@pmndrs/bvh
    Preparing search index...

    Interface HplocBatchPlanOptions

    Options of planGpuHplocBatches.

    interface HplocBatchPlanOptions {
        maxBatchBytes?: number;
        mortonCodes?: HplocMortonCodes;
        parents?: boolean;
        quantizeOnEmit?: boolean;
        triangleLeaves?: GpuLayoutTriangleLeaves;
    }

    Hierarchy (View Summary)

    Index
    maxBatchBytes? mortonCodes? parents? quantizeOnEmit? triangleLeaves?
    maxBatchBytes?: number

    Each batch's GPU bytes (inputs, scratch and outputs) stay at or under this, except a single mesh bigger than it, which gets a batch of its own. Default: no cap (batches as big as the device's buffers allow).

    mortonCodes?: HplocMortonCodes

    The Morton codes the batches will use ('tinybvh-63' needs 8 more bytes per triangle). Default 'cubic-30'.

    parents?: boolean

    The batches will write their parents output (GpuHplocBatchOptions.parents). Default true.

    quantizeOnEmit?: boolean

    The batches will write quantized nodes in their emit pass (HplocOptions.quantizeOnEmit): 32 bytes per node and 32 per mesh, not 64 per node. Default false.

    triangleLeaves?: GpuLayoutTriangleLeaves

    The trees' triangle format, whose output size the cap counts. Default 'records'.

    Settings
    Member Visibility

    On This Page
    Properties