I'm gonna get flamed to no end for this but...

Page 3 - Seeking answers? Join the Tom's Hardware community: where nearly two million members share solutions and discuss the latest tech.
Toggle sidebar Toggle sidebar
They are already creating programs that learn, so you can teach the program to be a spreadsheet etc, they are kind of pointless but they could prove useful for putting in exploratory vehicles where humans cannot go.

I agree with you lot that doing things randomly wont work.

Search google for learning programs, and Advanced Heuristics.
 
Ummm let me take my socks off
Click to expand...

Ha I just had to drop my pants. 😳

Back to the subject at hand 😳

A computer computes, to compute you need something to compute on.

The operator needs an operand.

To change state you must have state.

To solve the problem you must define the problem.

To represent something you need an to define it.

A tree falling….
 
what if memory could be made from a calculation or the answer to a calculation is repeated over and over again while another processor reads the calculation... this means you would need a heck of a lot processors to 'make' memory where there is none, thus the system is able to 'function' without memory even though you have 'memory'

Ara
Click to expand...

On the circuitry level, that is very similar to how the latches that make up memory ciruits work. What you are describing is essentially a watered down definition of memory.

You can in fact have a Processor without memory,for example, the ALU inside a processor, it would be useless though, with no source of input and output. And it would not be a computer. the most basic computer has three components: Processor, Memory, and an I/O system. If you don't have these, you don't have a computer.
 
I can prove that the tree falls. In space when a star many many lightyears away explodes we don't see it happening. No one experiences it, not until afterwards when the light from that star reaches us, for that light to do so the star has to have exploded, no one saw it happen, but it did. Time wins again!
 
Grasshopper your Zen needs training.

The saying goes

“If a tree falls in the woods and nobody is around to hear it, does it make a sound (noise)?"

The real quandary of this statement is the shared definition of sound (noise). Is sound just vibrations moving through a medium or is it the ability to pickup these vibrations that makes the sound? We often talk of the speed of sound but is it just the transmission of vibrations or the ability to produce and receive over a medium that constitutes speed of sound.

A good way of reflecting on the star would be to ask.

“If a far away star explodes and nobody sees it does it shine?”
 
don't take this in the wrong way, but I could think of a few more quests for you...for instance, if one so to say would have a stiffy and you would remove his ears and then you place him horizontally with his dick pointing down...if you spin him around, would the air blowing through his ears improve 3D-performance ?

don't be offended by this, it's just an idea
 
Look the tree falling in the woods and making a sound means...

Did it actually go through the process of happening thereby producing a sound, it means: Do things go through the process of happening if not witnessed by humans?

So if you interpret it that way the answer is yes, it did happen.

Sound is just a movement of molecules so is light therefore the real meaning is ," Did it affect the universe" and the answer is once again.

"You under-estimate skill of my Shaolin."-Some crappy movie (the 'the' was not there)

And:

¿SoD?
 
Do things go through the process of happening if not witnessed by humans?
Click to expand...

yes the laws of physics prove this to be true..

as for your star, we can witness a star exploding just what we see happend in the past cause of the amount of time the light took to get here.. ive looked up before and seen a star just dissapear.

theoretically the idea of a computer without memory is possible if you could enter all the instructions at the very same moment, and get the output at the very same moment you put in the input, but thus the output would flash onto the screen for less than a second and it would be useless anyway..
 
A purely hardware based computer could exist, where every program is in the hardware and not 'executing via the CPU / registers' and it could be powered by fire (heat -> energy -> power source).

Video cards / GPUs are 'similar' to this at 'some' level, but they still have memory and registers, although they do execute a predefined 'program' in hardware.

This differs greatly from just heating silicon to a level where anomolies occur (stray energy causing hardware to perform in a unexpected way).

If it where possible, without memory (of any kind), chances are it would be in the 60 - 120 degress Celcius range using specialist designed hardware that can only execute predefined functions (although that doesn't mean an entire OS plus application suite couldn't be stored on silicon, it would require a large amount of silicon, or other building block for the 'computer' though... then again early operating systems could fit in the cache of current CPUs, but it would still be classed as 'memory' at some level - as it would be more similar to executing a ROM without shadowing it within the same die that contains a execution unit).... but introducing random elements into a system designed to only execute things a certain way would lead to all sorts of instability (unless the heat is turned into a power source, at a nanostructure, or likely much larger level and not interfering with the 'computer' in dangerous ways).

In summary: Directing any form of energy through a circuit / process, in that there is a power source (fire, mechanical, electricity, etc) a meaningful input (controlled heat source creating energy -> in a certain way so that it is meaningful) a process, and a desired or meaningful output, then yes it could be classed as a computer. Some of the earliest computers where mechanical (not electronic), and just really rather large (half) automatic abacus. So yes, you could construct a rather large mechanical prototype and have it perform some calculations without 'storing' those calculations per say. Then the design could be scaled down to an electronic version, I doubt it would be efficient or powerful though.

