dispatch()

void MayaFlux::Core::VKComputePipeline::dispatch	(	vk::CommandBuffer	cmd,
		uint32_t	group_count_x,
		uint32_t	group_count_y,
		uint32_t	group_count_z
	)		const

Dispatch compute workgroups.

Parameters

cmd	Command buffer
group_count_x	Number of workgroups in X dimension
group_count_y	Number of workgroups in Y dimension
group_count_z	Number of workgroups in Z dimension

Executes the compute shader with the specified number of workgroups. Total invocations = group_count * local_size (from shader).

Example: // Shader: layout(local_size_x = 256) in; // Process 1M elements: 1,000,000 / 256 = 3,907 workgroups pipeline.dispatch(cmd, 3907, 1, 1);

Must be called after bind() and bind_descriptor_sets().

Definition at line 261 of file VKComputePipeline.cpp.

{
    if (!m_pipeline) {
        MF_ERROR(Journal::Component::Core, Journal::Context::GraphicsBackend,
            "Cannot dispatch invalid compute pipeline");
        return;
    }
 
    if (group_count_x == 0 || group_count_y == 0 || group_count_z == 0) {
        MF_WARN(Journal::Component::Core, Journal::Context::GraphicsBackend,
            "Dispatching with zero workgroups ({}x{}x{})",
            group_count_x, group_count_y, group_count_z);
        return;
    }
 
    cmd.dispatch(group_count_x, group_count_y, group_count_z);
}

References MayaFlux::Journal::Core, MayaFlux::Journal::GraphicsBackend, m_pipeline, MF_ERROR, and MF_WARN.

Referenced by dispatch_1d(), dispatch_2d(), and dispatch_3d().

Here is the caller graph for this function: