Intel will allow ARM chipmakers to use its 10nm manufacturing process

Intel made ARM chips in its foundries before though. See also XScale

>>56114736
is ARM botnet like x86 is?

>>56114759
varies by vendor

>>56114757
>Intel still holds an ARM license even after the sale of XScale; this license is at the architectural level.

I wouldn't be surprised if they had a stealth project to make an ARM chip.

>>56114759
TrustZone.

>>56114773
>>56114782
What makes ARM different from x86?

>>56114736
Do they contain backdoor like Intel x86/x64?

>>56114759
some arm chips require botnet firmware for graphics

amd arm chips have secure boot

>>56114801
Power consumption, I think.

>>56114757
Intel dumped XScale and specifically blocked any product that might compete against Intel's own from accessing its foundry services.

That's why Intel's foundry has only been baking FPGAs and niche network processor chips, not mainstream SoCs.

Since Intel now gave up the smartphone and tablet market (and fired thousands of employees in that division), the foundry is free to take orders of smartphone SoCs.

>implying Intel isn't just letting ARM guinea pig their process to see if it is indeed gonna fuck with electron flow
Nah OP.
Too much software that's written and optimized for it, x86 is here to stay.
x86 4lyfe

>inb4 the entire world moves over to Java for everything

>>56114872
Fabs that aren't running at full capacity are money sinks

I think they ended up mothballing a nearly complete fab as there wasn't enough chip demand for it to be opened

>>56114913
Intel was being retarded with that policy.

Intel has a finger in pretty much every semiconductor pie, so there are very few product segments that would require Intel's cutting edge manufacturing (and all the extra cost that comes with it) __and__ not potentially compete in some way with one of Intel's products.

Samsung did the right thing by letting everyone in, including its direct competitors (Qualcomm competing Samsung Exynos, Apple competing with the entire Samsung smartphone business).

good for them, phones are overpriced wifi modules, any actual computing is done on the server side

bring out the optane

AYYMD IS FINISHED & BANKRUPT

AMD K12 on Intel technology

>>56115168
Oh wait nevermind this is only for reference ARM cores nothing useful for AMD

>>56114782
>TrustZone.
Sauce on TrustZone requiring a separate processor running blobs like Intel's ME and AMD's PSP require?

>>56115317
It doesn't require a separate processor but it's still fun: https://www.blackhat.com/docs/us-15/materials/us-15-Shen-Attacking-Your-Trusted-Core-Exploiting-Trustzone-On-Android-wp.pdf

>>56115317
Don't listen to that vidya gaymer tech evangelist wannabe, he probably didn't even read http://www.arm.com/products/processors/technologies/trustzone/
it's basically a specification on how to architecture and arbiter different components of a SoC. No it's not as simple as slaping a micro core in there and connect it as master to the system wide bus.

>>56115425
Oh yeah and I forgot to mention that even with the trainings webinars and meetings, constructors like Qualcomm still fuck it up.

>>56115083
They make ARM server chips.

>>56114801
Cheaper to make, literally that's it

>>56114736
Kind of a meaningless statement, they've done the same in the past, but its murky as hell.
Intel isn't a company that will be taken at face value. They probably want someone to step forward with a 50mm2 ARM SoC they can use for victim test wafers. Tweak their process so they can then lie to investors and claim their own 10nm chips are ramping up even faster than prior 14nm parts, which were a disaster.

>>56115317
AMD' "PSP" is an ARM TrustZone processor. Its an ARM A5 core.
Though it is literally nothing like intel's backdoor. Its a far more complex version of trusted execution for securing devices on a network. Businesses working with confidential material can allow employees to use their own machines, online banking and shopping become more secure. Its neat stuff.

>>56114736
Need a way to recoup on the foundry, shrinking nodes is getting really fucking expensive

>>56117033
>AMD' "PSP" is an ARM TrustZone processor. Its an ARM A5 core.
It's not an ARM TrustZone processor, it's an ARM processor with the TrustZone feature. That doesn't make TrustZone bad, that makes AMD's security implementation bad.

>Though it is literally nothing like intel's backdoor.
AMD's PSP has the exact same problems as Intel's ME. Both have complete access to your hardware at a lower level than you will ever have and are controlled by signed proprietary blobs that you will never be able to look at the internal operations of or replace.

>>56114801
It's a RISC architecture rather than a CISC architecture. Much fewer instructions supported directly by hardware.

>>56117969
But modern x86 are RISC.

>>56114801
Fewer instruction sets, so they're simpler, cheaper to make, reduced complexity means reduced power consumption

>>56118034
wut

mmx, sse, sse2, avx etc.

>>56114736

I don't think Intel's fabs are running at full capacity and Intel losing increasingly more money and having the other fabs like TMSC, Samsung and GlobalFoundaries basically catching up means this was kind of a non-negotiable thing that was bound to happen.

I still don't know why they are making Kaby Lake when they could've just introduced improved steppings to bring like 90% of the benefits to Broadwell and outside of improved graphics and H.265 hardware, it's pretty much useless.

They should rename Kaby Lake to Coffee Break, since that's probably why 10 nm is taking so long from them taking so much of them.

Why is ARM still considered RISC?

Shouldn't it be called the Orthogonal but only Slightly Reduced Instruction Set Computer? OSRISC

Is it possible for ARM to ever out perform x86 chips?

>>56118240
Its unlikely on the high end, but not inherently impossible in any way.

>>56118093
He doesn't know what he's talking about - he's just regurgitating something he heard