Why 32bit float point texture cannot be filtered by hw?

Started by
7 comments, last by dextml 14 years, 7 months ago
16bit floating point texture filtering is supported on latest cards, but not for 32bit ones. Why they cannot be filtered? What is the difficulty at hardware level?
Advertisement
You'd have to slam that addition floating point modules to each texture unit... which are quite large for 32bit floating point math, in addition to the 16bit fp, 32bit integer units and stuff.

BTW, both the NVidia Geforce cards (starting with the 8000 series) and the ATI Radeon cards (starting with HD3000 series, I think) can filter and blend fp32 textures.
----------
Gonna try that "Indie" stuff I keep hearing about. Let's start with Splatter.
sorry for the clarity. By saying the latest, i mean the latest platforms we should concern about. That's usually from GF7~GF200 (similar generations for ati's).

Thanks. never expected the answer comes this quickly.
Kudos for u.
R600+ from Ati and G80+ from Nvidia can filter 32bit floating point textures. R500 and G70 are limited to 16bit. Older hardware cannot filter floating point textures at all.

G200 most certainly *can* filter 32bit textures.

[OpenTK: C# OpenGL 4.4, OpenGL ES 3.0 and OpenAL 1.1. Now with Linux/KMS support!]

R500 can't filter fp16 textures, but it can multisample and blend them. NV40 and G70 could both filter fp16, but couldn't multisample. I think they could blend...but I can't remember off the top of my head.
Is there a table/wiki page that has the info on what cards support what things? (like this)
Quote:Original post by zedz
Is there a table/wiki page that has the info on what cards support what things? (like this)


There's a CardCaps PDF/Excel spreadsheet in the folder for the ConfigSystem sample that comes with the SDK, but it hasn't been updated since September 2007.
Everything DX10+ can filter/MSAA all non-integer formats in practice (although not all of them are technically required by DX10). That includes fp32, fp16, and all of the unorm formats. MSAA is supported on everything IIRC, although integer targets have to be manually resolved (for obvious reasons, same with filtering).

So no, there really aren't limitations on "modern" cards. Anything in the last few years is pretty orthogonal.
i appreciate your help.
Looks like starting from G80 and R600, everything becomes clear.

This topic is closed to new replies.

Advertisement