Press and Reviews Thread

[Yak]

We had "a fuck ton of people" complaining when we wanted to go 3.0, so to satisfy both we implemented 2.0 AND 3.0

You'd have been the ones complaining if you'd done it 3.0 only and most of your customer base found they couldn't play.

Well duh. Which is why the baseline is sm2.

[PVB]

.........and modern Shader-Model 4 effects (if your graphics card supports DX10 effects, that is.......

Oops.

We DO NOT use Shader Model 4, we use Shader Model 3 AND 2.

As clearly specified in the Online README.TXT

We had "a fuck ton of people" complaining when we wanted to go 3.0, so to satisfy both we implemented 2.0 AND 3.0

4.0 was considered "too much high specs" and for the moment left apart, besides not really necessary so why to use
something "bigger" when we do not need it

My point was that they'd made a mistake, not that you should have made it for Shader 4.

[bcass]

In a nutshell, what's the difference between SM 2.0 and 3.0?

[gravy]

1.0

[bcass]

[stefcha]

In a nutshell, what's the difference between SM 2.0 and 3.0?

Silk and treacle. On my PC, anyway

[Yak]

If I had to sum it up in one word I'd say "loops". It's awesome enough being able to run significant coide per pixel, but being able to run significant iterations of significant code per pixel is even more awesome - and demanding of the GPU as well.

[gilesgoat]

In a nutshell, what's the difference between SM 2.0 and 3.0?

"Unlimited Instructions" VS "96 slots (? can't remember the number ) and such"
"More conditional jumps/things/allowed"
"Geometry generation" ( or better "geometry expansion" ) i.e. a shader can generate vertexes ( but we don't use that ).
"More shite available"
"More constants/vars/registers available"
( not sure, possibly ) "compiler and optimization bit better".


In short you can do LOT more complex code on 3.0, on 2.0 you are quite TIGHT on the number of instruction slots
you can execute plus I think there are also some limitations on certain things I can't remember how they are called.

Moral of the story : "you really can't do some stuff properly in 2.0 because there's not enough power/resources".

You can trim things down and/or make them "similar but not the same".

But 2.0 can still be faster to execute than 3.0, esepecially on older cards ( of course ).

[Dio]

Yeah, 2.0 is typically quite a bit easier to optimise - not least because it's a single basic block, so you don't have to worry about the multiple blocks that flow control introduces. (In fact anyone running a 9700 Pro or thenabouts who hasn't been wise enough to update their drivers since 2004 might still be using my optimiser ).

[Harlequin]

Eurogamer review: http://www.eurogamer.net/articles/gridr ... on-article

I find this interesting, as I have read many people commenting that (presumably on the basis of descriptions and screenshots) GRR reminds them of Geometry Wars:

[Geometry Wars] took many of its cues from Gridrunner++, Revolution's forerunner. Grid? Check. Crabby ship? Check. Retina-searing particulate explosions? You get the picture.

[gilesgoat]

YAY ! 10 on Metacritic so far

http://www.metacritic.com/games/platfor ... revolution

[IronGiant]

A mini review from someone:

http://funeffect.com/wp/2009/09/gridrunner-revolution/

[PVB]

YAY ! 10 on Metacritic so far

http://www.metacritic.com/games/platfor ... revolution

That's the public rating, not 'official' reviews.

[PVB]

A mini review from someone:

http://funeffect.com/wp/2009/09/gridrunner-revolution/

I'm sure I've read the same text on another site.

[oddbob0]

It's the TIGS article.