Keymaker
12-10-2003, 17:22
Da Beoynd3D:
Is the HDR rendering utilising float buffers? (i.e. will this work on NV hardware under the DX9 path if they haven't enabled float target support.)
[Brian Jacobson] Yes, although we're working with NVIDIA to get a non-float buffer version working.
[Gary McTaggart] Actually, with ATI, we use 16 bit integer buffers. We don't use a float buffer. NVIDIA is adding support for buffers that are lower than 16-bit integer.
What range does HDR use? (i.e. does it use [-8, 8] such that it could operate in a PS1.4 model?)
[Gary McTaggart] We currently represent light values in the [0,16] range. For HDR, we depend heavily on the fact that our dx9 shaders all render in a single pass. . This makes it difficult to get working with ps1.4.
The presentations states that one of the things the mixed mode does is "Trades off texture fetches for pixel shader instruction count" - is that using Vector normalization by cubemaps rather than PS math? Is there anything else like this as well?
[Gary McTaggart] Most of this (if not all) is for vector normalization.
Were the Shaders in the benchmark compiled with the latest version of HLSL (i.e. the version that orders the assembly for a more efficient use of the FX's register limitations)?
[Brian Jacobson] The ones used by the mixed mode path were.
Will the default DX9 HLSL game path utilise the HDR rendering?
[Brian Jacobson] Yes.
Does the "Mixed Mode" use any PS1.4 shaders in place of PS2.0 shaders used in the default DX9 path?
[Brian Jacobson] Not for the 5900. Yes for the 5600 and 5200.
Does the mixed mode use the precision hint on all available instructions, most instructions or are the majority still full precision?
[Brian Jacobson] The majority are partial precision; about 10%-20% are full precision.
Is the HDR rendering utilising float buffers? (i.e. will this work on NV hardware under the DX9 path if they haven't enabled float target support.)
[Brian Jacobson] Yes, although we're working with NVIDIA to get a non-float buffer version working.
[Gary McTaggart] Actually, with ATI, we use 16 bit integer buffers. We don't use a float buffer. NVIDIA is adding support for buffers that are lower than 16-bit integer.
What range does HDR use? (i.e. does it use [-8, 8] such that it could operate in a PS1.4 model?)
[Gary McTaggart] We currently represent light values in the [0,16] range. For HDR, we depend heavily on the fact that our dx9 shaders all render in a single pass. . This makes it difficult to get working with ps1.4.
The presentations states that one of the things the mixed mode does is "Trades off texture fetches for pixel shader instruction count" - is that using Vector normalization by cubemaps rather than PS math? Is there anything else like this as well?
[Gary McTaggart] Most of this (if not all) is for vector normalization.
Were the Shaders in the benchmark compiled with the latest version of HLSL (i.e. the version that orders the assembly for a more efficient use of the FX's register limitations)?
[Brian Jacobson] The ones used by the mixed mode path were.
Will the default DX9 HLSL game path utilise the HDR rendering?
[Brian Jacobson] Yes.
Does the "Mixed Mode" use any PS1.4 shaders in place of PS2.0 shaders used in the default DX9 path?
[Brian Jacobson] Not for the 5900. Yes for the 5600 and 5200.
Does the mixed mode use the precision hint on all available instructions, most instructions or are the majority still full precision?
[Brian Jacobson] The majority are partial precision; about 10%-20% are full precision.