I've heard Moor's law is ending in 2022. Why should i bother buying any tech until then if the final computers will be out so soon?
>>54189323
Ending in 2022?
It has already ended.
>>54189355
I mean the smallest transistor will be out by 2022. Not the slowing down until then.
So what is the next big breakthrough for consumer technology?
>>54189459
Superconductor CPUs so you can have ridiculously high speeds because you don't have to worry about heat
X GHz CPUs once graphene is able to be cheaply manufactured. Carbon nanotubes will be able to suck the heat out of the core efficiently enough to allow for more complex cooling systems (at the level of abstraction that modern heatsinks are on.)
I think if one were to take the lay understanding of Moores Law (doubling speed ever 18 months as opposed to the technical definition) then the law would actually produce a conservative prediction of processing power.
If anything Moores Law will die due to massive increases in speed in 20-30 years, although it is surely dead already due to speed increase slowdowns in this local time scale.
Glad I am young enough to see at least the beginning
>>54189665
>implying graphene will ever be cheaply manufactured in our life time.
>>54189459
em-drive equipped dragon dildos with built in 4G hotspot
>>54189459
Quantum photonics
>>54189459
I can't really come up with anything that would be plausible in today's world
3D Memory was really the only latest technological "breakthrough" from 2013 apparently with HBM and 3D XPoint starting to show some light in 2016
If I could stay in 2011 with the same technology, I would be satisfied, as long as I could buy and pirate new content.
>CPU is fast enough and cheap enough to play most current games.
>Internet connection is fast enough for 1080p or 720p pirate streaming.
>HDD space is cheap enough for all my digital library needs. Porn and entertainment luxury.
>SSD is cheap enough for boot drive.
>>54190019
>current games
>new content
>staying in 2011
I think moores law is finished. Four years ago I was looking into building a pc but couldnt do it because I was a poor high school student. Today I am finishing my build and ready to order, and almost nothing has changed in the past four years save the gpu. People are still buying i5 6600ks and i7 6600ks, 8 gigs of ram, etc.
Literally nothing has changed in four years, sans gpus.
ssd's evolve very quickly and will evolve for long too
hdd's evolve too
>>54189323
To preserve remnants of the past before the divergence
>>54190019
Why not let it get better?
Also what new content for games is their? Hasnt been a good AAA title in years
>>54189832
I sure hope so. The worst part about sites like Phys.org is that they get me so hyped but as you say most of that cutting edge stuff is all centuries away from being commercially viable
>>54189323
SSDs will get faster, cheaper, denser and smaller
GPUs will be more efficient, better and smaller(maybe)
Display technology will get sharper, cheaper and more efficient
New standards come out like USB C or muh no headphone jack
It's not all about the CPU you know.
>>54190265
>skylame
>4 years ago
>implying you aren't talking out of your ass
>>54189323
>Cirno abuse
>>54190456
What he meant is that processing power has barely changed.
An overclocked Sandy Bridge processor will keep up with a Skylake processor with ease, possibly even beat it.
There has been no need to update processors in the last 4 years for desktop users and until graphene transistors are commercially viable, the only increases in technology worth following will be storage and GPUs.
>>54189500
Oh, God... Forgive this man, for he is not ignorant on purpose
>>54190433
SSDs will run into the same problem where transistor size can't improve. Same goes for GPUs, which are essentially the same as CPUs.
>>54190474
THE STRONGEST...gpu?
>>54189323
people have been saying this shit for decades. Moor's law isn't ending.
>>54190575
Yeah, eventually. But not in 20 fucking 20
>>54190680
It is approaching its limit.
Traditional lithography only goes to 10nm.
They have to use very problematic extreme ultraviolet machines for the next years.
Quantum tunneling also becomes a huge problem.
if quantum computing is figured out then it won't end
>>54190575
GPU's suffer to a lesser extent due to form factor - they can afford to be much larger (even if to tackle density problems) and can easily bring far, far more cooling than most cpu's will be expected to have. Consider the Noctua D15 - built to handle 200w or so on a cpu next to the likes of pic related - a gpu cooler built to handle over 300w.
>>54189323
All Moore's law does is state that every 2 years computers become twice as powerful.
Moore's law has been done in the CPU space for half a decade now. Computers will keep getting faster/more complex, at a slower rate, so it doesn't really matter when you buy, it just means you can keep using the same computer for longer periods of time before it's slow enough that you really NEED to upgrade for it to remain viable.
>>54189323
You misunderstand the prediction.
Even if the smallest and fastest transistor is developed and put into commercial production by then, that does not preclude further advancements in chip architecture, more pipelines, more cores, and far faster chips using future breakthroughs like full 3D chips, a non-silicon semiconductor, photons instead of electrons, etc.
>>54190847
>All Moore's law does is state that every 2 years computers become twice as powerful.
No, it states that every 18 months, the number of transistors in a chip doubles. It says nothing about speed, just scale and complexity.
>>54190856
The problem is cost per transistor and power consumption.
Sure you can add more cores.
But it will get much more expensive, and you can forget about small portable devices.
>>54190680
Correct. It isn't ending. It ended years ago.
Not like it makes any difference. Dennard Scaling is also dead. Even if we had exponential increases in transistor count, that would no longer lead to exponential improvements in performance.
