Google and Qualcomm partner up to develop RISC-V based Wear OS chipsets
- V
- Vikass17
- CbF
- 23 Dec 2023
DeepIn2U, 18 Oct 2023Arm is NOT a monopoly did you fully read the article?
A... moreCustom core license is many times more expensive, so you have to stick with Arm cores. They are trying their best to stop custom core designs.
It is monopoly. Think deeper.
- Kangal
- Hq4
- 25 Oct 2023
Kycklingris, 24 Oct 2023Sure, there isn't all that much in the high performanc... moreWell, if we take phones as an example we saw a couple major standards:
- Inception: 2008 Android 1.5, 65nm, ARM 11
- Adoption: 2011 Android 4.0, 32nm, Cortex-A9
- Maturation: 2016 Android 5.1, 16nm, Cortex-A73
- Evolution: 2022 Android 13.0*, 6nm, Cortex-A715
...now I don't expect RISC-V to surpass ARM, let alone x86. But if they can at the minimum match the previous standard (maturation) set by ARM, I think they would be onto something. The efficiency of 16nm, speed of A73, and the optimisation of Android 5.1 I believe is the true benchmark. They need to match/surpass these in each category. I think efficiency is a given since they will use newer nodes, but we can't forget the software optimisations there. The performance is more challenging but the A73 is about on-par with the Intel Core-M, and I think this could be achieved using newer techniques. The last part is the most difficult; software support. RISC-V needs to get adoption, support, and optimisation on the level of Android 5.1 and I just don't see that happening. In terms of Linux, there is three platforms: Debian, RedHat, Arch. With many projects downstream from it, like Ubuntu, Fedora, or Manjaro. Getting unofficial support for those six plus Android, then official support feels like the minimum. Or getting one commercial software, and having that specifically optimised with great application support, is the minimum market viability.
If RISC-V only ever manages to get Cortex-A515 performance levels or lower, or has huge inefficiency problems, or is only able to get experimental software support.... well, it won't be going anywhere. ARM is already dirt cheap and decently supported, whilst x86 has a mountain of support behind it. Even in the SBPC demographic people won't be taking it seriously, and it will languish there. I think that's what's happening and will continue to happen. The major position we will see RISC-V play, is that it will get a demotion, and will be made into a coprocessor/accelerator alongside some x86 CPU, or dGPU, or in an ARM-Server. But the lack of direct development as I outlined above will mean it won't be viable as a computing platform onto itself; it needs money, effort, time.
- K
- Kycklingris
- StQ
- 24 Oct 2023
Kangal, 19 Oct 2023The problem with RISC-V is that it looks promising, but the... moreSure, there isn't all that much in the high performance space yet, that takes time, especially since the competition is always working themselves. But in that space we'll have to wait and see how good the HiFive pro p550 is for a good benchmark of where hpc RISC-V is currently.
That is not to say that RISC-V isn't being used, you see, RISC-V is already being used in a shit ton of devices like phones, laptops, household appliances and especially IOT. Of course in a lot of the parts that are RISC-V based are minor IO, storage controller or other sub features. But RISC-V is here to stay, that's almost certain.
- K
- Kycklingris
- StQ
- 24 Oct 2023
Adpocalyptic, 19 Oct 2023Here's hoping for RISC-V SBC's 🤞🏽
Well milk-v seems to have some options, but I don't know if the software is any good since I believe it's a Chinese company.
I also know pine64 has 1 or 2 options and theyre European with a good software support record.
Of course there's also whatever sifive is working on, but their stuff is mostly for companies to test out whatever new RISC-v version/addon that's being worked on, it's expensive but I'm pretty sure its also the fastest you can currently get. Of course probably won't stay that way for long since the Chinese have been working damn hard on RISC-V.
Now I know it's maybe technically not a computer but espressif also has options both RISC-V only and ones with xtensa or something and RISC-V. From what I remember they run rtos at least.
- 404
- thv
- 24 Oct 2023
you have to suffer, 22 Oct 2023so like i get the capability and performance as x86 process... moreWell... kinda.
- you have to suffer
- wru
- 22 Oct 2023
404, 22 Oct 2023With years of development, they might can be better than bo... moreso like i get the capability and performance as x86 processors, and power efficient as ARM chips? it would be mind blowing to have one.
- 404
- txF
- 22 Oct 2023
Kangal, 22 Oct 2023It's not fully open-source.
Companies can get the lic... moreFair enough.
Also technically if we counted from the 32-bit version of both x86 and ARM, they actually start developing on the same year, 1985, so technically development of the 32-bit and 64-bit processors already start from 38 years ago.
