>"What a D-Wave machine does in a second" would take a conventional computer with a single core "10,000 years" to perform a similar task
>1,097 qubits
But can it run Crysis?
>>51784141
It's fucking Vaporhardware
God I love memes
Quantum anything it's just a buzzword at this point
>2015
>single core computers
It can add 1+1 a quadrillion times a second
Exceptional
>>51784168
Maybe at 480p and opengl turned off.
A friend on Facebook said "...I imagine it could be released into the general public in the farish future with it's own OSs that could work for games and shit.
Would be interesting to see how it develops."
Yes or no?
>>51785914
no fuck that quantum scam shit, what we want is sub 1nm carbon transistors in the cpu.
>>51785978
What about muh quantum tunneling? Did they solve it yet?
>>51785978
And what are we at now?
>>51785914
Block your friend.
>>51785914
Your friend is an idiot and so are you for even thinking about such a retarded statement.
>>51785978
>what we want is sub 1nm carbon transistors in the cpu.
Not possible.
Anyway, D-Wave quantum computers can't run Shor's Algorithm so they're basically useless. I can't think of another quantum algorithm that is actually useful. It's certainly much slower than a normal computer.
>>51786050
Hey, don't lump me into it. I was just interested to see what you guys thought about it.
>>51785978
>Quantum scam shit
>Asks for 1nm
Do you even physics
>>51785914
fucking christ
>>51785914
Kill your friend.
>>51786051
>I can't think of another quantum algorithm that is actually useful.
Hooow about:
- Grover's algorithm (!)
- Quantum counting
- Simon's algorithm
And maybe more that WP will point you at:
https://en.wikipedia.org/wiki/Quantum_algorithm
Or is the problem with that D-Wave actually that it can't do Deutsch-Jozsa?
>>51785914
It's difficult to say where the technological ceiling lies with quantum computers, but as it is currently understood they will nit be better than regular computers at everything.
They are likely going to be useful for database queries and simulations of quantum physics, but they are unlikely to be useful for laymen.
>>51784194
>Implying thats not the basis of CS
>>51786233
No, it's AWP
>>51786026
like....14nm silicon, aka skylake.
>>51786390
Aren't CPU (and GPU, I guess) advancements slowing down? Sorry for the ignorance - I don't really keep up with the manufacturing processes. I've just read (on /g/, I think) that each successive generation isn't advancing as much as the one before it.
>>51786419
CPU, yes.
GPU, not as much. The chips aren't getting faster as much as the other components though.
"God doesn't play dice."
>>51785914
You're both fucking morons. Where are you going to isolate your qubits in your house, are you going to keep them in a box??
>>51786430
So what do you think the next big leap in CPU technology will be? Old mate above said smaller transistors, what do you reckon?
>>51786480
> he says the thing
> "you're both"
Mate come on don't be mean.
>>51786485
>the next big leap in CPU technology
Well, if graphine doesn't turn out to just be vaporware or meme technology then that would probably be it.
>>51786499
Can you elaborate as to why? I've heard of graphine and it's near-wizard-like properties, but I'm in no way an expert.
Not a real quantum computer. The "speedup" is just a measurement of the evolution of the particular system that is the D-wave, compared to the speed of solving the same problem directly on a classical computer. In other words, the only quantum problem it can solve (with a non-constant speedup) is the problem of simulating itself, which can be said of any quantum system. It's still pretty awesome, though.
>>51786485
I have read about the 3d transistors from Intel, can someone explain to me how this would affect computing in a layman standpoint?
http://www.scottaaronson.com/blog/?p=2555
>Thus, while there’s been genuine, interesting progress, it remains uncertain whether D-Wave’s approach will lead to speedups over the best known classical algorithms, let alone to speedups over the best known classical algorithms that are also asymptotic or also of practical importance. Indeed, all of these points also remain uncertain for quantum annealing as a whole.
>>51786502
once they get below a certain size, I think its like 5nm, quantum tunneling becomes a problem. Also there haven't really been many graphine prototypes yet so nothing is 100% proven.
>>51786419
ASML the biggest (~80% marketshare for current technology) chip machine manufacturer guarantees Moore's law will hold till 2022 and they're currently working on pushing it to 2024. So in the near future I think the pace isn't actually slowing down.
>>51786515
Well no real big revelations here. Just more transistors in a smaller space that can still dissipate enough heat.
We aren't anywhere near current performance technology with graphene or light transistors as far as I know.
