/tech/ - Technology

china pls save gaming/rendering pc's

>>13164
Ultimately to overthrow the dominance of x86 something that's faster and cheaper needs to come along. Currently RISC-V isn't very competitive performance-wise. It will take a while before RISC-V is mature enough for some manufacturer to bet on it and throw money at a high-end batch of chips, and those will likely be more expensive than similar x86 ones.

I can see China making a better and cheaper x86 in a shorter time-span than it would take for RISC-V to mature enough for me to put one into my desktop. And let's face it, if hobby/enthusiasts aren't using it, then it will never be as prolific as x86 is.

Long term however, open design just like open source code will win. Because it's already clear that this is the way to develop technology and will outpace anything developed in a secret silo. This will become even more true as Asia develops further ahead of the united states and becomes technologically dominant, and Intel's chips having a 100% failure rate is only going to accelerate this change of tech dominance.

I just think the Intel thing is part of a larger ongoing trend of everything in the west going to absolute shit. Think about it, no one cares about their job and does the bare minimum. That leads to complicated stuff like software and hardware not being as well understood or well-designed as it was when previous generations worked on it.

What we're in is a negative feedback loop of everything getting progressively shittier. And it doesn't help that managers want to extract money from everything: Google stopped giving you more and more gmail storage every year, google search results suck now (probably because of their management), Windows 11 is a nightmare to use compared to any of their releases in the past. The management thing makes workers not give a fuck which results in tech that is broken intentionally and unintentionally.

>>13165
>china pls save gaming/rendering pc's
lol, yeah that's not entirely wrong.
But the guys that are currently working at intel probably could go for RISC-V, in case Intel capsizes.

>>13166
>Ultimately to overthrow the dominance of x86 something that's faster and cheaper needs to come along. Currently RISC-V isn't very competitive performance-wise. It will take a while before RISC-V is mature enough for some manufacturer to bet on it and throw money at a high-end batch of chips, and those will likely be more expensive than similar x86 ones.
Why would it be more expensive ?

>I can see China making a better and cheaper x86 in a shorter time-span than it would take for RISC-V to mature enough for me to put one into my desktop. And let's face it, if hobby/enthusiasts aren't using it, then it will never be as prolific as x86 is.
Would you be able to get any of these x86 processors ? Isn't there some licensing shenanigans going on that prevents others from selling x86 processors ?

>I just think the Intel thing is part of a larger ongoing trend of everything in the west going to absolute shit
Feels true
>it doesn't help that managers want to extract money from everything
>What we're in is a negative feedback loop of everything getting progressively shittier.
The great enshitification

How universal is that tho ? i mean is there no counter tendency that goes for making good stuff ?

>>13168
>Why would it be more expensive?

Historically chips that try to seriously compete with x86 have been on more expensive platforms, IBM POWER and Sun SPARC come to mind.

I guess it's because it takes significant resources in R&D to become competitive and Intel and AMD already have economies of scale to keep their prices down.

>>13168
>Would you be able to get any of these x86 processors ? Isn't there some licensing shenanigans going on that prevents others from selling x86 processors ?
You know I thought it was just that China doesn't give a fuck about western licenses but it appears that Zhaoxin actually has a license to make x86 products. So I mean unless there are some crazy tarrifs and import bans in north america I don't see why you can't get them.

>is there no counter tendency that goes for making good stuff ?
Okay this is going to just be a rant about open source, because I consider open-source to be that counter-tendency, but consider this: everyone in tech is busy and burnt-out for these huge tech monopolies either directly or indirectly. And so even with open-source a lot of development time today goes into building and maintaining things that the industry needs. Remember even though it's "free and open source" that doesn't mean there isn't a lot of money that goes into it.

So you have software that is ever-increasingly more complicated: like take the stack of tech that you need to serve a modern website, you need (minimally): a linux server on one side and firefox on the other side. Both of these projects are humongous, like incomprehensibly large and border-line unmaintainable. So lots of people are trying to make these things better, but for what? So we can add yet another W3C spec into our already over-complicated browsers to render a facebook Like buttons a little faster? At the end of the day that's what it's driven by.

And I don't see people embracing simplicity. I was once really excited for truly open source linux smartphones like the Pinephone. I was expecting the whole community of Linux users to switch to something like that and rather quickly get it to be on the same feature-level as Android, but years later it's dubious that you can even use the modem for calling and LTE with a phone running Linux on the bare-metal. The interest just isn't there, and this is why I said I think most devs are just too busy with corporate jobs.

>>13169
>Historically chips that try to seriously compete with x86 have been on more expensive platforms, IBM POWER and Sun SPARC come to mind.
Yeah that's true when IBM open sourced the POWER ISA, a dev-board with a processor and some ram cost 4 grand. That said it was really powerfull, so the value proposition wasn't that bad, but not many can buy in at that price point.

