predictions for 2006

[opcode-foo]

My predictions for the coming year:

The internet will collapse. People will finally realize what a worthless medium it really is,
just a massive archive for trolls and spammers.

The internet has become the first prototype for a corporate police state. Moderation
and censorship have become the norm of life on the internet.

The dot com industry will boom on the next new thing.

AI will be applied to Google's archives and discover there is no intelligent life on earth.

ET will make a come back on the XBOX360

John Carmack will invent the next big thing
:sneaky:

[Reedbeta]

opcode-foo, perhaps you could post some actual content once in awhile instead of just random crap like this...?

[Jynks]

I remember reading once about a year ago. You know how gfx cards, or even sound cards now, come with there own chipsets to handle the gfx or sound independently from the computer cpu. Freeing up cpu time for other things and I am pretty sure they have some kind of magic "I can not understand it" type way of being built to do their tasks really well with ways to directly access the chips with special functions.

Anyway I heard that there will be AI cards. For the same reasons. A card you can get for your computer that will allow developers to make a better Ai for a strat game, say chess or dune clone.... .. . or finaly make a proper "grammer" check in a word proccessor.... or allow you to use more natural english in help files by asking questions of it... etc etc

I've not had any confirmation of this.... ..

[Onikhaosifix]

I believe AI cards are being processed as we speak but I could be wrong. But I am certain that physics processor chips already exist. I'm not 100% but I think either the PS3 or the X-Box 360 has that.

[Nick]

Physics cards, AI cards, it's all a bag of air...

Why? Well first of all read the Ageia (http://www.ageia.com/products/physx.html) website in detail. It's all marketing talk without content. Not a tiny detail on actual physics hardware. And all presented demos run on the CPU.

Secondly, dedicated hardware won't really help. The bandwith to and from the card is too small to make a difference. And don't underestimate the processing power of a CPU. Furthermore, in 2006 dual-core processors will become mainstream. What will we do with that double performance? Exactly, we can use it for things like physics, without demanding any other extra hardware.

Last but now least, for more interactive scenes with better physics we don't only need phsyics calculations. Every dynamically moving/deforming object requires recalculation of visibility. All major engines of the last year used some form of PVS, which will no longer work correctly when static geometry becomes dynamic geometry.

If dedicated processing power for physics (and related visibility) is really required, I expect the GPU to be a better candidate. The newest generation already features good GPGPU qualities. DirectX 10 even adds a 'geometry shader' to the pipeline, which could be excellent for physics...

[ProgramWizard]

opcode-foo, perhaps you could post some actual content once in awhile instead of just random crap like this...?
I believe this was made when he thought that "devamster was full of a bunch of twits, just like gamedev.net". He said it, not me. Anyway, for 2006, I predict that Uncyclopedia will be sued. Over something. Don't ask why, just got a gut feeling. One day, they'll piss someone off...

[.oisyn]

If dedicated processing power for physics (and related visibility) is really required, I expect the GPU to be a better candidate. The newest generation already features good GPGPU qualities. DirectX 10 even adds a 'geometry shader' to the pipeline, which could be excellent for physics...

If the goal is parallel processing, using the GPU to do physics is not that good a choice since you're taking away raw rendering power and creating a backdependency for physics on rendering.

[Mihail121]

I believe that fishes will learn to talk human languages!!!

[Nick]

If the goal is parallel processing, using the GPU to do physics is not that good a choice since you're taking away raw rendering power and creating a backdependency for physics on rendering.
Physics doesn't really take that much processing power. Besides, clearly DirectX 10 already prepares for it. Just regard it as having a physics processor inside the GPU. One card instead of two. And it's also quite likely that multi-core CPUs will help out. In any case a separate physics card is doomed to become the 'fifth wheel'. A waste of money that won't deliver on expectations. The only real role Ageia could play is to develop a physics API that is capable of both GPU assistance and (multi-core) CPU processing.

[.oisyn]

I agree with you with the fact that a seperate physics processing unit is a dead end, we'd be much better of with something like a cell processor. I just don't agree that the GPU should be used for it, for the two main reasons I mentioned: firstly, you don't want to take away raw graphics processing power to do some physics, and secondly it's so very bad for parallellism. You are creating a dependency on the GPU where you don't want one; you'll need the results before rendering, and with current hardware it can take a long time before you get the results (rendering one or two frames ahead is not uncommon).

