[Flipside]

It's actually very easy to do different heigh lakes etc if you are using a heightmap landscape, since it uses exactly the same equation as a flood-fill. you can fill up 'basins' with water quite easily, problem is, it won't handle volume, so there's an infinite amount of water channels running downhill etc would be impossible.

[aldo_14]

It's actually very easy to do different heigh lakes etc if you are using a heightmap landscape, since it uses exactly the same equation as a flood-fill. you can fill up 'basins' with water quite easily, problem is, it won't handle volume, so there's an infinite amount of water channels running downhill etc would be impossible.

Ah, but the pre-fluid dynamic (etc) approach would probably involve placing certain parts - directly or indirectly - that act akin to keyframes.  Like, having a hollow and x and y means water will go between x and y, etc.   not so much physics - in fact, not atall - but just drawing a path in a very abstract manner.

Now, how much having actual fluid dynamics would improve upon (a good implementation of) the above, vis-a-vis the hardware etc requirements, is quite an important question IMO.

[Flipside]

Yup Imagine a city builder where water has to be channeled to a reservoir etc

For the main part Fluid Dynamics is more an 'Eye Candy' feature though, apart from a few interesting 'Puzzle' type games, there are limits to what can be achieved with it. I see the biggest feature of the card the ability to handle deforming and gravity-based physics, as well as the possible boosts to such things as Particle emitters etc.  With good Geomodding, the idea of flying through an Asteroid field with the asteroids colliding and shatterig realistically is not all that far away.

Also, games like a 3D version of The Incredible Machine could be great fun

[Wild Fragaria]

This is cool.  I feel like I have made some contribution to the chip although I didn't work for IBM  I researched (about 7 years ago) alternate metals other than copper for better and cleaner conductivity and actually synthesized the first ever germenium diketone compound (for MOCVD) that could be used in nanotechnology applications.

[aldo_14]

This is cool.  I feel like I have made some contribution to the chip although I didn't work for IBM   I researched (about 7 years ago) alternate metals other the copper for better and cleaner conductivity and actually synthesized the first ever germenium diketone compound (for MOCVD) that could be used in nanotechnology applications.

Wow.

Cool

[Mefustae]

Indeed, giggidy.

[Ace Pace]

Just remember this is a rather simple chip made for demonstration purposes and can in no way compete with the normal desktop chips we use.

Another thing to remember, you can easily reach high clockspeeds with x86 chips, look at the Presscott design for the idea, very long pipelines enable you to ramp up clockspeeds(Tulsa will be the perfection of that idea). However, as we saw from Presscott, it dosn't pay off.

[aldo_14]

Just remember this is a rather simple chip made for demonstration purposes and can in no way compete with the normal desktop chips we use.

Another thing to remember, you can easily reach high clockspeeds with x86 chips, look at the Presscott design for the idea, very long pipelines enable you to ramp up clockspeeds(Tulsa will be the perfection of that idea). However, as we saw from Presscott, it dosn't pay off.

My expectation would be that any sort of research like this, wouldn't be using the old trick of non-optimal pipelines to fake speed ratings.  To do so would defeat the point of the excercise.

[Colonol Dekker]

How old are you? 7 years ago i was 17......

Wow damn i feel old.

[Ace Pace]

Just remember this is a rather simple chip made for demonstration purposes and can in no way compete with the normal desktop chips we use.

Another thing to remember, you can easily reach high clockspeeds with x86 chips, look at the Presscott design for the idea, very long pipelines enable you to ramp up clockspeeds(Tulsa will be the perfection of that idea). However, as we saw from Presscott, it dosn't pay off.

My expectation would be that any sort of research like this, wouldn't be using the old trick of non-optimal pipelines to fake speed ratings.  To do so would defeat the point of the excercise.

Chips like these are made for expirimental purposes rather then anything resembling a plan to bring this to market. This seems more like a 'lets see what we can do with new technology' with an eye towards applying parts of it to designs.

IBM does of this kind of work, concept design research.

[Flipside]

Yup, we won't be looking at 100Ghz CPU's for a long time yet, 10Ghz may well be closer than we think, though I suspect it's more the architecture of the Motherboard that stands in the way of that rather than technology.

And that's pretty cool stuff Wild Fragaria! You made a little piece of History

[Colonol Dekker]

Well a tenth of a Teraflop in my house is ok by my estimations.
I give t five years, maybe 7.

[Wild Fragaria]

How old are you? 7 years ago i was 17......

Wow damn i feel old.

I'm not that old.  I'm a child prodigy   How old do you think I am?

Yup, we won't be looking at 100Ghz CPU's for a long time yet, 10Ghz may well be closer than we think, though I suspect it's more the architecture of the Motherboard that stands in the way of that rather than technology.

And that's pretty cool stuff Wild Fragaria! You made a little piece of History

Thanks.  I was quite happy to read the news this morning

[Colonol Dekker]

LAte teens early 20s?........ I've never conversed that much with you, well not enough for me to bas an opinion. You are an Londoner or S/E Englander though, i can tell that much

[CP5670]

I think this is a simple signal processor, quite far off from a computer CPU as Ace Pace said. There is some interesting manufacturing tech involved though.

OK, that's enough exaggeration. FEAR runs beautifully at max settings on my rig, and it wasn't exactly top of the line when I built it a year ago. No overclocking or SLI, either.

Depends on what you call beautifully. I have upgraded my stuff twice since that game came out and although it now runs decently, the performance is nothing even close to what I would consider "beautiful," unless I play at some very low resolution.

[Wobble73]

LAte teens early 20s?........ I've never conversed that much with you, well not enough for me to bas an opinion. You are an Londoner or S/E Englander though, i can tell that much

Didn't you see her pics when she posted em on here Dekker OMG 

[Colonol Dekker]

Not that i remember, I tend not to look at peoples pics it shatters the illusion of the preconcieved handsome beutiful people i have. 

J/K

I'm sure you're all presentable.

[aldo_14]

LAte teens early 20s?........ I've never conversed that much with you, well not enough for me to bas an opinion. You are an Londoner or S/E Englander though, i can tell that much

[Ace Pace]

Yup, we won't be looking at 100Ghz CPU's for a long time yet, 10Ghz may well be closer than we think, though I suspect it's more the architecture of the Motherboard that stands in the way of that rather than technology.

Not really, making a faster bus on the motherboard is more technical then anything, the problem is the CPU itself and the memory it has to acess.

On one side, making CPUs work fast is connected to having more stages in the CPU pipeline(I.E Netburst design), but then you run into heat/power issues, as more pipelines requires more power. 

The other issue is RAM speed, the faster your CPU is, the more it has to wait for memory. Today latency for on die memory is down to single/lower tens of NSs, still a long time for the CPU, while RAM acess latency is slowly dropping from 150ns to 50ns(Athlons do well here). Thats still a very long time for the CPU to wait, so you need more cache to keep memory acess down. Problem spirals untill people will finnaly adopt XDR RAM.

[Wild Fragaria]

LAte teens early 20s?........ I've never conversed that much with you, well not enough for me to bas an opinion. You are an Londoner or S/E Englander though, i can tell that much

Me, an English from London (or from southeast England)??!  That is a very big news me

I know you're 24 (because you said it on HLP) and you have a 2 yr old kid (also, you said it on HLP).  I am around your age.