Other interesting advancement would bring forth an entire new computing paradigm which would unify a processor and memory in a persistent way. But I don't really know the estimates on this.
>>51786445
Hurr, so dat meenz God planned everything frum da start?
Atheists have been BTFO
BTFO
T
F
O
>>51786515
Just think it's like going from one floor apartments to multi-floor apartments. They will be using some tricks to get the heat out so it doesn't turn into a melting furnace, but it won't feel any fancier.
>>51785914
>A friend on facebook.
Fucking normies.
>>51786083
>>51786083
>hey, don't lump me into it
What a freakin normie
>>51784168
>Crysis
>not Doom
>>51786020
you dont "solve" a fundamental underlying principle of quantum mechanics.
>>51785914
>>51786083
This is some cringe right here.
>>51784141
Meh, I'll wait for Zen
>>51786419
Single core CPU performance is totally separate from process node.
The issue with serial integer performance is that all of the low hanging fruit have long been picked, the arch itself is changing little at a fundamental level. Continually trying to refine the exact same thing over and over yields only diminishing returns.
>>51786485
Conventional scaling will persist down to 5nm or 3nm. The Silicon roadmap has an awful lot of life left in it.
What happens after that is impossible to guess since it would be in the realm of theory at present.
>>51786515
Smaller planar transistors exert less control over their channel, current leakage increases, and they lose the ability to hold an absolute off state.
"3D" transistor topologies like intel's Trigate process and other FinFETs change the orientation of the gating structure and increase its height to deliver more gating surface area over the channel.
That structure allows you to continually shrink feature size without leakage continually increasing. Ultimately it improves performance per watt as performance per die area improves.
>>51784141
OP did you read all the comments on here and notice that you're trying to have a conversation with kids?
>>51788413
>5nm or 3nm
current hard limit seems to be 5nm, anything smaller and quantum tunnelling exists, this is not "fixable" as it is just a fact of how electrons work.
>>51788835
I don't know what ridiculous pop sci clickbait you read, but you need to stop. GAA structures exert virtually perfect control over their channel, and they've been fabbed smaller than anything we've ever produced. A 1nm GAA structure is not impossible, Applied Materials themselves have laid out a roadmap to production ready 3nm GAA nodes.
>>51788890
Well considering intel publicly said earlier this year their 7nm process will likely be indium gallium arsenide and NOT silicon based, so saying that 5nm and 3nm on silicon is going to happen is just stupid on your part, sure it MIGHT happen. But intel is already moving to other materials on 7nm because of quantum tunnelling, let alone 5nm or less.
>>51789034
You're demonstrating an absolute lack of knowledge on a subject.
What venues intel chooses to pursue are directly influenced by their own IC design requirements, and economics are secondary. They're already using triple patterning on their 14nm parts, and that obviously can't continue forever lacking an EUV source, but they can't continue increasing fin height either. That is the real issue. FinFETs are as much of a stop gap as everything else, and certain materials are better suited for use in one topology than another. They're going to employ more SiGe, but that is still fundamentally a silicon substrate.
Don't skim through articles way out of your league then try to correct someone actually studying solid state electrical engineering. You just sound like a child.
>>51786390
I thought some nerd at IBM made a 7nm silicon gallium
>>51790194
It wasn't some nerd, it was a full team, and they're now functionally GloFo employees. All of that IP belongs solely to Global Foundries as part of their recent acquisition.
The material you're thinking isn't gallium, its SiGe aka silicon-germanium. They're using it directly in the channel. A number of mobile chips today use it in various places, like around source and drain wells, its use in this 7nm process is very aggressive.
>>51786051
>>51786095
Pretty sure it was a joke
>>51789182
Do you think EUV lithography will ever be viable? From what I've read in various places, it's coming along, but it's really hard to ramp up and might end up not really being useful due to being too expensive/not efficient enough.
Also do you see any promising new production methods (beside EUV lithography) that are currently being researched?
>>51790614
At one point most everything seemed like an insurmountable feat, and here we are. Every technical hurdle simply comes down to a matter of man hours. EUV will become common fair just as 450mm wafers will be. EUV specifically seems to have one teething issue after another because primarily it is new, and more important the precision of the wavelength used is mind boggling. Imagine being an engineer sitting in a room sipping coffee only to have your boss walk in and tell you that you need to create a smaller wavelength light source then man has ever produced. Not only thing but variability has near no wiggle room, and you have to maintain this light source at such high intensity that it'll break a bunch of the conventional tooling you have.
