Anand has Farcry 1.2 patch and tests SM3.0

[Ante P]

some quick and dirty swedish SM3.0 lovin'

http://www.nordichardware.se/Artiklar/?skrivelse=252

Nice benchs. BTW, can you explain the "Dual Depth Shadow Mapping with 3x3 Bilinear Percentage Closer Filter" +333 % boost? (in english or even french )

I have no idea.

TB tried supplying a new build of the app but it refuses to work on my rig and the debug versions doesn't create an error log so we're kind of at a stand still now that we don't know how to solve.

 

[Hyp-X]

In case you missed that:

UPDATE: It has recently come to our attention that our 4xAA/8xAF benchmark numbers for NVIDIA 6800 series cards were incorrect when this article was first published. The control panel was used to set the antialiasing level, which doesn't work with FarCry unless set specifically in the FarCry profile (which was not done here). We appologize for the error, and have updated our graphs and analysis accordingly.

 

[nAo]

Nice benchs. BTW, can you explain the "Dual Depth Shadow Mapping with 3x3 Bilinear Percentage Closer Filter" +333 % boost? (in english or even french )

I have the feeling they used dynamic branching to avoid to sample 2 different resolutions depth maps in some cases. Samples are taken and percentage close filtered from a single depth map..and from another one. Sampled signal is then filtered between the 2 maps..like in trilinear filtering.
When the filtering coefficient is almost zero or almost one you just avoid to sample both and you just sample the higher or the lower resolution depth map, thus saving bandwith and ALU cycles. This would easily explain the big improvements in performance we're seeing on that test..
BriLinear Depth Shadow Mapping anyone?
Obviously this is just an hypothesis..sorry if I completely missed the target.

ciao,
Marco

 

[Pete]

Nice to see Derek respond so quickly and update the article so transparently. It's a total turnaround from last year's mess.

Still, I have to (stubbornly) admit that it's disappointing that a profitable site can still make such mistakes.

 

[Ailuros]

Nice to see Derek respond so quickly and update the article so transparently. It's a total turnaround from last year's mess.

Still, I have to (stubbornly) admit that it's disappointing that a profitable site can still make such mistakes.

Can't disagree there. Could be that time pressure is one of the possible explanations, yet still not a good enough excuse; not at least for a slip like that.

It takes fractions of seconds in a running timedemo to detect with the naked eye whether antialiasing is on or not. Presupposition would be of course that you actually watch the screen while it's running....*cough*

 

[dizietsma]

But don't you think that the reviewer should visibly NOTICE the lack of AA?!?!

Jeeze, I sure would. Mebbe this goes a long way to showing why Anand never saw a problem with the nV3x's image quality...

But remember your thoughts on temporal AA until someone pointed out you had not even enabled it yet ...

 

[Jack_Tripper]

Nice to see Derek respond so quickly and update the article so transparently. It's a total turnaround from last year's mess.

Still, I have to (stubbornly) admit that it's disappointing that a profitable site can still make such mistakes.

Can't disagree there. Could be that time pressure is one of the possible explanations, yet still not a good enough excuse; not at least for a slip like that.

It takes fractions of seconds in a running timedemo to detect with the naked eye whether antialiasing is on or not. Presupposition would be of course that you actually watch the screen while it's running....*cough*

Now you ARE doing that with you're 6800 review...right????



Jack

 

[Demigod]

Could someone explain to me why the ati shot here
http://www.anandtech.com/video/show...reviews/video/nvidia/farcrysm3/volcanoati.jpg

has less objects behind the phone than either of these two nv shots?

http://www.anandtech.com/video/show.../reviews/video/nvidia/farcrysm3/volcanonv.jpg
http://www.anandtech.com/video/show...reviews/video/nvidia/farcrysm3/volcanosm3.jpg

edit: there is an ammo box missing in the nv sm2 shot (on the shelf to the right)
Thanks

 

[dizietsma]

Anandtech have fixed the results but not the conclusion, first part of their conclusion reads

"Both of our custom benchmarks show ATI cards leading without anisotropic filtering and antialiasing enabled, with NVIDIA taking over when the options are enabled."

Which is completely out of phase with the new results.

Sigh.

My apolgies to jvd and rainz for saying now you get speed with AA/AF with SM3, nope Ati is still king when it comes to AA and AF in this game

 

[Ailuros]

Now you ARE doing that with you're 6800 review...right????



Jack

O'm tupongg bkindfodled

 

[Draconis]

Well now after reading this stuff about anandtech with the new results I wont go to that site anymore.

 

[digitalwanderer]

Anandtech have fixed the results but not the conclusion, first part of their conclusion reads

"Both of our custom benchmarks show ATI cards leading without anisotropic filtering and antialiasing enabled, with NVIDIA taking over when the options are enabled."

Which is completely out of phase with the new results.

Well, I still give 'em points for updating the graph so quickly. That is a much different reaction than last time and a muchly better way of handling it. 8)

Besides, if they changed the conclusion the changed graph results wouldn't be as "transparent".

 

[thop]

I smell a fish.

The word is already out there, so the update is too late.

 

[BRiT]

The sad thing is, they didn't even update their conclusion paragraphs. The evidence observed contradicts their conclustion:

Both of our custom benchmarks show ATI cards leading without anisotropic filtering and antialiasing enabled, with NVIDIA taking over when the options are enabled. We didn't see much improvement from the new SM3.0 path in our benchmarks either. Of course, it just so happened that we chose a level that didn't really benefit from the new features the first time we recorded a demo. And, with the mangoriver benchmark, we were looking for a level to benchmark that didn't follow the style of benchmarks that NVIDIA provided us with in order to add perspective.

EDIT: damn, I must still be sleeping since I didnt see this mentioned just a couple of posts earlier.

 

[Mr. Travis]

the sad part is how common basic errors are on hardware sites... would it kill anyone to take the time to do things right?

 

[Humus]

Anand seems to be very frequent on making errors.

 

[rainz]

Anand seems to be very frequent on making errors.

Yeh but i think it's not really an " errors ".. If you know what i mean.

EDIT: and Btw Humus ! nice demo again ! and continue your good job !

RainZ

 

[jvd]

My apolgies to jvd and rainz for saying now you get speed with AA/AF with SM3, nope Ati is still king when it comes to AA and AF in this game

no prob , thats what happens on the net .

Btw the 4.7 betas are even faster than these

 

[digitalwanderer]

Anand seems to be very frequent on making errors.

Yeh but i think it's not really an " errors ".. If you know what i mean.

It is a bit suspect that he didn't change his conclusion even though the new results invalidate it. :?

 

[Kombatant]

I believe they're not doing anything fishy... they were just sloppy with the said review (which imho is inexcusable for a "professional" site, but anyways)...