips when they consume significantly more power than other architectures?
Because as far as Intel's concerned there are already too many companies manufacturing x86 processors.
no
I just checked out their website. Their considering their CPU offerings, I don't understand how they're still in business.
VIA got their license when they acquired Cyrix and they negotiated a cross-licensing agreement with Intel and AMD to get access to their instruction set extensions.
Would a company even have a chance to beat Intel's IPC even if they DID give out AMD64 licenses? Also how the fuck can someone license an instruction set? I get that the architecture itself it protected by copyright, but that this point x86 is nothing but a fancy script language interpreted at machine level.
They sell ARM boards with native SATA 3gbps
I hadn't seen that. I only looked at the x86 offering, nevermind.
OP here it may not be the instruction set. I just know there is some kind of licensing involved. Like you can't just make a x86_64 chip on a whim.
It seems like DIY PC CPU sales are pretty strong I'm just confused as to why it's not more competitive. AMD struggles to compete with nvidia and intel but I would say they do a better job in keeping the GPU market fair than the CPU market.
No point getting into the x86 market now. Might as well wait for something better and be at the front.
Making a competitive x86 chip is a lot of work for not a lot of return, and yes, Intel really isn't into the idea of more competition, AMD and VIA both are grandfathered in, maybe you could read more about Transmeta or something to see how they got around the legal bullshit for their chips.
It still pales in comparison to the OEM market and neither they nor whitebox builders ultimately care for more choice, computers are cheap enough, model lineups are bewildering enough, and CPUs barely matter anymore anyway unless you're a compiler writer or similar poor sod working on keeping everything abstracted away.
there are a few more companies other than Intel and AMD that could legally produce x86 CPUs if they wanted to, but apart from VIA nobody really ever made anything.
>wait for something better
That's never a good idea. Rather they should just see what's going to be the next thing and they go with that.
Not sure what it's going to be, but I'm sure ARM will be related somehow.
What other companies still have an x86 licence?
IBM?
server and supercomputing market
IBM has a license yes, but it's a cuckold one
en.wikipedia.org
I don't see arm going doing anything other than what it already does really well. Intel has no business in the mobile market and ARM can't into anything other than mobile.
x86, the 32 bit instruction set, is only patent-free up to the Intel 80486. Any company will need to wait for at least 2 years until the next processor, the P5, is not patented anymore.
But here's the thing. P5 and newer processors, have new CPU instructions that is technically included in the x86 instruction set and not out of patent yet and they include some important instructions like the CPUID and CMOV instructions, the latter which is used extensively in compilers.
So until the last x86 only chip from Intel is not patented, the P6 line released in 1995, you would need to wait until 2020 to make a x86 chip without any instructions missing and 100% feature complete. Until then, you would have to license from Intel to make an x86 chip.
But ever since he P6, Intel has had more stuff like SSE released on their chips and AMD64, which is the 64 bit version of x86, would also need to be license although that is a lot easier since AMD would readily take money.
So yeah, lots of barrier and the cash margin just isn't there to justify getting into the x86 market.
Why wont just more people make ARM CPUs?
Because x86 is superior
>Do you guys think that companies like Samsung just aren't interested in making x86 cpu
The x86 CPU market growth rate has subsided significantly since the birth of the Smartphone/Tablet market. Samsung isn't going to invest their time and money into something with no good prospects.
Zen is coming in at just right time to fight over some scraps.
I know but with all the availability why not?
ARM CPUs are inferior yeah, but arent they easier to make?
it really isnt
But the Mobile market is the only market that's seeing growth.
I don't think there's anything preventing ARM cores from being as fast as x86 cores. ARM just carved out a niche in embedded stuff because they're more efficient while x86 carved out the entire desktop/laptop market because of DOS and Windows
ARM CPUs are absolutely perfect for what most people want a CPU to be. And energy efficiency over power is the way the wind is blowing for pretty much every single market. Remember how many people bought e-ink screen e-book readers simply for that reason?
