Oh lookie, more lundjew spam.

>>109563933
>"I HAVE A VERY STRONG OPINION ON SHIT NOBODY CARES ABOUT"
> — Literally WHO???
Buy an ad faggot

>>109563933
I was discussing it with Gemini yesterday.

Do you have a reason for not just posting the link without breaking it apart?

>>109564863
So he doesn't get autoflagged, must be a tourist

>>109564863
I tried several times, but it kept triggering 4chan's spam filters. I'm guessing it's due to the .gr domain and all the GET query parameters, I wish this guy hadn't structured his website this way.

>>109563933
We should design our own ISA, 4machine or 4M. I will make the logo.

>>109565809
There is no need. x86 is already perfect.

File: HO_W2czXgAEM7wn.jpg (216 KB, 1080x1349)

216 KB JPG

>>109565836
No.

> make a vague grandiose statement
> "just google it"

retard alarm

>>109565927
I posted a version of the link itself if you don't want to google it. Couldn't post the actual thing because of 4chan's stupid filters.

>>109563933
does not matter you target whatever gcc standardized on
and then you save 20% on your fan controller
(already won btw)

File: RVA23.png (91 KB, 426x1086)

91 KB PNG

>>109563933
>everything is optional
Half true. I mean, it's true that all instructions are optional, but the RISC-V Foundation released the RVA23 Profile which is what they consider the bare minimum a modern CPU needs to properly work.
And while it isn't a standard, it is basically a de facto one, considering how all Linux distros and other operating systems are targeting that very configuration as the main one.

>>109566321
>And while it isn't a standard, it is basically a de facto one
How can it be, when there's hardly any processors that actually support it?
And a standard isn't great if it sucks. The one benchmark I could find had a full RVA23 CPU doing worse than a raspberry pi.

> meanwhile taking over embedded entirely
Yep, looks like it failed. Tragic.

It's great because it's license-free and open to implement. It's supported by both major compiler suites. It's a whole ecosystem where companies can compete based on the quality of the implementation.

This pisses off the current tri-opoly of ARM, Intel, and AMD.

>>109566879
>It's great because it's license-free and open to implement.
Dude an ISA is just an interface. Nobody cares if you don't open sores cores.

>>109566895
Yes they do, they care very much. Everybody wants to make their own cores these days. China, Russia, France, the UK, Germany, Japan, even Korea, other places too. Taiwan. US.

Hobbyists also produce cores now, there are hobbyist-produced RISC-V cores implemented for FPGAs. Most first silicon is FPGA-based. Open source cores are the future for sure.

>>109566879
>>109566914
Well, there is also OpenPOWER, but it seems nobody wants to deal with IBM, anymore.

File: file.png (74 KB, 600x607)

74 KB PNG

>>109566914
>China
China's going with Loongson because it doesn't totally suck, and all those other countries will probably just license AArch64 if they want something high-performance, just like Japan with Fugaku.

>>109567021
Loongson is a meme. That architecture is literally a copy of MIPS64 and there are almost no new software for it.
They are also making new x86 cpus thanks to the VIA partnership, which at least makes more sense.

>>109567087
It's less of a meme than RISC-V is. The processors exist and are being actively used in desktops and servers, albeit in China. RISC-V can't even say that.

File: videoframe_484237.png (1.76 MB, 1920x1080)

1.76 MB PNG

>>109566914
Stop being delusional anon, I also want open source ISAs and stuff but RISC-V is not a good design.

>>109563933
>AArch64 will be a better open ISA as soon as the patents expire in a decade
If we go by that might as well wait for the x86_64 patents to expire, which is earlier than the AArch64 patents.

>>109567128
It's fine, it's MIPS-inspired.

>>109565809

what this fujitsu gpu connector be

>>109565840
great teeth, skin and clean clothes

let's pray there are no tattoos

>>109566879
This, RISC-V is already eating the microcontroller market, there is nothing like it as an alternative.

Glanced at his CV. He is an OS guy, as opposed to compilers or CPU microarchitecture. Opinion ignored. He has never been on a CPU development team. and has no idea how long it takes to design one from scratch.

Oh and by the way, I have friends who were hardware guys at Cruise. They said the software guys were entitled shits.

If the RISC people were on suicide watch you wouldn't make a thread attacking them and bringing attention to them. You only do things like that if you're trying to get attention for yourself. This thread has nothing to do with RISC and everything to do with OP fishing for attention. Look how they're arguing with their own quotes in their own post.

>>109563933
>here will never be a RISC-V CPU that beats an x86 or AArch64 at anything
It's already stamping over ARMs market share in the microcontroller space.
The laptops and devel boards in RISC-V are pretty much toys, now, but they was with ARM ten min ago and now look at what ARM doin'.

>>109565836
>x86 is already perfect.
So in five years time you won't be seeing RISC-V lifting similar weight with ⅜ less energy draw?


>>109567002
OpenPiton?
Supposedly that'll scale up to ½billion or so cores....

>>109567111
>RISC-V can't even say that.
Yet.
How long did it take ARM to end up in the datacenter?

>>109567128
>but RISC-V is not a good design.
OFC, the open sourced design you've released is so much better...

>>109569762
>It's already stamping over ARMs market share in the microcontroller space.
For money reasons, not because it's better.
>So in five years time you won't be seeing RISC-V lifting similar weight with ⅜ less energy draw?
Make it a decade, and I'll still take that bet against x86.
>How long did it take ARM to end up in the datacenter?
Still hasn't.
>OFC, the open sourced design you've released is so much better...
x86 and AArch64 are both better. Are you going to deny that?

>>109570624
>For money reasons
Explaining why ARM has the larger stocks already on the shelves, the more mature processes and manufacturing at scale.
Because it cost more.

Not because it's more power efficient for a given workload, which really matters in most microcontroller application.

>a decade, and I'll still take that bet against x86.
x86 is already losing out to ARM, which *will* be curbstomped by RISC-V.
Wise move.

>Still hasn't.
So there's not ARM based server solutions, that for all but a few specific workloads can do the same, or more for less?

>Are you going to deny that?
I note the way you casually adjust the subject due to your poor grasp of reality as you desperately attempt to reframe.

But: Yes.
Which metric of 'better' shall we choose to smash your delusion?

>>109569696
>entitled shits
People who spend their career around the concept of innate flexibility end up expecting the hardware to be as magically changeable as their 3-layers-extrapolated code.
All programmers should be forced to have a crash course in embedded and FPGA logic and no soft-side bigwig should be allowed to lead a dept without concurrent skill in EE.

But I assume you're talking about people who have never touched ISA and, in a practical sense, see firmware as the lowest level.