>>54189404
we are switching to graphite which can be clocked at thousands of ghz
>>54190967
>we are switching
If you work for IBM, sure.
For the rest of the world? Graphene is a looooooooooooong way off.
>>54190954
Cost per transistor can and will continue to drop.
Power consumption can also drop as well, through developments in power management and overall architecture. Just look at different chips on the same manufacturing technologies.
>>54189832
What is that thing supposed to be?
>>54190550
Intel is shifting away from the desktop market. Once low latency gigabit is widely available you'll see a massive shift away from desktops to cheap streaming devices.
>>54190207
2011 hardware + new software/consumables .
>>54190967
Quantum photonics aside from being quadrinary allow for hundreds of terahertz.
>>54189323
>the final computers
this is literally the worst board
>>54191310
>the final computer
That is like asking /k/ what is the best weapon.
(BTW, it is a 16 inch .223 wylde SIG MCX)
>>54191430
>>54189323
When they reach the limit of silicon, they'll probably switch to some other material. Otherwise, they'll build out and expand cooling solutions for a few years.
>>54189832
What is aluminum
>>54191430
Absolute shittiest taste in everything except manufacturer. Barrel length, caliber, and model all garbage. Apply yourself.
>>54191545
>What is aluminum
An abbreviated form of aluminium?
>>54191615
>>54190320
>Why not let it get better?
the width of semiconductors is approaching a width where there's unpredictable quantum behavior of electrons
>>54191658
So it becomes a quantum computer?
>>54192068
That's not how this works
that's not how any of this works.
>>54191658
That is a good point, but the cost of materials, cooling, scaling, distributed processing, and new ways to network transistors, like 3d, will likely be able to effectively carry moore's law past 2022.
When quantum computing hits, it'll likely do so in time to put us back on the moore's law curve, but we'll probably be lagigng behind it.
>>54190019
>developers learn to use what they got and instead of being inefficient pricks the learn efficiency
>>54192238
falling for the quantum computer meme
>>54192302
>i can't install gentoo on it so it's a meme meme
>>54189832
>implying iron will ever be cheaply manufactured in our life time
Honestly we should just stick with bronze, it's so much cheaper
>>54192891
Bronze age best age. Fuck these iron and steel shitters.
>>54189665
There's more to 2d nanosurface electronics than just graphene man. Graphene is only the conductive bit.
>>54192961
Honestly I think flint was better, more manly and less shiny bullshit.
>>54193516
Nah bro I live there.
Between the water and the violence I am surprised I ain't dead yet.
>>54192068
no. you can't reliably say "the current is on the left side of the conductor" when the distance is that small. quantum computing involves having more than 2 states of "bits"
So like what if it is a circle? We develop machines that are wood and shit and think they are the best then work our way up to crystals eventually and their quantum storage shit but all along we destroyed the crystals of the races before us?
>>54189459
SiGe 7mm chips.
>>54189323
There are two possible futures from this point forward. In the first, advancement slows greatly and everything slowly moves to the cloud so that the average consumer can still experience their devices getting thinner and more powerful, enthusiast computers will start being built using server hardware and multi processor builds will become common. In the second, an inexpensive way to produce graphmeme and carbon memetubes in large quantities is found soon and you get to see a quick advancement to a world that would seem science fiction with smartphones and laptops that can either charge in seconds (supposedly they'd be able to make capacitors with energy density comparable to lithium ion batteries) or have batteries that last 10x longer, electric cars becoming better than gas cars in every way due to the problems with range/recharge time being eliminated, traditional guns being replaced with rail guns or possibly other energy weapons, ridiculous body armor that manages to be both much stronger and much lighter than current body armor, and probably power armor to go with that.
>>54189323
You can't make a CPU node smaller than 5 nm. It's impossible. Moore's Law is already dead btw.
>>54193986
>You can't make a CPU node smaller than 5 nm
Why not? Can't they make transistors the size of a single atom, or around .5 nm?
At how many different years has/will Moore’s law end(ed)?
>>54194091
All of them.
>>54189459
AI according the the pajeet of Google
>>54194071
No because then you can never know where the electron is because of quantum mechanical uncertainty so things start to leak and everything goes to shit. IBM has already made a 7nm transistor.
>>54192360
>>54192302
>Not compiling the kernel to work on a quantum computer
>>54193545
You get one Internet.
>>54190474
I don't understand why he doesn't make a liquid cooling system with Cirno's piss
the strongest coolant!
>>54189323
>thinks a down global economy and monopolistic markets causing temporary tech stagnation is a long term trend
people thought everything that could be invented was in the 19th century too, bud.
>>54189323
It is already officially dead:
http://spectrum.ieee.org/semiconductors/design/the-death-of-moores-law-will-spur-innovation
http://spectrum.ieee.org/semiconductors/processors/the-multiple-lives-of-moores-law
http://spectrum.ieee.org/semiconductors/devices/the-status-of-moores-law-its-complicated
>>54191210
Its Eileen from Bloodborne sitting at a computer desk, then put through Google's dream-shit filter.
To be honest family, the thing that spurned the world to innovate in computers and silicon was WWII and the Cold War.
There aren't any wars, so innovation is stagnant, relatively speaking.
