Struct ComputeClient

pub struct ComputeClient<Server, Channel>
where
    Server: ComputeServer,
{ /* private fields */ }

The ComputeClient is the entry point to require tasks from the ComputeServer. It should be obtained for a specific device via the Compute struct.

Implementations

impl<Server, Channel> ComputeClient<Server, Channel>

where Server: ComputeServer, Channel: ComputeChannel<Server>,

pub fn new( channel: Channel, properties: DeviceProperties<<Server as ComputeServer>::Feature>, ) -> ComputeClient<Server, Channel>

Create a new client.

pub async fn read_async(&self, bindings: Vec<Binding>) -> Vec<Vec<u8>>

Given bindings, returns owned resources as bytes.

pub fn read(&self, bindings: Vec<Binding>) -> Vec<Vec<u8>>

Given bindings, returns owned resources as bytes.

Remarks

Panics if the read operation fails.

pub async fn read_one_async(&self, binding: Binding) -> Vec<u8>

Given a binding, returns owned resource as bytes.

pub fn read_one(&self, binding: Binding) -> Vec<u8>

Given a binding, returns owned resource as bytes.

Remarks

Panics if the read operation fails.

pub fn get_resource(&self, binding: Binding) -> BindingResource<Server>

Given a resource handle, returns the storage resource.

pub fn create(&self, data: &[u8]) -> Handle

Given a resource, stores it and returns the resource handle.

pub fn empty(&self, size: usize) -> Handle

Reserves size bytes in the storage, and returns a handle over them.

pub fn execute( &self, kernel: <Server as ComputeServer>::Kernel, count: CubeCount, bindings: Vec<Binding>, )

Executes the kernel over the given bindings.

pub unsafe fn execute_unchecked( &self, kernel: <Server as ComputeServer>::Kernel, count: CubeCount, bindings: Vec<Binding>, )

Executes the kernel over the given bindings without performing any bound checks.

Safety

Without checks, the out-of-bound reads and writes can happen.

pub fn flush(&self)

Flush all outstanding commands.

pub async fn sync(&self)

Wait for the completion of every task in the server.

pub async fn sync_elapsed(&self) -> Result<Duration, TimestampsError>

Wait for the completion of every task in the server.

pub fn properties( &self, ) -> &DeviceProperties<<Server as ComputeServer>::Feature>

Get the features supported by the compute server.

pub fn memory_usage(&self) -> MemoryUsage

Get the current memory usage of this client.

pub async fn profile<O, Fut, Func>(&self, func: Func) -> O

where Fut: Future<Output = O>, Func: FnOnce() -> Fut,

When executing operation within the profile scope, you can call sync_elapsed safely even in multithreaded workloads. Creates a profiling scope that enables safe timing measurements in concurrent contexts.

Operations executed within this scope can safely call sync_elapsed() to measure elapsed time, even in multithreaded environments. The measurements are thread-safe and properly synchronized.

pub fn enable_timestamps(&self)

Enable timestamp collection on the server for performance profiling.

This feature records precise timing data for server operations, which can be used for performance analysis and benchmarking.

Warning

This should only be used during development and benchmarking, not in production, as it significantly impacts server throughput and performance. The overhead comes from frequent timestamp collection and storage.

Example

server.enable_timestamps();
// Run your benchmarks/operations
let duration = server.sync_elapsed();

Trait Implementations

impl<S, C> Clone for ComputeClient<S, C>

where S: ComputeServer, C: ComputeChannel<S>,

fn clone(&self) -> ComputeClient<S, C>

Returns a copy of the value.

fn clone_from(&mut self, source: &Self)

Performs copy-assignment from source.

impl<Server, Channel> Debug for ComputeClient<Server, Channel>

where Server: Debug + ComputeServer, Channel: Debug,

fn fmt(&self, f: &mut Formatter<'_>) -> Result<(), Error>

Formats the value using the given formatter.