[Nick]

I agree with you with the fact that a seperate physics processing unit is a dead end, we'd be much better of with something like a cell processor. I just don't agree that the GPU should be used for it, for the two main reasons I mentioned: firstly, you don't want to take away raw graphics processing power to do some physics, and secondly it's so very bad for parallellism. You are creating a dependency on the GPU where you don't want one; you'll need the results before rendering, and with current hardware it can take a long time before you get the results (rendering one or two frames ahead is not uncommon).
Physics calculations are not that costly that it would significantly lower rendering performance. I mean, nowadays it's still all done on the CPU. Even if you have a whole CPU core for physics alone that's still only a fraction of the GPU's processing power. So I don't think that's anything to worry about. Besides, if physics are really going to revolutionize gameplay then I think it's worth it.

For good parallelism we need a short response time. This seems doable because the vertex and/or geometry shaders are at the start of the pipeline. And a GPU with a unified architecture could just reserve a few execution units for physics calculations, at any time, and instantly stream the results back to system memory. This would be no worse than a separate physics card.

Ideally there would be no dependency at all, with all data staying in graphics memory. Queries could be made asynchronously, like intersection tests. The result would be returned a couple frames later...

[gamerdude3045]

The internet is much more than "a massive archive for trolls and spammers." It is a place to shop, find information and contact people. It creates thousands of jobs for people all around the world. The government uses it to protect us.

[diejondie]

i predict.....

michael jackson will be sued in europe :lol:

mmofps will be the next biggest thing

wow will die for something better to take its place

hd tvs will become inexpensive and everywhere

nintendo will go bankrupt

bush wont enforce illegal immigration policys

televised online competitive gaming

mm... i guess ill see if any of this happens :w00t:

[Reedbeta]

The government uses it to protect us.

How does that work? :lol:

[.oisyn]

For good parallelism we need a short response time. This seems doable because the vertex and/or geometry shaders are at the start of the pipeline. And a GPU with a unified architecture could just reserve a few execution units for physics calculations, at any time, and instantly stream the results back to system memory. This would be no worse than a separate physics card.

Which is then more like a cell processor than an actual GPU, like I said ;).

[gamerdude3045]

How does that work? :lol:

I used the wrong word there. I didn't mean protect. I meant provide us with information on our safety.

[Mark]

In 2006 we will see the first successful moves to unify TV and the Home PC, through large HDTV displays. This will be the one thing to really get the ball rolling, for better or for worse.

Mozilla Firefox will become the browser of choice on 25% of all web capable machines.

Microsoft will ship Windows Vista which will do exactly the same things Win2K does, but eat 10 times more ram, disk space, network bandwidth and GPU time. There will be no way to uninstall or remove 'features' you don't want or need and IE (anyone remember that?) will be pre-installed to a ROM chip on your motherboard, so you can never get rid of it. Ever. Windows Media Player will use special fragmenting features to scatter it's files completely randomly about the machine, even worse than before. A system service for automating coffee machines will be enabled by default. It'll eat 156MB of ram, but make a fantastic cappuchino.

Office Vista will be released to accompany the new OS. It will ship on 5 Blue ray discs and still be massively inferior to Open Office. Users will confuse the price tag with the customer support number. A pig farmer in Ohio will be innundated with technical queries.

Nintendo Revolution will be scrapped. Nobody will care.

UK petrol prices will reach a kidney per litre. Nobody will complain.

[gamerdude3045]

Users will confuse the price tag with the customer support number.

hahahahaha

P.S. I'm just stickin with XP anyways...

[Reedbeta]

Speaking of Windows...has anyone here installed XP x64? I am thinking about putting it on my laptop (which has an Athlon 64) and am not sure if that's a good or a bad idea =D

[Vandervecken]

Speaking of Windows...has anyone here installed XP x64? I am thinking about putting it on my laptop (which has an Athlon 64) and am not sure if that's a good or a bad idea =D

A year or two ago I installed the open beta they released. It was absolutly horrible... my problem was that there were almost no 64bit drivers for all my hardware even the stuff nvidia released I couldnt get to work properly. I'm sure it's probally a lot better now but I'd search for all the drivers you need before downloading :) And yeah... none of the windows prebuilt drivers worked with my stuff either ><

[.oisyn]

Speaking of Windows...has anyone here installed XP x64? I am thinking about putting it on my laptop (which has an Athlon 64) and am not sure if that's a good or a bad idea =D
I'm actually in the exact same situation :lol:. My greatest concern is driver support, but I haven't checked the Acer site though (I have an Aspire 1501.. eh.. LMi.. something)