>I guess it's because it takes significant resources in R&D to become competitive and Intel and AMD already have economies of scale to keep their prices down.
Some of that has improved tho, chip foundries with great economies of scale will produce cost efficient chips (in terms of hardware manufacturing) for newcomers too. RISC-V already has a lot of the plumbing done, so that should chip away some of the R&D entry barriers.
Maybe i'm just being naive.

>>13170
>So you have software that is ever-increasingly more complicated: like take the stack of tech that you need to serve a modern website, you need (minimally): a linux server on one side and firefox on the other side. Both of these projects are humongous, like incomprehensibly large and border-line unmaintainable. So lots of people are trying to make these things better, but for what? So we can add yet another W3C spec into our already over-complicated browsers to render a facebook Like buttons a little faster? At the end of the day that's what it's driven by.
There's still people trying to make a new browser from scratch, like the project called lady bird. And there's some hope that more automation in coding might make the complexity of these more manageable.

>And I don't see people embracing simplicity.
There is stuff like the Gemini protocol. So it's not like there is no interest in simplification.
>I was once really excited for truly open source linux smartphones like the Pinephone. I was expecting the whole community of Linux users to switch to something like that and rather quickly get it to be on the same feature-level as Android,
I think that is a problem of expectations, Linux was marginal for a very long time until it took off. This probably has to cook for a while longer.

>but years later it's dubious that you can even use the modem for calling and LTE with a phone running Linux on the bare-metal. The interest just isn't there, and this is why I said I think most devs are just too busy with corporate jobs.
I think that the problem might be the cellular modem stack. It's soo much legacy cruft. Left-overs from ancient technology limitation that are nonsensical now, some of it is because of the way cell towers networks are organized and so on. Eventually a new fresh communication stack will become necessary and then it'll be possible to go the open source route from the start.

>>13165
>>13166
honestly the ideal would be asia sprinting towards technological supremacy just enough for the us to shit ourselves and start putting cold war level spending into r&d which leads to hyper deflation of chip prices in the world. diversity of chip production would mean a very nice era in consumer hardware for the consumer- that in addition to open source/open concept designs accelerating the opposing groups. you can't just delete RISC-V from existence. It's already out there.

With software like GNU/Linux and Android being open source, China and other countries could start developing consumer and business software to their specific needs and start exporting them around the world. This would force software everywhere including the west to compete on the same software.

China could start pumping out risc-v smartphones, tablets, and computers running android or linux and we could finally have the year of the open source linux desktop.

I'm very curious to see what will ultimately spark the big move to the next architecture. There's a lot of "baggage", a lot of inefficiency in computing that comes from technical design choices of 40 years ago. But it's what we have. And moving to something else means eating shit for an unknown amount of time, and so it won't happen until something very significant forces it.

Fault of gamers for demanding technology so complicated they couldn't understand it if they studied for 40 years.

So chip manufacturers put in all this cheating into the chips since the early 90s and now they are riddled with bugs, security holes, and breakage

What a joke

>>13183
>which leads to hyper deflation of chip prices
that sort off happened, for cost-per-processing
you can get a micro controller with 1980-85 era desktop computer specs for 25 cents

Or do you mean lowering the commodity price of chips all together ?
As in cost-per-chip-area
That would come from improving production technology to reduce the manufacturing cost.

>start pumping out risc-v
>finally the year of the open source linux desktop.
So Risc-v makes TYotLD happen ?
Maybe…

>>13185
>I'm very curious to see what will ultimately spark the big move to the next architecture.
same here

>>13188
good point that it already happened.

I think the first two have actually happened already and will continue to happen as long as there is competition.

As for the third that I mentioned, I'm honestly just super hopeful. I wish it would happen- assuming the us doesnt just sanction chinese chips and makes two markets with no competition

>>13218
>good point that it already happened.
It already happened for processing cost, but for the die-area of chips, those haven't really increased that much. There are large chips like the size of post-cards, but those are not cheap. The wavers (chip-blanks) haven't really scaled up that well, it's probably because growing large silicon crystals is kinda hard.

>I wish it would happen- assuming the us doesnt just sanction chinese chips and makes two markets with no competition
Yeah RISC5 is open-source, so if they segregate the market, both sides still can copy each others designs. Which kinda happens in tech a lot. Also its capitalism, sanctions can't stop anything from getting sold, if there's money to be made somebody will find a loophole. I think the point of these tech sanctions might be a shake-down. It could be people who are trying make money by first blocking Chinese goods so they can sell access for a premium.