Firestar99
|
ad99c14b29
|
Shaders: Provide WgpuExecutor via Scope (#3144)
shaders: automatically connect wgpu-executor
|
2025-09-05 18:51:17 +00:00 |
|
Firestar99
|
f12b4da549
|
Shaders: add `BufferStruct` to support bool and enums (#3109)
* node-macro: modernize `node` macro
* node-macro: add `CrateIdent` struct containing resolved crate paths
* shaders: add trait `BufferStruct` and derive macro
* shaders: `gamma_correction` and `channel_mixer` gpu nodes
* shaders: `selective_color` gpu node
* shaders: `brightness_contrast_classic` gpu node
* shaders: append GPU to display name
* node-macro: fixup doc links
* shaders: consistently append " GPU" to all shader node names
|
2025-09-05 16:32:41 +00:00 |
Dennis Kobert
|
acd7ba38cc
|
Graphene: Fine-grained context caching (#2500)
* RFC: Fine Grained Context Caching
* Fix typos
* Fix label
* Add description of inject traits
* Explicitly support context modification
* Start implementation of context invalidation
* Add inject trait variants
* Route Extract / Inject traits to the proto nodes
* Implement context dependency analysis
* Implement context modification node insertion
* Fix erronous force graph run message
* Fix Extract* Inject* annotations in the nodes
* Require Hash implementation for VarArgs
* Fix nullification node insertion
* Cross of done items unresolved questions section
* Update Cargo.lock
* Fix context features propagation
* Update demo artwork
* Remove BondlessFootprint and FreezeRealTime nodes
* Fix migration
* Add migrations for adding context features to old networks
* Always update real time regardless of animation state
* Cargo fmt
* Fix tests
* Readd sed command to hopefully fix profile result parsing
* Add debug output to profiling pr
* Use new totals instead of summaries for for iai results
* Even more debugging
* Use correct debug metrics (hopefully)
* Add more MemoNode implementations
* Add context features annotation to shader node macro
* Cleanup
* Time -> RealTime
* Code review
---------
Co-authored-by: Keavon Chambers <keavon@keavon.com>
|
2025-09-05 11:44:26 +00:00 |
|
Firestar99
|
5d441c2e18
|
Shaders: runtime and shader node codegen (#2985)
* shader-rt: initial
* shader-rt: fix recursion when generating shader node
* shader-rt: replace gpu node's args and ret types with `Raster<GPU>`
* shader-rt: properly cfg out the gpu node
* shader-rt: fix `impl Context` in the wrong places
* shader-rt: disable gpu blend node, needs two images
* shader-rt: connect shader runtime
* shader-rt: pass WgpuExecutor by reference
* shader-rt: correct bindings with derpy arg buffer
* shader-rt: manual pipeline layout, fixing errors when bindings got DCE'd
* shader-rt: correct RT format, working invert gpu node
* shader-rt: cleanup codegen with common sym struct
* shader-rt: correct arg buffer handling
* shader-nodes feature: put shader nodes behind feature gate
* shader-nodes feature: rename any `gpu_node` to `shader-node`
* shaders-rt: fix wgpu label name
* shaders-rt: explain fullscreen_vertex coordinates with a drawing
|
2025-09-05 08:33:53 +02:00 |
|
Firestar99
|
a10103311e
|
Shaders: rust-gpu integration to compile shader nodes to WGSL (#3097)
* shaders: shader compilation setup
* nix: use rustc_codegen_spirv.so from nix
* shaders: codegen for per_pixel_adjust shader nodes
* shaders: disable nodes needing bool
* shaders: `#[repr(u32)]` some enums
* shaders: add lint ignores from rust-gpu
* shaders: fix node-macro tests
* gcore-shaders: toml cleanup
* shader-nodes feature: put rust-gpu to wgsl compile behind feature gate
* shaders: fix use TokenStream2
* shaders: allow providing shader externally
* Update iai runner in workflow
---------
Co-authored-by: Timon Schelling <me@timon.zip>
Co-authored-by: Dennis Kobert <dennis@kobert.dev>
|
2025-09-02 14:10:32 +00:00 |