EUV machines exist and function today. Its a safe bet by 2017 the last kinks will be ironed out and they'll be used in production lines ramping up for volume production of consumer chips.
No one likes having to develop new tooling, developing a new process alone is now an investment of well over $2 billion, so everyone is trying to make everything compatible with conventional CMOS lithography tooling. EUV and larger wafers are going to be the big industry leaps for the next few years. Fabs themselves are going to explore III-Vs, TFETs, GAAs, and other novel structures to push area scaling and perf/watt. Thats where all the excitement is.
>hurr dhurr it's not interesting or impressive because it's not a fucking warp capable starship computer core right off the bat
/g/ is fucking retarded
>>51786209
>they are unlikely to be useful for laymen.
That's what we used to say about massively parallel processing systems. Now we're up to 6 - 8 cores pretty commonly, and likely going to break into 20+ territory in not too long.
Also, when I was in college we were taught that it was impossible to do any sort of predictive optimization to execute instructions simultaneously, but now there's HyperThreading, which is exactly that.
>>51790799
That's more or less what I was thinking. Thanks for the answer. I've read about the existing machines, apparently one of them broke early on which is what caused some delays.
You stupid faggots do realize this is the first step to artificial superintelligence right?
>>51791000
Even an ant has superintelligence compared to you.
>>51791050
Woah you totally burned me brah. #REKT
High five
>>51784168
Probably.
>>51791000
Synapse and the associated DARPA research are the first step to a real artificial intelligence.
Quantum computers are large number machines. Our first CPUs were integer workhorses, then we added FPUs, GPU add in cards, a dozen different media accelerators, audio DSPs, some Xeons now are including FPGAs even. Our approach to technology is an inclusionary one. We develop separate parts that excel at certain tasks then we integrate them together. Quantum processing units will eventually end up as another coprocessor to a larger system. As with everything else we will discover new ways to exploit their nuances and employ them in novel ways, but its incredibly unlikely that they will come to replace conventional systems.
>>51784141
what computer these days has a single core?
>>51791085
My mom's.
>>51791063
Probably not
It doesn't run on windows fucktard.
>>51791085
relevant photo
So where can I buy a quantum rig? my parents are super rich so possibly I could afford one for Christmas if I really beg my dad but I can't find anywhere to order online
can't wait to see all you envious faggots when I post a quantum battlestation
>>51791154
D Wave is a commercial venture, they'll sell you one of their systems for around $10,000,000.
I'm posting this from a D-wave rig. :^)
>>51791154
$10 or $20M would probably do it, contact D-Wave, they generally require you to have an idea of what you want to use it for though.
>>51784141
but
quantum computers can only at best do 3x5=15
>>51791339
Meh, mainly shitoposting and games :^^)
can't even imagine the graphics on this thing
>>51790406
Ah yes, thanks for correcting me - I can't wait to see 7nm come into mobile phones and tablets, that would be awesome
>>51785978
The CPU can't be that small. At that size you would have to deal with the uncertainty principle, which means that the CPU wouldn't work.
>>51784141
but can it run Minecraft?
>>51784141
ITT: people mad about their child porn not being secure anymore
>>51784141
Doesn't it only do what type of calculation though?
What's the point desu
The only real use for quantum computing wouldnt even be to make a independent computer system. It would be way better utilized in processing unit addons to real computers.
Since you know, you need a real computer to analyze the qbits in the first place.
So how long has the NSA had this?
Could I entangle photons with an alien on Alpha Centuri and communicate porn with him using this?
>>51784141
wait, is that real?
>>51791676
yes
>>51791676
Yes. All your encrypted volumes are now worthless.
>>51791687
>>51791690
I thought there would be a bigger reaction to a quantum computer...you guys don't seem surprised at all...
>>51791557
Electrons can be totally confined within a structure if the channel is fully insulated on all sides. Thats the point of a quantum well FET.
The underlying issue with conventional scaling is that matter is a finite thing of specific dimensions, and we can't make things smaller forever. When we get to that point where front end structures are single digit atoms across we have to use quantum junctions or optoelectronics to continue area scaling.
>>51791690
>>51791576
If Google wants to waste their time decrypting your Boku no Pico collection.
>>51791716
They had the quantum computer for a while now, but it still hasn't that much performance
>>51791716
they've been around for years, hell DARPA announced a year or two ago they have had fairly advanced quantum computers for over a decade they were just classified projects.
>>51791690
Time to use quadridimensional shapes to encrypt your files...
>>51784168
wow that was a really funny joke guy
