DX9.1, wtf?

cthellis42 said:
Did they make Quincunx the only allowable AA method in 9.1? ;)

If you can keep a secret...
In the upcoming Series 55 drivers, expected for the NV40 launch in February, they're adding a new AA mode to all NV3x GPUs: "2x 9-tap".

This has been supported in hardware for a long time, since 4x 9-tap is theorically possible, and it's really just a combination of a 2x and a 9 sample filter we've got for 2x 9-tap.


The breaking news though is that Microsoft is introducing a new AQ or "Automatic Quality" tool in DX9.1.
This looks at the FPS of your game ( the game is detected based on its window's identificator, as the Detonator drivers do ), and based on your FPS, the next time you run the game, AA and AF settings are automatically applied for you.

And, believe it or not, NVIDIA managed to convince Microsoft that their only AA solution capable of competing with ATI's 4x and 6x, edge quality-wise, is 2x 9-tap.

This feature can also be fully disabled, or, a certain favorite of reviewers, it can be "forced" based on the results of two log files, thus coming from two different cards. It then averages then and gives similar IQ settings for both cards.

The result? It is widely expected by IHVs and ISVs that this utility will suggest you to compare 2x 9-tap with ATI's 6x. Reviewers will be forced to use this utility by both NVIDIA and Microsoft.

But hey, if you publish this, I'd have to kill you :LOL:


Uttar

P.S.: Anyone who believed this for a single second is an idiot.
There, happy, Dig? :p
 
Uttar said:
cthellis42 said:
Did they make Quincunx the only allowable AA method in 9.1? ;)

If you can keep a secret...
In the upcoming Series 55 drivers, expected for the NV40 launch in February, they're adding a new AA mode to all NV3x GPUs: "2x 9-tap".

This has been supported in hardware for a long time, since 4x 9-tap is theorically possible, and it's really just a combination of a 2x and a 9 sample filter we've got for 2x 9-tap.


The breaking news though is that Microsoft is introducing a new AQ or "Automatic Quality" tool in DX9.1.
This looks at the FPS of your game ( the game is detected based on its window's identificator, as the Detonator drivers do ), and based on your FPS, the next time you run the game, AA and AF settings are automatically applied for you.

And, believe it or not, NVIDIA managed to convince Microsoft that their only AA solution capable of competing with ATI's 4x and 6x, edge quality-wise, is 2x 9-tap.

This feature can also be fully disabled, or, a certain favorite of reviewers, it can be "forced" based on the results of two log files, thus coming from two different cards. It then averages then and gives similar IQ settings for both cards.

The result? It is widely expected by IHVs and ISVs that this utility will suggest you to compare 2x 9-tap with ATI's 6x. Reviewers will be forced to use this utility by both NVIDIA and Microsoft.

But hey, if you publish this, I'd have to kill you :LOL:


Uttar
You enjoy making stuff up just to fuck with people, don't you?
 
I'm sure we all do. ;)

Here is some benches of DX9.1:

UT03: 10000000000000fps

Here is some benches of DX9.0b:

UT03: 1fps

See the difference? :LOL:
 
DaveBaumann said:
Basically, whats being referred to fo the most part is the VS/PS_2_a profile for the VS/PS2 extended features of GeForce FX - this already became available as aprt of the summer update to DX9's SDK and is part of the DX9.0b redistributable.
Technically, it isn't. Whole HLSL (and thus any profiles) aren't part of DX runtime libraries (they are part of SDK only). DX9.0b was released before Summer Update, and it differs from 9.0a just by fixing some security hole in DirectPlay.
 
Uttar thats not funny the number of nvidiots who would claim it looks superior would be annoying ( just look at all the starstorm XP users and their -15 mipmap LOD bais ). The other sad thing is blurring the crap out of image is considered atleast in some circles valid AA :/ if we were using better magnification filters then linear you could almost say render it to a quarter of the screen size texture and just magnify it.
 
rashly said:
i wouldnt mind a dx91 with fbuffer support.
I've seen a statement from ATI that they don't want to expose F-Buffer in DX and only use it for GLslang because of certain implications. It sure may be because of resource limits which have to be "hard" in DX, whereas in GL you can decide on a per-shader basis whether it will run in hardware or not.
 
Uttar said:
cthellis42 said:
Did they make Quincunx the only allowable AA method in 9.1? ;)

If you can keep a secret...
In the upcoming Series 55 drivers, expected for the NV40 launch in February, they're adding a new AA mode to all NV3x GPUs: "2x 9-tap".

This has been supported in hardware for a long time, since 4x 9-tap is theorically possible, and it's really just a combination of a 2x and a 9 sample filter we've got for 2x 9-tap.


The breaking news though is that Microsoft is introducing a new AQ or "Automatic Quality" tool in DX9.1.
This looks at the FPS of your game ( the game is detected based on its window's identificator, as the Detonator drivers do ), and based on your FPS, the next time you run the game, AA and AF settings are automatically applied for you.

And, believe it or not, NVIDIA managed to convince Microsoft that their only AA solution capable of competing with ATI's 4x and 6x, edge quality-wise, is 2x 9-tap.

This feature can also be fully disabled, or, a certain favorite of reviewers, it can be "forced" based on the results of two log files, thus coming from two different cards. It then averages then and gives similar IQ settings for both cards.

The result? It is widely expected by IHVs and ISVs that this utility will suggest you to compare 2x 9-tap with ATI's 6x. Reviewers will be forced to use this utility by both NVIDIA and Microsoft.

But hey, if you publish this, I'd have to kill you :LOL:


Uttar

How much do you want to bet that someone somewhere will take this seriously :LOL:
 
StealthHawk said:
How much do you want to bet that someone somewhere will take this seriously :LOL:
No smiley, I'm just waiting for someone to point this out to me when I'm arguing with 'em somewhere as "proof". :rolleyes:
 
digitalwanderer said:
StealthHawk said:
How much do you want to bet that someone somewhere will take this seriously :LOL:
No smiley, I'm just waiting for someone to point this out to me when I'm arguing with 'em somewhere as "proof". :rolleyes:
And that is exactly why I have created DSTP--the Dopeslap Transmission Protocol. Or we could send a Special Forces team deep into China to find the secret lab where they're keeping Vambat and let him loose upon the poor bastard.
 
The Baron said:
digitalwanderer said:
StealthHawk said:
How much do you want to bet that someone somewhere will take this seriously :LOL:
No smiley, I'm just waiting for someone to point this out to me when I'm arguing with 'em somewhere as "proof". :rolleyes:
And that is exactly why I have created DSTP--the Dopeslap Transmission Protocol. Or we could send a Special Forces team deep into China to find the secret lab where they're keeping Vambat and let him loose upon the poor bastard.
I still want a device that lets you stab people in the face over the internet. :rolleyes:
 
digitalwanderer said:
I still want a device that lets you stab people in the face over the internet. :rolleyes:

Right. 'Cause you'd be the only one who'd have one.
 
Back
Top