Struct GPUCompute 

pub struct GPUCompute {
    pub pipeline: ComputePipeline,
    pub bind_group: BindGroup,
    /* private fields */
}

The GPU-resident form an uploaded Compute produces — its compiled pipeline (possibly shared with other Computes, see ComputePipelineCache) plus its own bind group.

Fields

pipeline: ComputePipeline

bind_group: BindGroup

Implementations

impl GPUCompute

pub fn update(&self, name: &str, data: &[u8])

Overwrites a named uniform/storage buffer’s contents in place — avoids rebuilding the whole bind group for a per-frame update.

pub fn update_value<T>(&self, name: &str, value: &T)

where T: ShaderType + WriteInto,

Same as update, but takes a typed value instead of raw bytes — same encase layout Compute::with_uniform_value/ with_storage_value use.

pub fn buffer(&self, name: &str) -> Option<&Buffer>

pub fn dynamic_buffer(&self, name: &str) -> Option<&DynamicBuffer>

Same as buffer, for a binding made via .with_dynamic_buffer — use DynamicBuffer::write_element on the result to update one element in place.