Intel is dying and becoming irrelevant.

ARM motherboards when

You realize x86 software can't natively run on ARM or has to be run through emulators which incur heavy performance penalties, right?

Intel isn't dying especially when ARM has the worst performance/watt which makes it irrelevant for the enterprise market IGNORING the paragraph above.

>>58893062

It can. Microsoft has demonstrated full 32bit Photoshop running on a Samsung arm tablet running Windows 10 for arm.

ARM is a fucking joke for high performance workloads.

The problem of Intel is that they have a 45% profit margin on desktop, notebook cpus and even higher on servers. It's hard to invest billions to become competitive in a market where the profit margins are a lot lower like mobile.

>>58893088
I can "run" PS on a 12yo intel cpu too. Now apply some heavy filters and watch your samshit melts.

>>58893088
>"Microsoft revealed today that it plans to support existing desktop apps on ARM chipsets through a new emulation layer."
go kill yourself you sack of retarded shit

http://www.theverge.com/2016/12/8/13881800/microsoft-demonstrates-full-windows-10-with-photoshop-on-arm-chips

>Intel Confirms 8th Gen Core on 14nm

>As confirmed to Ashraf Eassa by Intel at the event, Intel’s 8th Generation Core microarchitecture will remain on the 14nm node.

R.I.P. Intel

http://www.anandtech.com/show/11115/intel-confirms-8th-gen-core-on-14nm-data-center-first-to-new-nodes

>>58893110

#BTFO

Microsoft demonstrates full Windows 10 with Photoshop on ARM chips

http://www.theverge.com/2016/12/8/13881800/microsoft-demonstrates-full-windows-10-with-photoshop-on-arm-chips

Time to dump Intel stocks

>>58893135
What the fuck does that have to do with performance?

I overclocked my 2700K to 4.6 GHz years ago and it still runs fine yet it's based on 32nm cpu lithography.

>>58893173

Arm chips are more power efficient for tablets and laptops

Intel puts backdoors in their CPUs so fuck 'em

>>58893194
everybody does
on modern arms it's much worse

>>58893194
ARM had backdoors long before intel you underage b&

File: intel-xeon-phi-compare-table-1.jpg (89KB, 660x339px) Image search: [Google]

89KB, 660x339px

>>58893152
see
>>58893128
>>58893062


Find me a single ARM chip that can output 3,000+ GFLOPS of FP64 performance for less than 300 watts then I'll take ARM seriously.

http://www.theverge.com/2016/12/8/13881800/microsoft-demonstrates-full-windows-10-with-photoshop-on-arm-chips

>>58893019
99% of Windows/OS X machines run x86 purely because of backwards comparability. Most people don't need it for any other reason. It makes me laugh when Intel tries to get into the mobile market with low power x86, even if it was better than ARM, no one would use it because no one voluntarily wants to enter into an agreement with Intel to get ass raped with their monopoly on x86.

Intel's burned every bridge with manufacturers already for their decades long business of ass fucking everyone, and with x86 emulation on ARM the chickens are coming home to roost.

>>58893190
Care to share the credible source to back up your statements?

>>58893250
You do realize there's a huge performance penalty for running x86 code on ARM processors, right?

>>58893286
For 99% of people that doesn't matter. Most people don't use demanding software on their PCs.

shittel in suicide watch.

>>58893318
They will feel noticeable lag and whine about it and it will all end with them going back to computers with x86 processors

t. ex-chromebook usee

>>58893331
see
>>58893221

intel makes the big bucks in enterprise where ARM does a terrible job of competing in. Im fact I don't think ARM has been relevant in enterprise since like 06

>>58893088
>32bit

File: 1482863053406.jpg (27KB, 1000x750px) Image search: [Google]

27KB, 1000x750px

>>58893019
Intel could have prevented all of that had they accepted Steve Job's offer to manufacture chips for iPhone. Paul Otellini said no and now Intel's almost dead.

You say no to Jobs and you fucking die.

>>58893457
Otellini is alive and Jobs died from asscancer

>>58893469
pancreatic cancer. and we're talking about companies.

Intel's dead because they have no play in mobile.

>>58893404
Regular people want a computer for a internet browser, a word processor, a music player and other basic software like that. In all most of cases like this there is already native ARM software to do that role. The only reason shitbooks still ship with x86 is because manufacturers don't want their consumers to install some random stupid shit and have it not work and complain. Emulation is enough to fill the gap for most people.

