Struct OffscreenRenderTiming

pub struct OffscreenRenderTiming {Show 47 fields
    pub vello_submit_us: u128,
    pub copy_submit_us: u128,
    pub settle_wait_us: u128,
    pub full_attempts: u32,
    pub damage_setup_us: u128,
    pub damage_target_creations: u32,
    pub sparse_submissions: u32,
    pub sparse_coarse_active_bins_submissions: u32,
    pub sparse_coarse_adaptive_submissions: u32,
    pub sparse_coarse_adaptive_dense_submissions: u32,
    pub sparse_coarse_native_bins: u64,
    pub sparse_coarse_selected_native_bins: u64,
    pub sparse_path_tile_bounds_submissions: u32,
    pub sparse_path_tile_bounds_tiles: u64,
    pub sparse_coarse_opaque_prune_submissions: u32,
    pub sparse_solid_fuse_submissions: u32,
    pub sparse_segment_row_mask_submissions: u32,
    pub canonical_segment_order_sparse_selected_attempts: u32,
    pub canonical_segment_order_full_selected_attempts: u32,
    pub area_ppt8_sparse_selected_attempts: u32,
    pub area_ppt8_full_selected_attempts: u32,
    pub clip_single_group_sparse_selected_attempts: u32,
    pub clip_single_group_full_selected_attempts: u32,
    pub sparse_fine_quad_skip_submissions: u32,
    pub sparse_fine_quad_total_quads: u64,
    pub sparse_fine_quad_active_quads: u64,
    pub sparse_fine_quad_skipped_quads: u64,
    pub sparse_fine_tiles: u64,
    pub sparse_written_pixels: u64,
    pub coarse_prefix_pixels: u64,
    pub gpu_timestamp_span_ns: u128,
    pub gpu_timestamp_pass_ns: u128,
    pub gpu_timestamp_last_pass_ns: u128,
    pub gpu_timestamp_other_pass_ns: u128,
    pub gpu_timestamp_pass_count: u32,
    pub gpu_timestamp_expected_submissions: u32,
    pub gpu_timestamp_sampled_submissions: u32,
    pub gpu_timestamp_missing_submissions: u32,
    pub gpu_dispatch_timestamp_expected_submissions: u32,
    pub gpu_dispatch_timestamp_sampled_submissions: u32,
    pub gpu_dispatch_timestamp_missing_submissions: u32,
    pub gpu_dispatch_timestamp_stale_submissions: u32,
    pub gpu_dispatch_timestamp_present_stages: u32,
    pub gpu_dispatch_timestamp_stage_duration_ns: [u128; 32],
    pub gpu_dispatch_timestamp_stage_dispatch_count: [u32; 32],
    pub gpu_dispatch_timestamp_unknown_duration_ns: u128,
    pub gpu_dispatch_timestamp_unknown_dispatch_count: u32,
}

Expand description

Host-side phases of one completed offscreen render, populated only by OffscreenRenderer::render_damage_regions_gpu_wait_profiled.

vello_submit_us includes scene resolution, uploads, and Vello’s shared queue submission, excluding separately measured copy encoding. copy_submit_us covers only dirty-copy command encoding in Vello’s post-render callback. settle_wait_us is the blocking GPU-complete wait while collecting Vello’s bump counters.

Fields§

§vello_submit_us: u128§copy_submit_us: u128§settle_wait_us: u128§full_attempts: u32§damage_setup_us: u128§damage_target_creations: u32§sparse_submissions: u32§sparse_coarse_active_bins_submissions: u32

Sparse submissions that actually selected the opt-in native active-bin coarse shader. Capability alone is not counted because oversized selectors fall back to legacy coarse.

§sparse_coarse_adaptive_submissions: u32

Sparse submissions made while Vello’s opt-in adaptive coarse policy was active.

§sparse_coarse_adaptive_dense_submissions: u32

Adaptive sparse submissions which deliberately retained dense coarse work.

§sparse_coarse_native_bins: u64

Sum of representable native coarse bins across adaptive sparse submissions.

