150+ companies backing

>150+ companies backing
yeah i get that it's royalty-free and cheap for the corps, but does it benefit the average computer user?
cheaper hardware? software? whats in it for us?

Attached: riscv.png (512x512, 28K)

Other urls found in this thread:

en.m.wikipedia.org/wiki/Open_Firmware
twitter.com/AnonBabble

nothing
RISC-V will just be copied into proprietary designs and the status quo will resume.

It doesn't

The potential for fully open source hardware to run your open source software on. Also the potential for formally verified designs.

how does open hardware translate exactly? better drivers? custom bootloaders? coreboot n stuff?

Zero potential of that. Hardware is made by large companies that will always use properitary blobs and hardware extensions. It's not like Risc-v even uses the gpl.

Technologically it's nothing groundbreaking or new at all.
The only difference is licensing. Software world is blessing since you can make large server architectures from nothing but free software and minimal investment to licenses.
But hardware and embedded world is still similar on few places to other industries with expensive licenses, patents and locked standards which publishing giant companies change every so often to get most revenue out of them and working on the edge of anti-monopoly laws. ARM is not _that_ expensive but RISC-V is definitely a good step from this licensing madness.
But it's still 90s technology just with wider registers and buses.

yeah whats stopping riscv from turning into an ARM dumpsterfire with fragmentation and device trees

It is intrinsically safe from Spectre/Meltdown.
The down side is that it is excessively orthodox. For instance there is no autoincrement/decrement of pointers.

again, nothing. Intel will just steal the designs and manufacture their own proprietary version.

reminder that research in risc started in the 70s

Attached: perk.jpg (236x386, 18K)

Literally nothing. The only difference is that companies don't have to pay a central authority for the rights to use it. Oh and it'll be easy for students to learn how to program and make Risc-v processors because there won't be any red tape like arm and x86.

I don't get why this information has such significance that you feel a need to remind it

Intel has their own risc and alternative designs already. Every time they've tried to switch to them they've sold like shit. Despite the risc propaganda , simple risc designs actually can't match the performance of complex cisc designs.

>It is intrinsically safe from Spectre/Meltdown.
Source? Not that it really matters anyway. Side channel vulnerabilities only matter if you're running a hypervisor because you're AWS or GCP.

He's trying to say that risc is old and busted. 50 years of hardware advancements and we're still using the same designs from the 70s.

that's why DPS and simpler MCUs with no pipelining or caches are the major adopter and the sales speech is often low power consumption

but there was some progress from 70s, it started in 70s because that's when CISC emerged and those two were distinguished

No "management engines" for the glow in the darkies

lower starting barrier for new vendors (no fees and less own research)

>hardware extensions
If the hardware extensions are too fragmented then no software will use them. Companies wont want to distribute 20 programs when they could just do one that supports everything

Cisc processors are older than the 70s. The IBM mainframe computers of the 50s were cisc processors. RISC was kind of a middle finger to traditional processor design like Unix was a middle finger to traditional operating system design and like C was a middle finger to traditional higher level language design. Notice how they all emerged at around the same time.
Exactly if Intel and AMD wanted processor with God awful performance they too could get good power consumption.

shill mill

Attached: mill.jpg (474x266, 26K)

Since it's not copyleft, companies are just gonna grab it and make their own proprietary versions and it will be a clusterfuck to develop for it with a billion different incompatible extensions.

>If the hardware extensions are too fragmented then no software will use them
It's not stopping people from making software for arm devices. All this stuff is already true for ARM devices. Only high end arm servers make use of consumer friendly standards like uefi. Intel and AMD CPUs aren't open but it's far easier to create open source software for their hardware than for ARM's fragmented properitary echo system.

why the fuck would they do it. the unified ISA is good since you get compiler support for bono

whats stopping every company from implementing their own management engine into the design? is there a standard boot like x86?

*Ecosystem

so did x86

Because they're doing it right now for arm systems. Tell me does your phone use uefi? Can you load any Linux distro you like on it. Chances are it has a bunch of companion chips with no documentation to make up for ARM's trash performance to.

This is the only bad thing about it. THERE NEEDS TO BE A STANDARD BOOT!!
I'd be more than glad to leave x86 behind, as it's extremely botnet what with the ME and PSP shit , as well as poorly designed as can be seen from both Intel and AMD (mainly intel)'s hardware vulnerabilities that people are having to duct tape with all kinds of patches.
But there needs to be standard boot. That is key.

why the fuck didnt arm use uefi or at least implement own version?

no boot? wtf i dont want another platform with custom romshit like android

bump

No one except x86 has ever implemented a standard like that and it wasn't them you can thank IBM for that innovation. They wanted more than one x86 vendor for their platform so they made a bios standard they all had to use.

>Since it's not copyleft
There is however a trademark and to get a license they need to follow the rules. Also, what do they have to gain in making excessively divergent versions? After all many just want to make cheap embedded stuff.

Proprietary lockdown.

It allows for more competition to enter the market. You'll have more alternatives to jewtel and ayymd.

>they've sold like shit
More like nobody wants something that isn't either x86(-64) nor ARM.

But Risc-V already has a large amount of software working on it.

Didn't SPARC have something like that too?

Yeah but is anyone gonna write code utilizing those proprietary extensions in the first place? Considering how much of an issue crossplatformness is I doubt anyone will do it voluntarily.

150+ companies suppressing and delaying

I have been watching compiler Gandalfs videos for years...
i know this stuff takes time but COME ON.. RELEASE SOMETHING GODDANGIT!

yes, it will fork

but the main objective, run a OS, any OS, will be the goal of any design

each design modification will be able to compite against each other without having to worry about not having market access for not beeing AMD or Intel

Attached: 800px-Linux_kernel_map.png (800x600, 377K)

Meldown and Spectre are microarchitecture vulnerabilities, not ISA vulnerabilities, so it's not intrinsically safe.

Yep and it's even better than IBM standard. en.m.wikipedia.org/wiki/Open_Firmware

The fact that nobody has pointed out how laughably retarded this post is just goes to show the average caliber of RISC-V shills.