>>58893019
>this fucker does't actually know what ARM or RISC means

File: 25c.jpg (46KB, 484x417px) Image search: [Google]

46KB, 484x417px

>tfw Intel is building a huge facility in Arizona

File: 14310633383220.jpg (22KB, 600x346px) Image search: [Google]

22KB, 600x346px

>>58893019
What's there to like though?

x86/x64 is relatively well standartized and open. ARM has none of that. It's literally custom hardware, often with intentionaly reduced compatibility, down to custom instructions.

Crappy NIH non-standard one of a kind hardware with obscure poorly supported drivers written by pajeets which have half of their features barely working. Unremovable hardware DRM. Tivoized bootloaders, walled gardens, much more planned obsolescence. Server hardware must be at least half-open since it depends on linux support, but on the consumer devices you'll see all this crap because it's much more profitable for corporations.

File: arm2016.png (144KB, 1638x921px) Image search: [Google]

144KB, 1638x921px

http://www.phoronix.com/scan.php?page=article&item=amd_phenom_am1&num=4
http://www.phoronix.com/scan.php?page=article&item=nvidia_tk1_amdam1&num=2

Tegra K1(Cortex A15 variant) compared to the Jaguar based Athlon 5350.
Then the 5350 vs the Phenom 9500.

I think Cortex A57 cores are pretty close to Jaguar in over all performance per clock, A72 are significantly faster. Jaguar is pretty close to the performance of a larger K8 core, sometimes faster, sometimes slightly slower.
A newer smartphone has more power in the CPU than your desktop.

>>58893509
careful what you wish for

the road of doing "what regular people want" eventually leads to planned obsolescence and artificially crappified products

basically you will get an eternal cycle of crap

just look at modern laptops, there's a reason why there's no good laptops anymore, even apple makes crap nowadays

or look at xiaomi, they started as hackable and open, now they are locked down and are making indistinguishable chinkphones with broken firmware every month

>>58893542
>>x86/x64 is relatively well standartized and open

Lol, wat. AMD is only allowed to manufacture x86 cpus because of cross-licensing deals with Intel. X86 is anything but open.

>>58893667
It's very open for software developers.
Also, screencap this. You'll remember my words in several years of ARM-dominated market.

>>58893592
>A newer smartphone has more power in the CPU than your desktop.
nice meme right there faggot

>>58893062
lol, it will be just rebuilt for arm

>>58893653
It's not necessarily what I wish for, but it's the reality of what is going to happen in the future. Fact is, increasing in the future regular people are going to be running just ARM and people with a specific demand for high performance hardware are going to be running x86, most likely at higher prices and even less generation on generation improvement because the market is smaller.

>>58893019
i thought she was going to shoot a booger out of her nose

>>58893692
I really don't even know what point you're making. arm will work with devs and even apples soc deviates from the standard arm architecture. x86 has patents behind it and licensing. I don't really now what you mean by being more open. Arm needs to be pretty friendly for developers for the low level embedded coding that it's often demanded for.

>>58893457
>Paul Otellini said no and now Intel's almost dead.
>You say no to Jobs and you fucking die.

Paul Otellini said no and now Steve Job's dead.
Intel says no to you and you fucking die.

File: 1484360833318.jpg (129KB, 924x976px) Image search: [Google]

129KB, 924x976px

>new laptops come out
>can't even handle 4k
>new phones come out
>do 4k recording/playback/editing easily

really activates my almonds

Thats right all normies and desktop computers will be replaced with weak armshit. Everything will become cloud based and your personal computer will basically be a thin client.

>>58893775
Point is, there's no such thing as ARM, besides the weak defined ISA it's just everybody doing everything they want. On x86, there's at least some expectations that most of the standards will be respected by your PC/laptop. And it's much easier for vendors to vendor lock everyone with ARM since it's almost completely custom and nothing is assumed/expected.

>Arm needs to be pretty friendly for developers for the low level embedded coding that it's often demanded for.
That is also what I'm talking about. It will never be friendly.

>arm will work with devs
ARM is just a company which is licensing the ISA, it has nothing to do with devs. You have to implement the hardware yourself. Sure vendors will work with devs... problem is most of their crap will be hidden under the hood, and you won't be able to install anything else on your own hardware. Licensing for devs? Treating them like apple does? Competition does nothing when everybody is at lowest common denominator.

>>58893019
Will ARM be able to handle? CISC to RISC is a pretty big deal I believe, so, how would those ARM CPUs would look like that would handle multiporpouse OS and softwares.

I mean, ARM is amazing and efficient and all those things, but all we see is it doing on thing at time, iOS is a very basic OS but Android wich is a closer OS to a desktop OS needs an 8 core ARM to runs decent.

