vello/piet-gpu-types/src/lib.rs

// Structures used only internally probably don't need to be pub.

pub mod annotated;
pub mod bins;
pub mod encoder;
pub mod pathseg;
pub mod ptcl;
pub mod scene;
pub mod state;
pub mod test;
pub mod tile;
pub mod tilegroup;
Binning stage Adds a binning stage. This is a first draft, and a number of loose ends exist. 2020-05-13 06:38:26 +10:00			`// Structures used only internally probably don't need to be pub.`

Store annotated elements Apply transform to paths and annotate with computed linewidth and bounding box information, storing the result. 2020-05-13 03:53:54 +10:00			`pub mod annotated;`
Binning stage Adds a binning stage. This is a first draft, and a number of loose ends exist. 2020-05-13 06:38:26 +10:00			`pub mod bins;`
Add scene types This patch adds a module that contains both scene and ptcl types (very lightly adapted from piet-metal), as well as infrastructure for encoding Rust-side. WIP, it's not wired up in either the shader or on the Rust side. 2020-04-17 11:14:09 +10:00			`pub mod encoder;`
Experiment with new sorting scheme Path segments are unsorted, but other elements are using the same sort-middle approach as before. This is a checkpoint. At this point, there are unoptimized versions of tile init and coarse path raster, but it isn't wired up into a working pipeline. Also observing about a 3x performance regression in element processing, which needs to be investigated. 2020-06-03 10:10:20 +10:00			`pub mod pathseg;`
Add scene types This patch adds a module that contains both scene and ptcl types (very lightly adapted from piet-metal), as well as infrastructure for encoding Rust-side. WIP, it's not wired up in either the shader or on the Rust side. 2020-04-17 11:14:09 +10:00			`pub mod ptcl;`
			`pub mod scene;`
Experimenting with sort-middle Starting a prototype that explores the sort-middle approach. This commit has a prefix sum pass computing state per element. 2020-05-12 13:01:06 +10:00			`pub mod state;`
Add f16 support. Handling f16 requires special work, compared to other scalars, as the minimum conversion operation for u32->f16 in GLSL (unpackHalf2x16) loads two f16s from one u32. This means that in order to minimize unnecessary calls to unpackHalf2x16, we should look-ahead to see if the current f16 has already been extracted in the process of dealing with the last f16. Similar considerations exist for write operations, where we want to pack, when possible, two f16s in one go (using packHalf2x16). 2020-04-19 10:04:19 +10:00			`pub mod test;`
Experiment with new sorting scheme Path segments are unsorted, but other elements are using the same sort-middle approach as before. This is a checkpoint. At this point, there are unoptimized versions of tile init and coarse path raster, but it isn't wired up into a working pipeline. Also observing about a 3x performance regression in element processing, which needs to be investigated. 2020-06-03 10:10:20 +10:00			`pub mod tile;`
First draft of kernel 1 Output of kernel 1 is validated by simple inspection, next step is to wire it up properly. 2020-04-21 10:15:36 +10:00			`pub mod tilegroup;`