Language:
switch to room list switch to menu My folders
Go to page: First ... 33 34 35 36 [37]
[#] Tue Jan 03 2023 13:52:33 EST from Nurb432

[Reply] [ReplyQuoted] [Headers] [Print]

For the IoT market, i think the Eco system is there. At this point plenty of options with full support across the spectrum. Now its just a matter of slowly eating away at the ARM market share. 

For end user devices, beyond the lower/medium end like i'm playing with it may take another year, but the momentum is there and its turning out to be a logarithmic curve now that its in-gear and the threshold was met, so to speak. Tho we will see when that tablet comes out, supposed to be this summer.  I expect to see competitive desktops summer of 2024.

Server side, i heard that some UK thing tied to the government just committed 300m ti start producing high-end 'chiplets' for HPC this summer.

And technically its not encumbrance free. The downside is vendors CAN add proprietary extensions on their own. If that starts happening it will fracture things, badly.

Tue Jan 03 2023 01:39:56 PM EST from IGnatius T Foobar
Right, so it's at a point now where it is being used simply to avoid the cost of a license, which is where the open source software world was 25 years ago.
It has not yet become "the" ecosystem that is everyone's first choice becuase it is encumbrance free and permanently vendor neutral.

Hopefully we'll get there and it doesn't take quite that long. Software that is hard-built to a particular architecture and platform is not as mainstream as it was a generation ago. Open source infrastructure and web based software have made it easier than ever to jump around. Perhaps my current motherboard will be my last AMD64.

 



[#] Wed Jan 04 2023 11:07:28 EST from Nurb432

[Reply] [ReplyQuoted] [Headers] [Print]

Not tried it yet, but looks like Firefox is availble for the starfive v2.  

 

and a few other nice things like libreoffice and QT5.

 

 



[#] Wed Jan 04 2023 12:24:01 EST from Nurb432

[Reply] [ReplyQuoted] [Headers] [Print]

Its begun.  Tipping point in 3...2...1....

Like them or hate them, google promoting RISC-V to tier 1 for android will change the game.

 

 

https://arstechnica.com/gadgets/2023/01/google-announces-official-android-support-for-risc-v/amp/



[#] Wed Jan 04 2023 13:45:37 EST from IGnatius T Foobar

[Reply] [ReplyQuoted] [Headers] [Print]

No doubt, there are industry-wide concerns about the future ownership of ARM.
The takeover attempt by Nvidia was a shot across the bow; it could have turned into a real shitshow very quickly.

And I do agree that IoT will precede consumer devices, since they generally only have to run a single application that is built by the appliance vendor.
We've already entered an era where Linux has replaced Windows (or OS/2) as the dominant operating system for "things with screens" -- such as airplane seatbacks, retail self-checkout systems, ATMs, etc. -- and the Android userland seems a better fit for these devices than the conventional one. (Oblig: there is not, and cannot be, a universal user interface that works well on both small touchscreens and large upright desktops, as even M$ discovered after decades of failures.)

So yes, these devices would be a great opportunity for RISC-V.
It could take a while before mobile moves away from ARM, for the same reasons it's taken so long for desktops to start moving away from x86. And no, I don't hate RISC-V; I like the idea of it and am excited to see what happens next.

[#] Wed Jan 04 2023 17:02:37 EST from Nurb432

[Reply] [ReplyQuoted] [Headers] [Print]

Agreed it wont happen tomorrow, but with the sudden acceleration of things, i think it will happen far sooner than anyone expected.  I might even be around to see it.

ARM took decades to get where they are.  RISC-V, while isn't new, reached maturity much faster and market penetration almost overnight.

I think RISC-v had some things at play ARM never did.

  • Chinese Sanctions.. That single event alone made it almost inevitable. Further developing RISC-V was the only real option.
  • NVIDIA buy-out of ARM.   I may be in the minority, but i think that would have been their last chance of survival as the money flow and tech resources to improve things would have been immense.  Might have been some market concerns later for licensees/competitors of NVIDIA, but i suspect in the end it would have been OK as they want the revenue and governments woudl be breathing down their necks. 
  • Changes in how ARM licenses after the failed buy-out.  Now if you license their CPU, you have to use/license their GPU and FPU, and TPU, and you cant use their own. If they try to do that to NVIDIA or Apple when they renew, its game over.
  • General cost / control.  Always better if you can pay less AND keep total control over your stuff.  ( see above item :) )
and a few technical reasons i think too, but not as dramatic as the above

 

How long will it take to topple x86? Of course much longer than ARM.