>>58893540
>sells the factory cause profits are bad

>>58893019
>>58893128
>verge
http://archive.is/nrbNX

>>58893915
Every modern CPU is RISC with microcode, for 15 years already.

File: 1485478000859.gif (692KB, 432x250px) Image search: [Google]

692KB, 432x250px

>Matt hi-resolution screen
>At least 500GB SSD
>Linux friendly

Is it too much to ask?

>>58893897
>On x86, there's at least some expectations that most of the standards will be respected by your PC/laptop.

that's mainly just because of a co-dependent anti-competitive relationship with intel and microsoft. It feels well standardized because there's essentially only one company making all the shots and AMD just kind of being disruptive every once in a while.

>>58893096

Fastest super computer in the world
ARM

http://www.computerworld.com/article/3085483/high-performance-computing/china-builds-world-s-fastest-supercomputer-without-u-s-chips.html

>>58893950
Yeah, but still works like a CISC processor, its instructions just are subdivided into a group of RISC instructions but it doesnt change the fact that CISC uses less memory to to the same job.

>>58894008
uh-oh, I'm just going to wait here and watch the spergs move the goalpost someplace else.

>>58894008
for what cost and for what number of cores? what's the performance per watt? because that's what they mean when saying "arm sucks at supercomputers"

chinks are doing this because they are embargoed by ITAR in dual use tech, not because it makes economic sense

>>58893019
Why does she look mad :'(

>>58893542
>is relatively well standartized and open
It's neither

>>58893897
>just a company which is licensing the ISA
They also have designs like the a-73 or cortex-m4..... along with contributing to development tools like GCC and clang....

>>58894155
It's leagues ahead of arm. How's that custom rom working for your samshit exynos?

>>58894155
oh please
you can't even assume the instruction set or cpu features on arm
i'm not even talking about things like acpi

>>58894008
RISC high performance with vectorized units.

>>58893542
>Crappy NIH non-standard one of a kind hardware with obscure poorly supported drivers written by pajeets which have half of their features barely working
By the way, that's basically the description of PC hardware in the DOS times, up until late 90s or so.

>>58893702
That was copypasta'd from what I wrote in a different thread. It was in response to a guy with a Phenom 9500, which is in fact slower than flagship ARM SoCs today.

>>58893019
>>Microsoft is bringing Windows desktop apps to mobile ARM processors
Fucking yay!
I do wonder if this is using an x86 translator on die, or if it's gonna be done in software alone.
It'd be the biggest fucking giggle if you could just load this shit up on an original Surface RT

As for arm macbooks, I'd say it's silly, but if it's finally an arm device as unrestricted in use as an x86 laptop that could be nice, especially from apple.
I'd whinge no x86, but I mostly just use GNU shit on mine so that wouldn't be a problem.

>>58893019
>Windows desktop apps

By windows desktop apps you mean the default apps from the appstore and nothing more.

>Apple could be working

Might as well go back to powerPC.

>>58893457
>iPhones and all smartphones since could have been x86, "iPhone runs OSX" could have been real, but intel fucked it up
I mad

>>58893019
>done right this time
We shall see about that
>everyones switching to arm
>including apple
Nobody here in their right mind would waste $3000+ on stupid macs

>>58894316

No, any app. Full Photoshop isn't in the app store

Windows itself becoming irrelevant will be what kills Intel.

Almost nobody gets viruses and malware on anything besides Windows, it's something everybody wants to get rid of.

>>58894367
is Microsoft and Intel the IBM of the 21st century?

>>58894367
>Almost nobody gets viruses and malware on anything besides Windows, it's something everybody wants to get rid of.
That's the result, not the cause.

>>58893019
IT'S OVER, INTEL IS FINISHED

>>58894367
>Almost nobody gets viruses and malware on anything besides Windows

Not true. Android gets way more

File: moonman_by_neetsfagging322297-dacp9vh.png (359KB, 486x569px) Image search: [Google]

359KB, 486x569px

>>58893592
>A newer smartphone has more power in the CPU than your desktop.
Unfortunately no, ARM processors have terrible IPC

2-core 3.7GHz i3-6100 GB4 multi-core FP score: 9289

4-core 2.1Ghz (only one cpu module can be active during heavy load) SD 821 GB4 multi-core FP score: 5391

So despite 4 fucking ARM cores struggling past 2GHz they still manage to be only ~58% as powerful as 2 x86 intel cores running at 3.7GHz. Don't get me started on that shitty single-core performance.


http://browser.primatelabs.com/v4/cpu/1826847
http://browser.primatelabs.com/v4/cpu/1445749

This is ignoring the 4 and 8 core intel processors that people use on desktops.