- Kangal
- Hqa
- 22 Oct 2023
404, 22 Oct 2023It's just a matter of time I guess. RISC-V is still a ... moreIt's not fully open-source.
Companies can get the license freely, but they are able to make their own modifications AND not-release the design/details. That makes it open-source in theory, but proprietary in practice. In fact, such methods usually lead to things being more locked-down and not less.
Also don't forget that fragmentation is a real thing. RISC-V isn't really out yet, and it already stuffers from this. It will get even worse if it keeps moving this way. It's really hard to develop for a platform which has these issues at a core level. For all it's faults, the x86 32-64 platform is rock solid with 35 years of development and trillions thrown at it, the ARM v8-v9 platform is a bit looser, it has around 15 years of development., and has billions thrown at it. RISC-V in comparison has development phase in the months/years, and has funding in the millions, which is why it's so shaky.
...so don't hold your breath !!
- 404
- txF
- 22 Oct 2023
Kangal, 19 Oct 2023The problem with RISC-V is that it looks promising, but the... moreIt's just a matter of time I guess. RISC-V is still a relatively new architecture, compared to ARM that already existed since 1985, and even x86 that already existed since 1978. It will surely took time, but it will be a great alternative, since it's fully open source.
- 404
- txF
- 22 Oct 2023
you have to suffer, 18 Oct 2023looks like risc v has put into use for consumer electronics... moreWith years of development, they might can be better than both x86 and ARM, if we looks at the downsides of the x86 and ARM architecture. Think about it - a more customizable architecture than ARM, while being more power efficient than x86 (so literally a combination between x86 and ARM), isn't it great?
- S
- SERIOUSBLACK47
- 6Qn
- 19 Oct 2023
[deleted post]That must have been negotiated under acquisition & qualcomms original agreement with arm
I think there is a lot of ground since both designs are either fully or partially using arm design
Qualcomm should not hurry on the verdict & let the chips fall
& Qualcomm hasn't released any product
Arm claims to destroy the product is blathering & foolish
- A
- Adpocalyptic
- raQ
- 19 Oct 2023
Didn't Intel also have a project that involved them working on RISC-V? 🤔
- Kangal
- Hqa
- 19 Oct 2023
Lean.N07, 18 Oct 2023Finally, it's a good starting point. Let's see ho... moreThe problem with RISC-V is that it looks promising, but there's really no hardware out there. They have white papers and announcements, of these great accomplishments but nothing to substantiate them.
As an analogy, think of MediaTek chipsets and they will promise a synthetic and theoretical properties of the Dimensity 8200... but in real-world practical properties they might only deliver the likes of the Helio P60.
People have to remember that software is important, there's no need for all that power if you lack the necessary control. But you also can't have superb control but with very little power. You want to maximise both.
- YUKI93
- IVG
- 19 Oct 2023
PhoneFreak45, 19 Oct 2023Can you reccomend some. I've been watching somegadgetg... moreI also like to watch SomeGadgetGuy. He's one of the rare techies who loves seeing competition. I also follow TK Bay, The Tech Preacher, and Flossy Carter.
- P
- PhoneFreak45
- qbk
- 19 Oct 2023
YUKI93, 19 Oct 2023Yup, he did it in one of his smartphone fails videos. I als... moreCan you reccomend some. I've been watching somegadgetguy. I enjoy his tech takes but the woke stuff being inserted is getting a bit ridiculous, especially with his views on quote-un-quote "price jacking" He doesn't understand how business works and I find it really very frustrating at times. So some suggestions would be greatly appreciated.
- YUKI93
- IVG
- 19 Oct 2023
PhoneFreak45, 19 Oct 2023I didn't even know he did that. So glad I unsubscribed... moreYup, he did it in one of his smartphone fails videos. I also stopped following him, alongside other popular sellouts like MKBHD, after his LG Mobile videos. It's nothing more than SEO clickfest and I'm totally not regretting leaving them for good. I now realize there are far better indie reviewers out there to follow.
- P
- PhoneFreak45
- qbk
- 19 Oct 2023
YUKI93, 19 Oct 2023Mrwhosetheboss doesn't see it that way, or else he wou... moreI didn't even know he did that. So glad I unsubscribed. That's [redacted] CRINGE!
- YUKI93
- IVG
- 19 Oct 2023
PhoneFreak45, 18 Oct 2023Hey guys, you see that pronunciation guide? It's very ... moreMrwhosetheboss doesn't see it that way, or else he wouldn't make a Friends skit just to mock Sony's smartphone naming scheme. 🤭
- YUKI93
- IVG
- 19 Oct 2023
Thinking about it, this can be a good start for RISC-V adoption on mobile devices. If we want to see how good RISC-V can be, doing it on something with a small battery can give a bigger picture of its battery life.