[#] Thu Jan 05 2023 07:06:13 EST from Nurb432

[Reply] [ReplyQuoted] [Headers] [Print]

And yes, i know, i'm a borderline fanboy.  Its my EE background. 

So disillusioned with ARM holdings its not funny.   They had the goose in hand, with all those decades of work to get where they were, just to squander it due to (current) bad management.



[#] Thu Jan 12 2023 09:54:33 EST from IGnatius T Foobar

[Reply] [ReplyQuoted] [Headers] [Print]

Actually if there is a transition to RISC-V it will take substantially *less* time than a transition to ARM. Or they will both be part of the same transition.
We're moving rapidly into an era where the ISA doesn't matter quite as much, as long as it's got a Linux or BSD kernel available for it, you're good to go. The era of requring a computer that is, to use an outdated phrase, "IBM Compatible" is waning.

The combination of open source operating systems and cloud technology are making all of this possible. And it's wonderful.

Agreed that ARM Holdings is throwing away all of the capital they've earned by being a solid, flexible, low power, and multi source alternative to the x86 monoculture. They're a leader but they don't exactly have lock-in. Add in the fact that RISC-V is zero license cost and they may have already blown it.

[#] Wed Jan 18 2023 17:44:23 EST from LadySerenaKitty

[Reply] [ReplyQuoted] [Headers] [Print]

Since its release, I have suspected the coming of RISC-V meant ARM's days are numbered.  With ARM's recent "you have to use our entire platform or nothing" decision is just going to accelerate their demise and then the real pressure for rv64 to have amd64-comparable purrformance will really ramp up.

 

Thu Jan 12 2023 09:54:33 EST from IGnatius T Foobar
Actually if there is a transition to RISC-V it will take substantially *less* time than a transition to ARM. Or they will both be part of the same transition.
We're moving rapidly into an era where the ISA doesn't matter quite as much, as long as it's got a Linux or BSD kernel available for it, you're good to go. The era of requring a computer that is, to use an outdated phrase, "IBM Compatible" is waning.

The combination of open source operating systems and cloud technology are making all of this possible. And it's wonderful.

Agreed that ARM Holdings is throwing away all of the capital they've earned by being a solid, flexible, low power, and multi source alternative to the x86 monoculture. They're a leader but they don't exactly have lock-in. Add in the fact that RISC-V is zero license cost and they may have already blown it.

 



[#] Mon Jan 23 2023 09:40:06 EST from IGnatius T Foobar

[Reply] [ReplyQuoted] [Headers] [Print]

Since its release, I have suspected the coming of RISC-V meant ARM's
days are numbered.  With ARM's recent "you have to use our entire

You're right, of course. ARM has enjoyed the breakout years of "you don't have to use AMD64" but they've scarcely achieved any sort of lock-in. Android is already architecture-independent (everything is in Java). iPhone/iPad is ironically tied to ARM but Apple is known to ignore that kind of problem even if it causes users some short-term pain. Linux and BSD have a history of easily following along on any architecture. That leaves Windows continuing its inexorable sink into the tarpits of doom, but no one really cares about that these days.

That means ARM isn't in a position where they can enjoy the benefits of vendor lock-in -- but as we all seem to be observing, they are acting like they are.
This does not end well for them unless they're simply planning to maximize profits in the short term at the expense of driving away all future customers.

[#] Sat Jan 28 2023 04:51:57 EST from LadySerenaKitty

[Reply] [ReplyQuoted] [Headers] [Print]

This kind of strategy rarely works, and is one I have seen send companies into dissolution.  Silicon Graphics, Lehman Bros, Enron, the US Fed Gov (hol' up ...)

Mon Jan 23 2023 09:40:06 EST from IGnatius T Foobar
This does not end well for them unless they're simply planning to maximize profits in the short term at the expense of driving away all future customers.

 



Go to page: First ... 33 34 35 36 [37]