There's Still No Sign Of AMD's Low-Cost ARM Development Boards

[pal666]

Once Zen is in the market, I think they will be able to scale down

lol, like intel was able to scale down in smartphones for billion dollar losses and eventual quit

[pal666]

long term x86

x86 is only relevant for legacy windows apps, everything which could be recompiled, does not depend on x86, and x86 (as isa) has no other benefits

[pal666]

their CPU division is basically disbanded and AMD becomes a GPU-only company)

that will make intel monopoly, so it is not that simple

[Zola]

More like 3 years, and it is a total redesign.
finishing whatever they begun as changing course now would have wasted the development effort.
That said, ARM space is a very high-competition low-margin market in the consumer segment you talked about.
Big companies don't give a shit about lack of support, they want stuff that works decently now and costs a few bucks. Mediatek usually wins here.

AMD GPUs aren't in high demand in embedded market where at most you need to display 2D content, and I doubt they can make a decent GPU for mobile (decent = powerful AND low-power enough).

Wouldn't call a Smart TV SoC market neither lo margin nor overflowing with competition one. It's a nice steady one with a decent size & it's lagging behind almost every other. The Satellite receiver one is totally another pair of slippers & MIPS rules there.
Zen is in active development more than 3 years now & we only had a glimpse at prototype recently so all in all you will find my prediction of 5 years as pretty much accurate one.
What development efforts? Licensing a POP IP cores from ARM & putting them together with available GPU's from APU's. I ensure you how that require a bare minimum of development efforts. Actually MTK risen doing it like that.
& I don't give a shit what big companies give shit about so I will always give my money rather to Qualcomm than MTK & certainly not because I love Qualcomm but because I love to know what I am putting in & that I am able to control & alter it (Code Aurora).
I am not certain that you even know what embedded is. Their mostly aren't any dedicated GPU's in embedded microcontroller products not even 2D ones.

[darkblu]

The problem with AMD's A1100 paper launch (after a two year-delay) sorted itself out for me with SolidRun's Armada 8040 board. In the meantime I've managed to port a good deal of code to ARMv8 on my "workhorse" RK3368, so I can start benchmarking and performance-tuning as soon as the 8040 arrives in October. When there's a desire there's a solution.

[edwaleni]

This explains why the Linaro Cloud hasn't responded to my request for an AMD A1100 based CentOS host yet.

The 14nm 65W 16C/32T Xeon D-1587 is around, but expect to pay north of $1500 for the SoC and its only in BGA.

While people seem to have access to pre-release fabs, its still not on Intel ARK.

http://wccftech.com/intel-16-core-xeon-d-1587-benchmarks/

[Max Payne]

AMD, appears to have partially suceced in the goal of making the A1100 arm opterons, they scared NVIDIA and Samsung off the server market, made Qualcomm spend more time in designing a decent chip that won't enter the market in the short term, so when Zen based server products enter the market it isen't overflooded with arm server products.

The A1100 doesen't even look like a serius effort from AMD, It only adresses the low power, low end server market for 25-35Watts, and AMD has the expertise to design a custom core if they really wanted to enter the arm server market seriuslly, this really looks like a low cost design just meant to scare and confuse the competition.

But still I must ask, do these cpu's support KVM or just plain Qemu?

[starshipeleven]

That's what I'm saying, hw is shit, but the sw crucially isn't (relatively speaking).

Ah ok, sorry for that.

Toned down the need for standardization, as I think the 96boards spec is part of the problem.

You are an idiot.
Your edit makes no fucking sense. Apple is doing masterful artwork tightly embedded hardware that is totally useless beyond their intended scope (single-user, single-disk, not-terribly-powerful dumb-end-user-devices), which is exactly what most manufacturers are doing already (minus the "masterwork artwork" part)

Standardization is the only thing that allows to keep costs down enough for stuff to sell, AND that allows people to make different devices out of the same components. Then sure there are good and bad standards. But standardization should be the goal.

96boards does not have any kind of clout like say Microsoft or Intel back then that bascially DOMINATED (and still do in x86 land), so they cannot really make designs that are too inconvenient for manufacturers or they risk auto-failing as none shows up.

The road to get the manufacturers to reach some kind of standardized non-shit bootloader sequence and firmware like with UEFI is long and perilous.

[pal666]

That said, with CentOS behind them now, I would not be surprised if the community spends some time on a AArch64 port (which I think the AltArch SIG is doing at this point) that gets Red Hat's notice latter.

??? https://fedoraproject.org/wiki/Archi...4/Installation

[pal666]

may I suggest an alternative that's truly an SoC with strong power efficiency: http://www.anandtech.com/show/10574/...f-atom-iot-kit

nobody wants intel's socs, even sold at loss