Why don't qualcomm and samsung make desktop cpus?

so is it an ARM desktop CPU?

Uhh I don't know anyone who throws away a Samsung S6 or iPhone 7 in the bin, desu. They all have resell value - mind elaborating what the fuck you're on about?

Yep. People were speculating that it was going to be a higher clocked/binned SD850, but Qualcomm confirmed its actually a unique die. Made specifically for small form factor desktops.

Those laptops are Windows 10 which doesn't belong on anything that isn't x86 IMO. Running Windows 10 even with x86 emulation is still going to feel neutered and at that point I'd rather run android

throw away is not just used for meaning putting something in the trash. It is also used in the sense to get rid of something you do not want.

I think the main problem is there's no market to make use of them. People get desktop processors specifically to do demanding x86 workloads and as far as I'm aware there's nothing with the ARM ISA that can even match 14nm zen 1 performance/watt ignoring the performance penalty of emulating x86.

Either:
4 very high performance cores at 4 GHz with half the IPC of zen 1 for like 65W
or
8 very energy efficient cores at 2 GHz with a quarter the IPC of zen 1 for like 35W

Both implementations are very ugly and will appeal to no one.

ARM is a meme anyways, all the other big tech companies are backing RISC-V as the ISA of the future.

That's even worse, RISC-V might be gaining ground on efficiency but they can't provide high performance alternatives to zen 1 SKUs yet. Even then ARM pretty much dominates embedded.

COSE RETARDED CASTOMERS WANT "TOP CPU FOR NOT UPGRADE IN DECADE"

Of course it dominates embedded, it's established. However we've already seen test systems with 512-core RISC-V systems, beefy mobile chips and have better performance on benchmarks, etc. There's no reason it can't compete on desktop. We'll get there. Fuck ARM.

ARM has had more time to grow, obviously. RISC-V is in good hands and has made better decisions in ISA design. Who's going to win, a bunch of UK cucks or literally every other tech company with a vested interest in high-performance silicon that is literally throwing moneys and engineering time at the RISC-V collaboration? Join the mailing lists, there is tons of activity.

Well by the time you get anywhere near there ARM will be right behind pulling the strings. Also if RISC-V is to have any chance of even 1% pc market share it better have efficient x86 emulation and a huge tech company behind it working day and night to make an OS that will work directly on the risc-v silicon.

In all probability we'll just get cheaper FPGAs competing with todays ASICs assuming ARM lets them.

That was literally said of AV1 and all they have to show for it is 3,000X longer encoding time than HEVC.

Good for you that you want things to be cheaper but stop expecting miracles.

The whole point of the RISC-V exercise, other than being "free", is that it eliminates the need to pay ARM any licensing fees. Hence, no, ARM will be kept at ARM's length away from the project by the gate keeping companies.

x86 emulation is a problem anyone would need to tackle to support existing desktop binaries, whether it be RISC-V or ARM or POWER. But ultimately this isn't a problem for most users as applications can simply be retargetted towards new architectures. If there's a market for it, companies will port their software. The idea is that the companies backing RISC-V eventually want to push x86 out, and they have the economic means of starting this transition.

hmm interesting. isnt ARM even more proprietary than x86 though?

Then risc-v better start working on time travel because 7nm zen 2 and 5nm silicon production starts next year. By the time we get to zen 3 we'll practically have facy high performance asics.

You realize that you can take a RISC-V design, or make one yourself, and send it to a fab with whatever tech node you want? Digital physical design has never been all that difficult once the tech node and standard cell library has been developed. What a weird thing to say.

Oh cool so you can take publicly available zen IP and slap it on risc-V? dumbass

You don't even know what you're saying.

More proprietary? No. The situation is actually somewhat similar despite the glaring differences.
The core ARM ISA, even fully designed cores offered by ARM themselves, can be produced by anyone who pays the fee. Technically this is the case for X86 as well.
With ARM however we have individual vendors who make various changes to their own SoCs, they might add their own accelerators, add their own interconnect fabric, so its a bundle of undocumented IP mixed in with the proprietary yet available o license ARM IP.

A company could make a chip with stock ARM cores, ARM fabric, ARM GPU, ARM everything, but no one is doing that. So its kind of a shitshow.

Unless you're jim "fuck shit up" keller himself in the flesh you won't get near even the first iteration of x86 zen ISA performance and efficiency for desktop chips if you're starting from scratch or copy and pasting things.

Attached: 25d.jpg (951x972, 104K)

No one said anything about copy-and-pasting things, and I don't think your understand what "ISA performance" means. To be frank, you come off as a paid AMD shill who has never been in the silicon industry and has no idea about what he's saying.

RISC-V is DOA for pc market, screencap me for all eternity.