occlusion computes its three channels as one vector instead of writing a
vector component through a loop index, which FXC cannot address and tried
to unroll every enclosing loop for (X3550/X3511). classic3d now also
builds like the effect kernels: if a compiler rejects it, 3D runs render
on the CPU instead of the whole compositor failing to start. The ignored
FXC diagnostic test now checks classic3d instead of the stale pointwise
failure.
Resolve the conflict in crates/gpu/src/fx_stylize.rs: main moved ops::warp to
the ops::WarpOptions parameter struct; the branch's per-sample transform loop
(motion-blur samples, now passing the sample by reference) uses the new
signature with the branch's reference argument kept.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Accelerator::comp_frame_yuv420 (default None) renders a top-level comp frame and
converts it over the background to the H.264 export's 8-bit BT.709 limited-range
4:2:0, every byte within one level of the CPU conversion (the GPU kernel follows
it operation for operation, including NaN and half-away-from-zero ties, but a
driver may contract its arithmetic). A 4K frame then reads back 12 MB of planes
instead of 133 MB of RGBA f32.
The H.264 export asks the accelerator for planes when a frame needs nothing after
the render (progressive, project depth, no crop/resize, rendered at the output
size) and otherwise renders and converts on the CPU as before. Deferred
readbacks (a browser worker) keep the pass-based path. Sizes the kernel can't
take (width not a multiple of 8, odd height) are declined before any GPU work.
Like Cx::frames, h264_frames caps GPU frames in flight only while the
accelerator is up (map_frames_while with the same liveness check): once the GPU
stops for good, the remaining frames render on the CPU side by side instead of
one at a time.
rgba_to_yuv420 moves to effectcraft_raster::yuv so the export and the GPU tests
share one reference. The GPU tests skip without an adapter (tests::gpu()), as
the crate's others do.
Turbulent Displace read its noise through value_noise, whose quintic fade also
eased the evolution axis. Near every whole revolution the pattern hardly moved:
Evolution 0 and 37 degrees (or 360 and 400) differed by well under a hundredth of
the frame, so different Evolution values looked identical, and the motion sped up
and slowed down within each revolution (the #1816 symptom).
Add value_noise_linear_z (x and y keep the fade, z is linear) on the CPU and in the
WGSL kernels and use it for Turbulent Displace on both paths. value_noise is
unchanged, so the other effects and the pinned simulation goldens are unaffected.
Tests: CPU test that non-equivalent Evolution values (fractional, past 360, past
720) all differ and that Cycle Evolution still matches 0/360; unit test of the
linear z axis; GPU-vs-CPU cases at small and large Evolution.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Claude-Session: https://claude.ai/code/session_01FYXhNN7W7nB6sKjwwU7CDM
The WGSL grammar test scanner indexed token characters directly, so a malformed shader could panic the test instead of failing it. Checked access turns that into a reported violation.
The PCF tap jitter was keyed on the receiver's 1/8-texel cell in the
shadow map. The GPU computes the receiver's map position with last-bit
differences from the CPU, so a receiver next to a cell line could take
another cell's jitter on one backend, and every tap moved. In a wide
penumbra that changed the coverage by several 8-bit levels:
gpu_matches_cpu_wide_penumbrae failed on Windows with 15 of 4159
shadowed pixels on an NVIDIA RTX 5080 and 22 on an AMD Radeon iGPU,
against a budget of 8.
tap_jitter now blends the hashes at the four corners of the cell by
the receiver's position in it, in shade.rs and in advanced3d.wgsl. At
the lattice points the jitter is the per-cell hash it was before. A
CPU test checks that the coverage does not jump at a lattice line.
Poison recovery, OsStr::to_str and char predicates only; S4325 casts are
f64 to f32 and stay.
Co-Authored-By: Claude Sonnet 5.5 <noreply@anthropic.com>
Content r from Center smears out to r / (1 - s), so the buffer is padded by
the farthest corner's reach (capped at 4096, skipped on adjustment layers),
on the CPU and GPU paths alike.
* cargo fmt
* GPU tests: llvmpipe on Vulkan gets the boundary-flip allowance; skip the Reshape boundary cases that crash lavapipe (#2238)
* GPU tests: the llvmpipe boundary-flip allowance is an opt-in of the two tests that need it
bflatastic