Think about how a transistor works, and valves, etc, and before that, and so on, perhaps a mechanical transistor powered by fire or steam did once exist.

If you just had a giant maze / grid of transistors, etc, and 'damaged' some from heat, and some where powered by the heat (raises complexity but not impossible), then if done in a highly controlled way, statistically speaking one could create a kind of computer by destroying part of the 'grid' so that other parts 'route' in a particular way, to produce something with a power source, a input, a process and a output.

The point though ? - It would be totally pointless line of research as we have lithography concepts (similar to the above) already in use for manufacturing far more advanced (and far more useful / more powerful designs) integrated circuits / microprocessors. (Which have standards like registers and memory as it improves the usefulness and efficiency of the end product)..... so there is absolutely no point 'researching' this since we've surpassed the concept several generations already. (to the point where people have forgotten it, and think it is 'crazy')

You are a strange one, but who knows, 100 (or so) years ago you may have been onto something.
That's just my 2c

http://users.on.net/~darkpeace
 
You wanted constructive criticism here goes

Constructive
Before you melt any more computers try thinking your idea through a little bit. Aside from the fact that most of us will agree computers cant exist/wont do anything without memory. You need to ask yourself how overheating a computer to the point of death will prove that they can work without memory. A better test would be to remove as much memory as possible from your computer and see if it does anything constructive when u turn it on. I think u will find it doesn’t.
Also you need to separate your ideas better you have 2 points 1. Computers can exist without memory and 2. When you overheat a computer and funny things start to happen it’s the computer doing it rather than just random errors caused by it melting!

Criticism
I am laughing so hard right now that i am crying! At best you are struggling to get your idea across in a coherent fashion and at worst you are an idiot/crazy!
 
The most you will accomplish by using temperature change as an input is a robot that likes or dislikes warmth /light/sound/whatever

The reason the star just vanished is cause it was probably not a star, and if it did vanish, you are lucky to have seen it.
 
Hello, again!

I haven't read the original post (from a few months ago) but, as far as I can understand (and letting aside the [burn] process), the main issue here is to have a [very fast] processing device, working ramdomly (with no-matter what kind of input) AND without memory (as i've stated before, not to be confounded with plain storage...).
If it is the case, then it'ss possible to have a computing device with and/or without 'memory' (although with INHERENT storage!). Quantum Mechanics allows for such a device, as i tried to show in my previous post; however, storage must always exist (vector-processing is faster than general-purpose, scalar processing, because there's much less need for 'memory' to be used... but storage is always present!). For instance, one might consider electrons to be stored within wires (is this 'memory'?) and, by the way, "to be stored in memory" is a redundancy, since 'memory' is an abstraction for a process obeying definite rules, whereby 'storage' is already included.
Hence, i propose the following:
1. For the audience: stick with the basic idea (and be indulgent; after all, "we" have discussed for ages, the roundness of the earth, its location within the solar system and in the entire universe, for what matters!) and try to find consistent arguments to disprove it (i.e., put your minds at work!);
2. Back to basics: it's fundamental for the understanding of how computing works, to know its physical basis;
3. Finally, if there's someone with the competent expertise (like an electronic engineer...), please do forward your opinion. I'm sure you've already discussed less attractive items.

As for you, mr. "oolceeoo", i'd rather take some physics lessons; it will help you be consistent with what you're venting out, help you to structure your thoughts and, as [one] consequence, be able to express yourself better & to be able to COMMUNICATE, coherently. Who knows, maybe you'll become a future computing engineer or architect...


CORRECTION 1: In my previous post, i've mentioned "The Mach's Principle", by Ernst, Max; it's rather by Ernst Mach.

CORRECTION 2: Dear fellow, alternative geometries (or Non-Euclidean Geometries, as it's supposed to be called), have r-e-a-l-l-y nothing to do with atomic or otherwise nuclear devices. Historically, Carl Friedrich Gauss and, later, Nicolai Lobatchevskii, Jánus Bolyai and Bernhard Riemann (between the XVIII & XIX centuries), among others, laid the foundations for the actual non-euclidean geometries which, had a determinant role in Einstein's GENERAL Theory of Relativity, not his SPECIAL Theory of Relativity, from which the most famous equation came out.
 
Oh oh oh..... let me try one (Jumping up and down waving my hands in the air)...
If a man is alone in the forest and says something and there is no woman around to hear him, is he still wrong?


Schmide... you dropped your pants but still came up one short. 8O
 
Maybe you should read the replies again. There is a valid reason given as to why this can't be done using existing engineering techniques.

This isn't to say that there isn't some undiscovered/obscure tech which could allow this, but just that using current standard design practices it isn't possible.

