Originally posted by: moonlightcheese
The real feature of interest is the ability to load and filter 4 texture addresses from a single channel texture map.
would someone mind explaining this to me? it doesn't seem quite clear. what exactly is meant by a "single channel" texture map? what is the alternative to a "single channel" texture map? and what exactly is meant by "texture address"? is that just a texel?
Textures which describe color generally have four components at every location in the texture, and normally the hardware will load an address from a texture map, split the 4 channels and filter them independently.
by four components in a location, is he refering to "RGBA" (red, green, blue, alpha) channels on each pixel? and again... what does he mean by "address" from a "texture map".
In cases where single channel textures are used (ATI likes to use the example of a shadow map), the R520 will look up the appropriate address and will filter the single channel (letting the hardware's ability to filter 3 other components go to waste).
i thought R520 didn't have the quad pixel shader...? if R520 can also filter 4 "components" (whatever that's supposed to be) then what's the big deal? what's so new about it?
In what ATI calls it's Fetch4 feature, the R580 is capable of loading 3 other adjacent single channel values from the texture and filtering these at the same time.
this sentence isn't even proper english... ("filtering" should be "filter"). so by adjacent does he mean loading adjacent texels for whatever filtering scheme is used? (eg. anisotropic, bicubic, etc.)
This effectively loads 4 and filters four times the texture data when working with single channel formats. Traditional color textures, or textures describing vector fields (which make use of more than one channel per position in the texture) will not see any performance improvement, but for some soft shadowing algorithms performance increases could be significant.
again... don't know what is meant by "single channel" formats. and what is the difference between a traditional texture and one that describes vector fields? and what is meant by "uses" in that sentence? does he mean "read"? and what is "one channel per position in the texture"? is that one color channel per texel?
i mean no disrespect to the author in anyway and this is in no way a criticism. i'm just trying to show where i am having trouble understanding the wording and terms in this article. it just seems very vague to me and i am incredibly confused by this paragraph. i'd really like to understand the article much more but this piece of information just seems to elude my understanding of the R580 architecture. can anyone explain all of the parts in question there?