>>58894556
>4-core 2.1Ghz (only one cpu module can be active during heavy load) SD 821 GB4 multi-core FP score: 5391

So you're saying this is one core under heavy load? So 5391/core? At 0.56x the clock, and it got 0.58x the score of a 2-core x86?

Sounds like the snapdragon has got intel BTFO'd.

Especially because laptops/desktops can tolerate a lot more TDP than a phone, expect to see those clocks go up and the throttling not happen

>>58894586
>So you're saying this is one core under heavy load? So 5391/core? At 0.56x the clock, and it got 0.58x the score of a 2-core x86?
No you dumbass, that was the multi-core result (all 4 active cores were stressed).

>>58894556
See: >>58894295
I was telling a guy that newer ARM chips are more powerful than his Phenom 9500. Someone else just copypasted one of my posts from yesterday.

>>58893915
>Android wich is a closer OS to a desktop OS needs an 8 core ARM to runs decent
8 core smartphone ARMs are actually 4 core. 4 big cores for performance, 4 little ones for energy saving.

>>58894586
https://en.wikipedia.org/wiki/ARM_big.LITTLE

File: Screenshot_20170211-011848(1).jpg (45KB, 519x270px) Image search: [Google]

45KB, 519x270px

>>58894635
>only 65% as powerful as the snapdragon 821's performance cluster on the oneplus 3T
that bad huh...

File: 1417060323629.jpg (72KB, 788x685px) Image search: [Google]

72KB, 788x685px

Intel is not dying you stupid fucks....they have a pretty damn good market share percentage.

"Intel's x86 processor market share was 87.7 percent the fourth quarter of 2015, growing from 86.3 percent a year earlier. AMD held just a 12.1 percent share, falling from 13.6 percent, according to Mercury Research.Apr 22, 2016"

>>58894858

Famous last words

>>58894518
Android gets malware too but owners can literally just backup and reset their phone.

The problem with a PC is having shit go wrong and some retard that thinks he has skills repairs it for $200 until it gets infected again in 3 months because they made the customer pay for Norton instead of actually fixing the problem (Adblocking).

>>58894858
X86 is dying u nipee

>>58893200
>>58893062
>>58893096
>>58893110
>>58893213
>>58893221
>>58893286
>>58893542
>>58893653
>>58893692
>>58894048
>>58894556
How much are guys paid to shill for intel? Honestly you're looking pretty desperate now

>>58893592
>tfw Cortex-A53 seems more and more like some alien tech in terms of insane power efficiency

File: A53 core size.png (567KB, 1527x1047px) Image search: [Google]

567KB, 1527x1047px

>>58896125
Its just small, and in order. Tiny 8 stage pipeline.

>>58893725
>she

Offtopic, but wtf is going on with the intel website?

For example:
https://ark.intel.com/products/52210/Intel-Core-i5-2500K-Processor-6M-Cache-up-to-3_70-GHz

>>58893062
Mate, just run the code through gcc-arm or something lol.

>>58895117
>>58895784
>>58896125
see

>>58893221
>>58894556

>>58898385
Ok tell me where to the source code to adobe after effects.

(((shut it down)))

File: comfy-cat.jpg (49KB, 524x700px) Image search: [Google]

49KB, 524x700px

>tfw you haven't bought Intel in 9 years

>>58893062
>You realize x86 software can't natively run on ARM or has to be run through emulators which incur heavy performance penalties, right?
It has been done over and over, and the emulators are getting better and better
>what is Rosetta (PPC->x86)
>what is Xbox 360 xbox emu (x86->PPC)
>what is Xbox One 360 emu (PPC->x86)
>what is qemu
>what is UltraHLE (N64 emu on hardware that was only 2 years older)

You also have to consider that if Windows switches to ARM, all big software vendors will switch too. The only software to emulate is old/legacy software, which wont require a lot of CPU power

>>58895106
>instead of actually fixing the problem(walling of the garden full of vegetables and fruits)

Fixed that for you. The problem with PC is the user has too much control over the system. Forced updates and replacing Commonsense 2017 with a computer program are just stop gap measures until we get enough DRM instructions built in to the CPU.

>>58894556
ARM mobile chips have weak FPU, the trend is to offload FP workloads to dedicated ASICs, DSP and GPUs. Because companies can just license ARM cores and make their own chip with custom accelerators on chip they have less need for large FPU units.