Maybe this explanation makes will help: controllers exist in discrete states, not unlike the bubbles in a flow chart. To perform a useful function there are at least 2 states necessary likely more. For example when a machine is powered up it might initialize into a state that reads the inputs and tests for some condition. A transition might occur when the condition is met. During the transition the machine moves from one state to another. The new state might be to assert some output. The state a controller exists in at any given time determines what job it is doing at that time whether it is sample inputs, process data, or present outputs. In a controller some event must trigger when it is time to change to a new state (or a new bubble in the flowchart analogy) and the machine must be designed right in order to transition to the intended state.

As you can see it would be easy to design a controller to do something simple like: when the switch is flipped turn the monitor red. The same type of machine can be very complex even on the level of being a PC. These state machines are implemented using flip flops (memories building blocks).

This is how it is done. In order to do it differently some fundamental paradigm shift will need to occur. The memories are necessary no way around it.
 
I agree with the other posters who say you are not going to find out anything useful by heating up a bunch of old computers until they die.

If you want just a basic input->computing->output relationship, there is always the (increasingly small) world of analog computing! I agree that you cannot get much done digitally without some form of memory, but a nice continuous analog system will give you a continuous output to a continuous input with none of that pesky memory. I guess you could call inductance and capacitance memory of a sort, but if that's memory, forget electricity alltogether.

The problem, and the reason almost noone uses analog computers, is they are quite difficult to program (read build) and usually only do one thing.

Maybe you just like the smell of hot plastic and metal?
 
Right here we go

All modern computer use the "stored program concept" (AKA von nueman machines, probably spelt wrong)

as this defines
"programs are stored in memory, fetched decoded and executed"

if you remove the memory, then what, theres nothing to fetch, decode or execute

the CPU may have 1mb L2 cache (for eg), but somthing has to get into that cache

The way computers are designed, is for it to go into memory (hence DMA HDD's and other devices)

With DMA (direct memory access) components in the computer can input data into the ram directly, instead of using CPU clock cycles (PIO mode)

But in DMA or PIO mode, data is sent from devices in the computer to the memory (1 way or another)

All this is done by design and cannot be bypassed

Also removing the L1 cache from the equation is impossible, as the L1 cache is what the CPU uses to store bits of data as it performs cacluations on them, and is hard wired, without it, the CPU wouldnt function

However, your experiment is stupid
Beacause you are not operating the computer without memory, you are proposing an experiment where the memory has been used, and the computer is running, however your then removing the memory from the equation

In all, the computer would not work, u cant even POST without ram, u cant boot up without a HDD and u certainly cant do jack without a CPU
 
Everyone (those who deny memory) is forgetting that state is memory and memory is state. Whether it is analog, digital, quantum, or pants. The first state is off the second state is on. This is the first bit in memory. If a second state (on) turns on something else to (on) a new state is acquired and so forth. When it comes down to it, a computer is just a series of switches that act on other switches and so forth. Since the precursor to computing is state (memory), NO YOU CAN’T HAVE A COMPUTER WITHOUT MEMORY!!!!

No more jokes about the state in my pants.
 
Everyone (those who deny memory) is forgetting that state is memory and memory is state. Whether it is analog, digital, quantum, or pants. The first state is off the second state is on. This is the first bit in memory. If a second state (on) turns on something else to (on) a new state has is acquired and so forth. When it comes down to it, a computer is just a series of switches that act on other switches and so forth. Since the precursor to computing is state (memory), NO YOU CAN’T HAVE A COMPUTER WITHOUT MEMORY!!!!

No more jokes about the state in my pants.
Click to expand...
in a computer as its all hard wired

its like havign a train track
and removing a nice big seciton of it
 
Wait a second it is not hard wired. It’s mechanical and has moving parts. A physically interconnected switch uses varying properties of substrate. I think the Analytical Engine uses moving parts (rotating disks and cogs) as electricity (transistor) was barely understood at that time.
 
I used to be a physics major, and I've taken a course on computer architecture.

I try to look at computers from a simplistic point of view. I understand that the circuitry of the computer is inherently complex with shifters, multiplexors, decoders, flip flops, etc. I used to read books on computer architecture and although I could understand the individual parts one at a time, the bigger picture never really took shape.

I've had lots of ideas of how computing could be different aside from this one. I was sitting in my data structures class one day and I think my professor was going over doubly linked lists, and the thought just came to me. "This seems unnecessarily complex, there has to be an easier way."

I am doing this out of curiosity, to see artifacts take shape and how they look. How this relates to my theory of computers without memory I don't know. The harder I try and think about it the harder it is to find the link. Maybe there is no link. It could lead in another direction entirely.

The point I'm trying to make is I'm taking things one step at a time. It's sort of like stairs with missing steps. I'm at the bottom right now, but I can't get to the top until I fill those missing steps. I know that my explanation is vague, but right now it is vague to me also.

I'm getting another computer on Friday, and possibly several more in the coming weeks. I'm using Debian for these tests and gnome for my window manager. I'll let you guys know what happens when I get it.
 
You must log in or register to reply here.

TRENDING THREADS

Latest posts

Moderators online

Share this page

COMPANY
RESOURCES
FOLLOW
Top Bottom