The Handheld Of The Future

[Phil]

Hopefully, in the future we will be all talking about gp3200.

 

[divadsci]

so surely the speed would then be multiplied by the amount of pins outputting?? unless you're talking about the lag between the processor and the other parts but that would still be there for co processors so that would speed it up at all. then again if it was working on one calculation the lag would merely have to be added into the beginning then it would be outputting constantly.

or do i still not get it? :unsure:

 

[JDUK]

The multiple pins/transistors would improve power not speed.

for example 10 VW Golfs with a top speed of 120MPH doesn't mean VW Golfs have a top speed of 1200MPH but 10 VW's could carry 10 times more than a single car.

For example an equivelent AMD does more work than the same GHZ P4 as it does more per processor cycle.

 

[divadsci]

ok i think i understand now.

by the way the vw golfs could be made to go faster than 120mph. just connect all their axles to a large cog connected to a small cog on a wheel, et voila

 

[0-bake]

Isnt 15 Ghz close to the speed of light?

erm...speed has the unit [meter/second], 3*10^8 m/s would be lightspeed.
the "speed" of processors isnt really speed, but frequency, which means "per second" unit [1/s].

so, you cant compare a frequency with a speed.

 

[divadsci]

Isnt 15 Ghz close to the speed of light?

erm...speed has the unit [meter/second], 3*10^8 m/s would be lightspeed.
the "speed" of processors isnt really speed, but frequency, which means "per second" unit [1/s].

so, you cant compare a frequency with a speed.

so does that mean i was right in my thinking?

if so i hate my lack of stubborness <_<

 

[JDUK]

i wanst saying 15 GHZ as such was close to the speed of light but was more trying to simplify the point that an electron in a conventional CPU would have to move at that speed to achive 15GHZ.

 

[a1210]

maybe the cpu's cud be in some sort of weird medium and we may get processors processing stuff in the past and we get the results in the future :rolleyes:

well if we look at what technology is in development right now, theres..
hologram memory? <--- future harddrive?
paper thin screens? <---- future screen?

and someone has to think of a better design for handheld game machines...

 

[divadsci]

i wanst saying 15 GHZ as such was close to the speed of light but was more trying to simplify the point that an electron in a conventional CPU would have to move at that speed to achive 15GHZ.

so what if one electron can only move at that speed? if you have the pins in groups like say



(pin1) (pin2)


when one transistor opens it ouputs on (pin 1) and then instead of waiting for that transistor to close so it can open again use a different transistor on (pin 2) that would be doubling the speed of it in exactly the same way as having two seperate processors. which you said would be the only way to increase the speed past 15 ghz.

 

[JDUK]

maybe the cpu's cud be in some sort of weird medium and we may get processors processing stuff in the past and we get the results in the future :rolleyes:

a type of Quantum computer that works a little like that is in the works... to grossly over simplify the 4 page article from New Scientist they basicly take special kinds of particle (non-Abelian anyons) with a "memory" and weave them round each other into mathamatical "knots" as they travel trough space-time. The particles can then be examined and the knots they wove in the past can be observed in the state of the particles current charge.

Mega complex stuff, best grab a back issue or search their web site as i cant articulate it to well not being an Qauntum Physicis boffin my self and i've had a few beers.

 

[JDUK]

i wanst saying 15 GHZ as such was close to the speed of light but was more trying to simplify the point that an electron in a conventional CPU would have to move at that speed to achive 15GHZ.

so what if one electron can only move at that speed? if you have the pins in groups like say



(pin1) (pin2)


when one transistor opens it ouputs on (pin 1) and then instead of waiting for that transistor to close so it can open again use a different transistor on (pin 2) that would be doubling the speed of it in exactly the same way as having two seperate processors. which you said would be the only way to increase the speed past 15 ghz.

Not increase the speed. increase the processing power.

Imagine those 10 VW's again. each one only goes 120 but if you load them all up with important computations and make them travel from point A to point B then you could transmit 10 times more data but that data would never travel faster that 120mph.

More VW's is like more transistors or more processor cycles you dont get more speed you just get more done at the current speed.

If you and I work non stop for 24 hours that 48 "man-hours" of work. We havent magicaly crammed 48 hours into a sigle day we have just doubled the amount of work done in that time

 

[divadsci]

yes but if the pins are timed so that they output consequtively(sp?) and then go round the the beginnign again then that would be increasing the speed of the processor

 

[JDUK]

Again thats combined speed it doesnt happen any faster it just gets jobs done sooner. Imagine shooting me with a machine gun (im sure loads of you would like to) you hold the trigger down, each bullet is fired of in quick succesion(like your pins). The bullets all travel at the same speed but the job (of killing me) is done quicker than that done by a non automatic weapon.

You cant make any things happen faster than the speed of light (by current standards.... unless you slow the speed of light as done in underwater neutrino "telescopes") but you can effect the amount of work done at that speed.

a 15Ghz processor doing things the current way may be considered not that fast but as technology gets better they are finding better ways to get more work done per cycle so amount of work done is not always relative to Ghz and will porbably be much better by the time you standing in line for you AMD-Omega15

 

[a1210]

oh i was just looking at some article about wearable technology and maybe instead of handheld gaming, it could be wearable gaming....
like glasses with a screen (eyetop glasses)
speakers in the hood
bluetooth
uses body heat to run
and maybe even solar power?

 

[GiG]

Hey hey! I started the topic to talk about a future handheld, not a piece of hardware so fast that can be used as a time-machine!

Well, the glasses idea is good, but what happens when you want to look around, I mean... outside the game? The glasses should detect the focus of the eyes, and dinamically change the opacity of the lens. The controller could be a small joystick with a few buttons on it: pressing a button while looking at an object in the game would start an action.

The idea of the body like a battery... is a bit frightening
...where should I plug my handheld?!

 

[a1210]

hmm it'd suck to see handhelds in the future having 32mhz CPU that only runs PS1 games when overclocked to 200mhz <_<

 

[a1210]

i think with the eyetop glasses you can still look around.. it just has a screen on the right...
well if we were the battery.. it might be kinda like exercising while playing games...

maybe theyll start having handheld/wearable gaming lessons in PE class...

 

[GiG]

It would be cool if the game was displayed on the lens, substituting the real vision, don't you think? The mini-screen on the right is terrible imo: after 1 hour of gaming my eyes would be blocked to the right :lol:

 

[Charge]

Terrahertz processors? quantum processors? Gigs of RAM?

I'd be happy with a floating point unit myself B)

 

[rcx21000]

Quantum processors can do brute force caluclations fast... If you want to do something like find out what x is where x+2=4 with a quantum processor, you'd do it like this: You shoot 8 particles into the processor (if you want all 8bit nums calculated.) One side up = 0, the other side = 1. The processor does the calculation for the combination that you shot through. According to the quantum theory, if a state hasn't been measured, the state is in all possible states, so doing the calculation on the 8 unmeasured particles will automaticly do every calculation....... It's like infinity processors all calculating beside eachother on one of the possibilities

At least I think thats how they work

You could do raycasting (or tracing, or whatever) and calculate each ray all at the same time

You could also make something that calculates every possible combanation fo chess.... Thats what the P/TB of mem is needed