>DOS and Windows
Both of which aren't what they used to be. What most people used to use computers for are now done on their tablet and phone, with only a few productivity based applications still keeping them tied to the old model because web solutions aren't totally viable quite yet, but that's not too far off.
There will always be desktops powered by x86 I'm sure. If ARM can get the software then it could take the laptop market as well.
The thing is that computers have always trickled in from other places after dominance has been assured in another arena.
ARM could probably take off on desktop/laptops as well if there was a way to run x86 software on them without a noticeable difference in performance.
AND also if this emulator was seamless to the end user. Something as user-friendly as "Windows on Windows"
>Advanced Color ePaper (ACeP) was announced at SID Display Week in May 2016, which can display up to 32,000 colors. It will be initially targeted at the in-store signage market, with 20-inch displays with a resolution of 1600 by 2500 pixels at 150 ppi. An ACeP display will have a 2-second refresh rate.[49] E Ink expects commercial production in the next two years.
Wait, so this means E-Ink screens have 0.5Hz refresh rate?
It's part of why they use so little power compared to LCD or IPS
Well its not like they need faster refresh rate since they are mostly for reading but still seems rather slow i guess.
Are all e-ink screens like this?
You don't need a high refresh rate when you are displaying static images for long periods of time
E-ink displays are usually in places where they don't change very often anyway. The neat thing about them is they can retain their image with no power if you just want it to be a static image. You only need power to change the image. They're perfect for billboards and shit like that. Also they don't need lighting during the day unlike LEDs which emit light or LCDs which require back or front lighting
>Making a competitive x86 chip is a lot of work for not a lot of return, and yes, Intel really isn't into the idea of more competition
This is probably the real reason, the amount they'd have to pay to license x86 and then actually field a product that's competitive is simply too much. Intel has decades of experience making x86 processors, very deep pockets, and ultimately controls the terms of the license.
Plus, x86 processors are becoming less popular every year, why invest billions into a market that will ultimately lose out to ARM?
1. No one new can get a license to make x86 or x86_64 CPUs.
2. Anyone who did would face the same problems AMD is having: Intel is way ahead of you in performance, and will stay way ahead of you for quite some time.
Not necessarily. Look at Nvidia's project Denver. It's a non-ARM processor compatible with the ARM instruction set. Should Nvidia obtain an x86 license they could do the same thing and not have to start over with an x86 design.
Denver was shit, just like the Transmeta chips Denver is based on were back in the day.
Licensing aside, the cost of entering the x86 CPU market is astronomical. Samsung would have no chance competing with both Intel and AMD because there isn't enough of a market. They're perfectly happy just making a few notebooks and slapping Intel CPUs in them.
Zen is just around the corner anyway, and hopefully it will create some competition in the x86 market.
Designing a competitive chip isn't necessarily difficult, ARM itself was a study of this in the '80s, and many major improvements are plenty well documented enough.
But designing a chip that actually brings something to the table worthwhile enough to win over OEMs from the more inexpensive contracts of the big names is another story, modern CPUs are so incredibly fast on their own that there's little you can really do to improve them for the majority of use cases enough to entrench your product, especially when it will undoubtedly be more expensive for what you yield.
>Plus, x86 processors are becoming less popular every year, why invest billions into a market that will ultimately lose out to ARM?
It's not going to happen, OEMs would have moved over to ARM long ago if it was as viable as the Sup Forums collective seems to think it is.
x86 has never been the world's most "popular" architecture anyway, ARM's surging is in the embedded markets where few took x86 seriously in the first place, especially its modern incarnations.
What makes you think that the next generation of home and office computers wont be embedded devices?
Because the next generation is wearables
>It's a non-ARM processor compatible with the ARM instruction set.
Denver is ARMv8. ARM doesn't manufacture processors at all.
Yeah and look at how great Intel is doing
Technically it's not.
It just has a hardware translator for arm instructions onboard
You're dumb.