Is there 128 bits technology in the works?

Is there 128 bits technology in the works?

Other urls found in this thread:

en.wikibooks.org/wiki/X86_Assembly/SSE
twitter.com/NSFWRedditVideo

lol,no

lol,yes

Who would ever need it.

CERN already has it, you just don't know about it

you won't ever need it. It's for them.

[GO]

Developers haven't even maxed out 32 bit yet so what's the point. Hell, even 16 bit is still viable in a lot of cases.

128 and 256 bit architectures exist, even 512.

High end GPU architectures are 256 or 512 bit

this guy

The architectures exist, there's just absolutely no reason to use it. Making pointers absolutely enormous so you can address more an 16 exabytes of RAM is totally useless for almost all computing currently done (possibly all? I'm not sure).

Who would ever need 64-bit?

Servers to properly address their 256GB RAM

You can order them right now but the urls are so large you need a 128bit compatible browser/machine already to place the orders.

The expansion in RAM capability with increasing address length is massively exponential, just because some people didn't see the need for 64 bit coming doesn't make 128 bit less insane.

lol, what are you talking about!
64 is only twice as many as 32. I don't want my 2017 computers limited to only twice as many addresses and registry space as a computer from the fucking 60s. It's ridiculous how slowly this has progressed! Other components are BILLIONS of times faster, but we're stuck with CPUs that handle twice the bits. We should be at 3GB architecture by now, but instead the industry just twiddles their thumbs.

this is bait right, please let this be bait

fucking idiot.

maybe they haven't really for individual programs, but have fun doing any sort of multitasking with 4gb or less of ram. And no, notepad and calculator doesn't count

Moor's law says that computers double every 2 years. So why did it take 50 years for CPU architecture to double?

Btw: I know this graph looks like it's increasing by set amounts, but it's because it's on a log scale.

My hard drive is 500gig. Who the fuck still uses bits?

excessive use of exclamation marks, of course it's not serious

already here but 0 application

we dont have hardware efficient enough

Moore's law is not an actual law but a theory

Yeah. That's the only problem with that post.

> So why did it take 50 years for CPU architecture to double?
8086 is 16-bit

because the need for higher cpu architecture isn't really there yet. Like, why do they still put 150-200hp 4-cylinder engines in cars when engine architecture has quintupled? Because its a waste when most roads dont accommodate higher speeds

There is 128 bits technology inside your computer right now.

>I need more than 4 GB of ram in 2017 so I can run 180 MB js/electron based shitware like Slack

RISCV-128.

But there isn't a use for 128-bit CPUs yet.

Advantage for 64-bit CPUs far exceeds the mere advantages of an expanded address space, due to the changes made for both Intel and ARM's transition to 64-bit. More registers, guaranteed SIMD instructions, and a few other benefits, depending on which ISA you're talking about.

128-bit would have uses for speeding mathematical programs, but not much else. 64-bit is pretty much perfect.

crysis

4gibibytes of memory

Word length and memory addressing my man. 128bits could address more memory than there are stars in the galaxy. 256? More memory than there are atoms in the whole universe.

I'm still hoping this isn't bait and you aren't this retarded.

Lets just look at the numbers

32 bit addressing:
2^32 = 4.294.967.296bit / 8 = 536.870.921 byte / 1.000.000.000 = 0.536 gb
64 bit addressing:
2^64 = 18.446.744.073.709.551.616bit / 8 = 2.305.843.009.213.693.952byte / 1.000.000.000 = 2.305.843.009,21 gb

now tell me, do you think we need to address more than 2.305.843.009,21 gb of data any time soon?

>wasting your address space on memory

In my architecture, the first of an address are the IPv6 address of the computer containing the device you want to address.

SSE

In principle, a 64-bit microprocessor can address 16 EiBs (16 × 10246 = 264 = 18,446,744,073,709,551,616 bytes, or about 18.4 exabytes).
> tfw fell for the 16EB meme

We've had it for decades.

Moor's law is also full of shit you retard. The rate that computing technology improves is declining over time.

>Moor's law

t. alberto barbosa

In this sense we had 128 bits since MMX, right?

in b4 quantum computers more than make up for all the delay of last years

Yeah sure, in the year 642600.

The road plan for 128bit memory is fairly extensive.

My 8GBs of memory needs it

We already have 512b internal data networks for cache lines and SIMD systems with 128-512b total register widths.

There's just no near-term need for 128b integer math, 128b floating point (outside a very small number of scientific simulations) or 128b address spaces.

yes
en.wikibooks.org/wiki/X86_Assembly/SSE

Nope, mmx uses 64bit-sized registers.

Wasn't ZFS 128? It's been around since the mid 2000s. Are you just not paying attention?

>the only reason to have a larger word size is to get more memory

It's funny hearing people talk like that about crysis now. It's almost nothing to run these days.

kys you're self my man

NASA and CERN have them for very very precise and memory intensive calculations. Otherwise there's simply no need

>open mandelbrot
>zoom in
>hit max
>sure wish I had 128-bit

there's no need for 128b architecture for an average customer. twice as many bits to take care of and twice as much memory being used

These are just general architecture and ISA improvements and have not much to do with a change from 32 to 64 bits.

Just use PAE :^)

> looking at thing famous for appearing as basically the same at ever zoom level.
> annoyed that you can only zoom in a few million times to see something that would, again, be the exact same fucking thing except twisted slightly differently

Sup Forums, everybody.

verdun stella

I kek'd

They're talking about transistors and logic gates lol

We've had 128 bit technology since '99, user.

...

you don't seem to understand powers of two. going from 32 to 64 bits doesn't double the number of values you can represent, it multiplies it by 2^32. 32 bits can represent 4 billion values, 64 bits can represent 18 quintillion.

There is literally no need for it at the moment. The only real benefit we would get in the near future would be more registers.

a computer is addressing bytes, not bits, you shouldn't divide by 8

Just use multiple precision integers then. It's dead fucking simple