Deano Calver's Article online

[nelg]

http://www.pcpro.co.uk/custompc/features/71102/3d-perspective.html

 

[Geo]

http://www.pcpro.co.uk/custompc/features/71102/3d-perspective.html

Uh, is this the one where the editor came on here and was fishing for an author? Or am I confuddled again?

 

[Jawed]

I like this bit:

As you can see, neither technique is ideal, and this is why graphics hardware now offers various features specifically for accelerating both types of shadow rendering. Nvidia chips actually have more in the way of specialist shadow acceleration features than ATi chips, and it is often joked among game developers that Nvidia devotes so much chip space by accelerating Doom 3's shadows that every other game suffers as a consequence.

Jawed

 

[nelg]

http://www.pcpro.co.uk/custompc/features/71102/3d-perspective.html

Uh, is this the one where the editor came on here and was fishing for an author? Or am I confuddled again?

Yep.

 

[AlNom]

Thanks for the heads up

 

[Reverend]

Easy money, Deano.

 

[nelg]

I wonder if Deano was refering to the R520 here.

Shader Model 4.0

The first major change to arrive with Shader Model 4.0 will be the unification of shaders, meaning that any differences between vertex and pixel shaders will completely disappear. Indeed, rumours even suggest that ATi's new chips will only support one type of shader unit, and the card will instead allocate shader units to specific jobs (vertex or pixel), depending on the current workload.

 

[Reverend]

I wonder if Deano was refering to the R520 here.

Shader Model 4.0

The first major change to arrive with Shader Model 4.0 will be the unification of shaders, meaning that any differences between vertex and pixel shaders will completely disappear. Indeed, rumours even suggest that ATi's new chips will only support one type of shader unit, and the card will instead allocate shader units to specific jobs (vertex or pixel), depending on the current workload.

I admit I didn't read every page -- I X'ed the window when I was asked to register after reading 3 or 4 pages (and these pages I read were the last few pages, which I was disappointed with, being far too general and consisting of things I -- and perhaps those here -- already know... of course, this could be because that's exactly what that site asked of Dean, to make it simple and a very basic summary of the subject... Dean must have practiced really great restraint to not expand) but :

meaning that any differences between vertex and pixel shaders will completely disappear

I'm sorry but "differences" between the two will exist even with SM4. Dean didn't expand about what he meant by "differences" (it's such too huge to explain) so I'll just STFU.

 

[AlNom]

I wonder if Deano was refering to the R520 here.

rumours even suggest that ATi's new chips will only support one type of shader unit

Even so, R500 still counts as a new chip, and I think he only just hinted at this alone with this statement. Otherwise, he mentions that ATi will have SM3 chips (R520, of course).

 

[nelg]

Even so, R500 still counts as a new chip, and I think he only just hinted at this alone with this statement. Otherwise, he mentions that ATi will have SM3 chips (R520, of course).

Considering the PC centric nature of the magazine and article I thought the mention was interesting.

this could be because that's exactly what that site asked of Dean, to make it simple and a very basic summary of the subject... Dean must have practiced really great restraint to not expand)

I think thats a given and that he did a good job of it.

 

[hovz]

id like to hear deanos personal opinion on the x800 series v the nv40 series. curious what a devloper thinks about them without worrying about controversy.

 

[vb]

I admit I didn't read every page -- I X'ed the window when I was asked to register after reading 3 or 4 pages

www.bugmenot.com

 

[DeanoC]

[Dean must have practiced really great restraint to not expand) but :

The editors simplified a fair bit from my original version, to be honest its not entirely accurate in a few places due to being so simple. I found it much harder to write assuming no 3D knowledge, its much easier to write for the Beyond3D crowd.

meaning that any differences between vertex and pixel shaders will completely disappear

I'm sorry but "differences" between the two will exist even with SM4. Dean didn't expand about what he meant by "differences" (it's such too huge to explain) so I'll just STFU.

Obviously there are some difference mainly to do with data routing and differentials but as the API apperance will be unified and at least one set of hardware will use the same execution units for shader work I felt it was a fair way of explaining SM4 in simple terms...

 

[_xxx_]

I admit I didn't read every page -- I X'ed the window when I was asked to register after reading 3 or 4 pages

www.bugmenot.com

Cool!

 

[DeanoC]

Its worth noting that the print article suggests anybody who wants more details on graphics check out Beyond3D as the place on the web for 3D card info.

As Rev pointed out it isn't meant to the people who already read Beyond3D, its for people who don't know what a normal is or even a GPU.

I expect most of the regulars here will find it really basic and annoying and have a few inaccuracys but its meant to be really simple (and fit it the required space) so I'm fairly happy with it.

 

[Unknown Soldier]

I admit I didn't read every page -- I X'ed the window when I was asked to register after reading 3 or 4 pages

www.bugmenot.com

Cool!

Way Cool!



US

 

[DeanoC]

id like to hear deanos personal opinion on the x800 series v the nv40 series. curious what a devloper thinks about them without worrying about controversy.

I like my GF6800, SM3.0 (I've written my first shader with dynamic pixel branching the other day...) and FP16 blending make me a happy developer...

But I'm a console bunny at the moment so I don't have to consider the whole PC market issues ;-)

 

[Unknown Soldier]

ATi technically won this round, a victory that even Nvidia will admit to now.

Have Nvidia ever admitted they lost the round? I don't think so .. sorry DeanoC. That's just how I see it of course.

Using 3Dc basically reduces them to a bare minimum, so that you can compress normal maps to a usable size. A few games support it so far, including Sid Meier's Pirates!, Serious Sam 2, Far Cry, Tribes Vengeance, darkSector, Project Freedom, Black and White 2 and Freedom Force 2, and it is expected to become the standard normal compression technique in the future.

Serious Sam 2 is out??

Nice Article though .. I like.

US

 

[vb]

there is also a firefox extension

 

[digitalwanderer]

there is also a firefox extension

For bugmenot? Link please!