§sparse_coarse_selected_native_bins: u64

Sum of selected native coarse bins across adaptive sparse submissions.

§sparse_path_tile_bounds_submissions: u32

Sparse submissions where Vello actually applied path-level native tile bounds.

§sparse_path_tile_bounds_tiles: u64

Sum of native bounding-box tile areas from actually bounded sparse submissions. This is diagnostic coverage accounting, not an allocation or memory-saving metric.

§sparse_coarse_opaque_prune_submissions: u32

Sparse submissions where the coarse opaque-prune shader was actually selected. This does not claim a number of tiles or commands pruned, or any resource savings.

§sparse_solid_fuse_submissions: u32

Sparse submissions where the solid-fuse producer/consumer shader path was selected. This does not claim command fusion counts or any resource savings.

§sparse_segment_row_mask_submissions: u32

Sparse submissions where the segment-row-mask shader path was actually selected. This does not claim a number of rows skipped or any resource savings.

§canonical_segment_order_sparse_selected_attempts: u32

Sparse submissions where canonical segment ordering was actually recorded. This is route-attempt telemetry, not a completion or speed claim.

§canonical_segment_order_full_selected_attempts: u32

Full submissions where canonical segment ordering was actually recorded. This is route-attempt telemetry, not a completion or speed claim.

§area_ppt8_sparse_selected_attempts: u32

Sparse submissions where the Area PPT8 shader was actually selected. This is route-attempt telemetry, not a completion or speed claim.

§area_ppt8_full_selected_attempts: u32

Full submissions where the Area PPT8 shader was actually selected. This is route-attempt telemetry, not a completion or speed claim.

§clip_single_group_sparse_selected_attempts: u32

Sparse submissions where the single-workgroup clip leaf shader was selected. This is route-attempt telemetry, not a completion or speed claim.

§clip_single_group_full_selected_attempts: u32

Full submissions where the single-workgroup clip leaf shader was selected. This is route-attempt telemetry, not a completion or speed claim.

§sparse_fine_quad_skip_submissions: u32

Sparse submissions where the Area fine shader actually skipped zero-mask quad invocations.

§sparse_fine_quad_total_quads: u64

Total fine invocation groups considered by the applied quad-skip path.

§sparse_fine_quad_active_quads: u64

Non-zero-mask fine invocation groups evaluated by the applied quad-skip path.

§sparse_fine_quad_skipped_quads: u64

Zero-mask fine invocation groups skipped by the applied quad-skip path.

§sparse_fine_tiles: u64§sparse_written_pixels: u64§coarse_prefix_pixels: u64§gpu_timestamp_span_ns: u128

Rounded GPU timestamp span across verified submissions.

§gpu_timestamp_pass_ns: u128

Rounded sum of individual GPU pass durations across verified submissions.

§gpu_timestamp_last_pass_ns: u128

Rounded duration of the final recorded compute pass in verified submissions.

§gpu_timestamp_other_pass_ns: u128

Rounded duration of all earlier recorded compute passes in verified submissions.

§gpu_timestamp_pass_count: u32§gpu_timestamp_expected_submissions: u32§gpu_timestamp_sampled_submissions: u32§gpu_timestamp_missing_submissions: u32

Submissions for which timestamp data was disabled, unavailable, stale, or failed.

§gpu_dispatch_timestamp_expected_submissions: u32

Matched submissions expected to have an opt-in per-dispatch sample.

§gpu_dispatch_timestamp_sampled_submissions: u32§gpu_dispatch_timestamp_missing_submissions: u32

Missing dispatch samples are not treated as zero GPU time.

§gpu_dispatch_timestamp_stale_submissions: u32

A completed sample belonging to an older submission was discarded.

§gpu_dispatch_timestamp_present_stages: u32

Bit stage_id - 1 is set when that stable stage was present.

§gpu_dispatch_timestamp_stage_duration_ns: [u128; 32]§gpu_dispatch_timestamp_stage_dispatch_count: [u32; 32]§gpu_dispatch_timestamp_unknown_duration_ns: u128§gpu_dispatch_timestamp_unknown_dispatch_count: u32