The Rydermark Thread (TM)

[Sharkfood]

Eh. There are plenty of examples of both
Then again, this is the first time I've heard somebody say B3D was biased in favor of NVIDIA

Well, plenty of examples of pages of deflection off the context anyways.

Oh, heavens no. I've always enjoyed Dave's objectivity and the absolute extremes the B3D staff goes through to find objective answers without any bias one way or another.

It's more a crack at forum communities.. and often times leads one back to that first link you've conveniently provided above.

I also think Rys's research is *exactly* what I would have expected such accusations to result in... as they always have in the past. A mocking reply or indignant link would, of course, be the unacceptable difference of deflection versus some curiosity/objective finding research to see what may be afoot. Thanks for that Rys.

I'd have done it myself but alas, I just moved into a new home.. the NVIDIA system isn't hooked up yet. Work PC and HTPC hooked-up are ATI-based, and the LAN/Gaming room still has paint tarps and naked ganglia of CAT5's sticking out of holes in fresh painted walls... awaiting the three game PC's and furniture sitting in my backyard. Here's to hoping nobody hops the fence and makes out with a ton of killer gear.

 

[IgnorancePersonified]

Bingo!

I can be relied on too provide the pertinent data in any technical thread

 

[Tridam]

Everytime we work on a new GPU review, I check shading precision etc. with handwritten ASM shaders I coded for this purpose. I do too when a driver release brings significant performance improvements. Since the FX era I never found any real issue (just a couple of bugs that were not precision related).

I heard once a developper using half in his code and complaining because Nvidia's precision was lover than ATI's precision (ATI ignores partial precision). But it seems impossible that serious benchmark developers do such a error. As I said to Fudo, it's more likely that he did not understand correctly what they told him.

 

[chavvdarrr]

http://theinquirer.net/default.aspx?article=33097
http://www.theinquirer.net/images/articles/nshader.jpg

http://www.theinquirer.net/images/articles/atishader.jpg

 

[Jawed]

Looks to me like someone's yanking L'Inq's chain. Someone said "Let's just make this selection here and change this..."

I've taken the difference between the two and magnified it to show the selection edges.

Jawed

 

[zed]

hmm looks like someones applyed a sharpness filter to one of the screenshots (+ very sloppy work as well) check the ground under the tourist in the center of the screenshot looking this way

We don’t yak a lot about it, we will just show you two pictures and let you judge for yourself.

 

[Jawed]

Actually the so-called picture is some kind of cut-n-paste crap.

This whole thing is a complete sham.

Jawed

 

[atlantis]

yeah i've seen the difference too... what a shitty hand made selection...

time for theinq to go to adblock too, after theregister


what a crap reading

 

[KimB]

There's no possible way that the rest of the image is exactly the same. The different texture filtering algorithms alone should result in some very noticeable changes between the two in the difference shot.

 

[chavvdarrr]

hmm looks like someones applyed a sharpness filter to one of the screenshots (+ very sloppy work as well) check the ground under the tourist in the center of the screenshot looking this way

lol, I thought that someone applied blurring to "nv" image ...

 

[_xxx_]

Lets assume there were real, where are the supposed artifacts? All I see is a bit blurrier and a bit darker water in the nV pic.

 

[Jawed]

Lets assume there were real, where are the supposed artifacts? All I see is a bit blurrier and a bit darker water in the nV pic.

It's a total sham.

I wouldn't be surprised if this is a campaign to show how easily L'Inq can be mislead.

The pictures are composites: examine the edges of the dome, look at the photographic people, notice the entirely inconsistent aliasing, etc.

Jawed

 

[_xxx_]

I know, but regardless of that, how could this be related to _pp? I see neither bending nor anything like that, just bluriness which has nothing to do with _pp anyway. I just don't get it how this should prove anything _pp related, even if real.

 

[Jawed]

This isn't 3D computer graphics being generated in realtime!!!!!

Jawed

 

[_xxx_]

Ok, forget this article/pics for a second.

Let me rephrase: if there was _pp cheating, wouldn't we see a totally different kind of artefacts/inconsistencies there? I'm simply not sure what the "locked" _pp would look like. Bending? Else?

 

[Jawed]

_PP errors are only going to occur over the duration of operations within fairly tortuous shaders - even then I think the errors are prolly going to be minor.

Jawed

 

[_xxx_]

Ok, thanks

 

[Razor1]

if the picture was sent through the smart sharpen filter in photoshop with around 1-2 radius to remove gaussian blur then dropped contrast and increased brightness that ATi pic would look exactly like the nV pic.

 

[trinibwoy]

H I L A R I O U S

 

[INKster]

Look's like someone pulled a prank on them...