I think they just announced optional enhanced FPUs for non-mobile use recently. The ability to have custom hardware on chip is killing intel, no one will buy atoms for NAS and other embedded things anymore, a more powerful FPU can't compete with dedicated silicon that is 20x faster at a specific task.

>>58898760
Man, you're fucking delusional.

ARM has horrible IPC, terrible energy efficiency not suitable for general computing, and you expect this shit to overtake x86 when it's pumping 3+ TFLOPS of FP64 per chip and using less than 300 watts?

x86 is old and deprecated but ARM is nowhere near a replacement for it.

>>58894556
I like this comparison because it looks like an arm core pulling maybe 5W peak is getting only slightly less performance per core-clock an an Intel pulling 51W

>>58899317
>ARM has horrible IPC, terrible energy efficiency not suitable for general computing
You're just talking out of your ass right now
>and you expect this shit to overtake x86 when it's pumping 3+ TFLOPS of FP64 per chip and using less than 300 watts?
Thats not even remotely related the point I was trying to make. Regardless, how many consumer devices have those specs?

>>58899335
SD821 can pull up to 10 watts under maximum load. Also look at the clock frequencies. Do you think your shitty ARM cores would still consume 10 watts max if they were clocked go 3.8GHz?

If energy efficiency is what you truly want then 72 1GHz atom cores with modified FPU units that do 512-bit AVX instruction units is what you want (ie knights landing processors). ARM can't reach anywhere near this level of efficiency.

File: P_setting_000_1_90_end_500.jpg (47KB, 500x500px) Image search: [Google]

47KB, 500x500px

>>58893019
My GTX1070 says otherwise

>>58899357
You can literally buy them for your desktop right now. Just plug them into a 16x pci-express slot and you now have access to ~3TFLOPS of FP64.

https://www.amazon.com/Intel-Xeon-Phi-7120P-Coprocessor/dp/B00FKG9R2Q

>>58899415
>>>/v/

>>58899469
Face it, arm is only good for writing FizzBuzz or using Office

>>58899485
Only 99% of offices need just Office and basic spread sheets. maybe like 99.9% of all software development would be comfortable on ARM systems.

Just think about all the people developing AAA game engines who can only use Intel, that has to be like a huge number in the 100s of developers.

What about the largest development companies like Amazon and Microsoft. I'm sure they'll never find efficiency gains from moving their terabytes dynamically JIT compiled code to whatever platform they wish with a single mouse click. It's not like they're even constantly looking for margins. Intel must be very comfortable right now.

>>58899583
99.9% of webdev maybe

>>58899583
see >>58893221

Give me one good reason why companies like amazon shouldn't instead just write AVX-512 instructions in their existing x86 code and use xeon-phi co-processors.

>protip: you can't

File: snapchat.png (104KB, 720x1280px) Image search: [Google]

104KB, 720x1280px

>>58899485


ARM is so shit.

Literally it only runs botnet, walled garden, piece of shit software like LAgdroid, Snapchat, Whatsupp... etc

That's why Android never gets any updates nor can you do a clean installation with any ISO like you can on X86

File: 1395590936369.png (25KB, 482x539px) Image search: [Google]

25KB, 482x539px

>>58899141
>The problem with PC is the user has too much control over the system
gee I wonder who could be behind this post

>>58899614
So they can pay the Intel tax of course. I'm sure a companies like amazon are just dying to build massive datacenters at the highest price per CPU. roi and power usage are probably not even no their radar.

>>58899651
>discussion of a processor technology
>brings up apps he doesn't like to support his claim that the said technology is shit
imbecile

>>58899614
Do you think Amazon's proprietary signed communication protocol would gain much speed with AVX512? Maybe they will put it in their propitiatory NICs that do the actual old fashioned 512-bit signing for comms outside their network. oh wait, they already have ASICs for that.

afaik, signing all their packets adds 2-4 ms latency which they determined was unacceptable. i could go on since i'm not under NDA any longer but I'm wasting my time here.

File: lmao.gif (353KB, 500x357px) Image search: [Google]

353KB, 500x357px

>>58898573
>he bought AMD AFTER they became shit

>>58900033
not him but I am arm retarded. What "power user" OS, application, or SBC uses ARM architecture?

File: 1405939015993.png (302KB, 450x488px) Image search: [Google]

302KB, 450x488px

> fighting intel jews with ARM
this is literally the funniest thing I've read in a week

>>58898385
bait

R.I.P. Wintel

>>58893019
Intel needs to start selling IP to other companies with limited licenses to forcibly exchange technologies back to Intel.

>>58894331
nah, they've always said they arent pulling a convergence/windows 10 thing.