Struct TrackedCudaSlice 

pub struct TrackedCudaSlice<T: DeviceRepr> { /* private fields */ }

A typed view retaining its actual device allocation and memory budget.

Native owning slices cannot be extracted through a safe mutable borrow:

use xlog_cuda::memory::TrackedCudaSlice;
use cudarc::driver::CudaSlice;
fn native_owner(slice: &mut TrackedCudaSlice<u8>) -> &mut CudaSlice<u8> {
    slice
}

Device access must be prepared by XLOG; a passive allocation handle cannot bypass its access reservation through cudarc’s safe pointer API:

use xlog_cuda::memory::TrackedCudaSlice;
use cudarc::driver::DevicePtr;
fn native_read(slice: &TrackedCudaSlice<u8>) {
    fn accepts_native<T: DevicePtr<u8>>(_: &T) {}
    accepts_native(slice);
}

Implementations

impl<T: DeviceRepr> TrackedCudaSlice<T>

pub fn len(&self) -> usize

pub fn is_empty(&self) -> bool

pub fn stream(&self) -> &Arc<CudaStream>

pub fn device_ptr(&self) -> &CUdeviceptr

pub fn device_ptr_value(&self) -> CUdeviceptr

pub fn view(&self) -> DeviceMemoryView<T>

A passive view retaining this allocation; obtaining it performs no device access.

pub fn try_slice( &self, range: impl RangeBounds<usize>, ) -> Option<DeviceMemoryView<T>>

pub fn slice(&self, range: impl RangeBounds<usize>) -> DeviceMemoryView<T>

pub fn slice_mut( &mut self, range: impl RangeBounds<usize>, ) -> DeviceMemoryView<T>

pub fn memory_manager_ptr_value(&self) -> usize

Stable address of the memory manager that owns this allocation.

pub fn runtime_block(&self) -> Option<&DeviceBlock>

Runtime block identity, when present. Device-native and foreign owners have no runtime block ID but still participate in shared range admission.

pub fn into_bytes(self) -> TrackedCudaSlice<u8>

Reinterpret this typed allocation as a raw byte allocation.

This is a zero-copy conversion used by XLOG’s columnar CudaBuffer representation, which stores device memory as untyped bytes + a schema. The conversion preserves the underlying [Backing] — runtime-routed slices remain runtime-routed, legacy cudarc slices remain cudarc-routed — so deallocation continues to match the original allocator.

Trait Implementations

impl<T: DeviceRepr> AsKernelParam for &TrackedCudaSlice<T>

fn as_kernel_param(&self) -> *mut c_void

impl<T: DeviceRepr> AsKernelParam for &mut TrackedCudaSlice<T>

fn as_kernel_param(&self) -> *mut c_void

impl<T: DeviceRepr> DeviceRead<T> for TrackedCudaSlice<T>

fn device_view(&self) -> DeviceMemoryView<T>

impl<T: DeviceRepr> DeviceSlice<T> for TrackedCudaSlice<T>

fn len(&self) -> usize

fn stream(&self) -> &Arc<CudaStream>

fn num_bytes(&self) -> usize

fn is_empty(&self) -> bool

impl<T: DeviceRepr> DeviceWrite<T> for TrackedCudaSlice<T>

impl From<TrackedCudaSlice<u8>> for CudaColumn

fn from(value: TrackedCudaSlice<u8>) -> Self

Converts to this type from the input type.

impl<'a, T: DeviceRepr> IntoKernelParamStorage for &'a TrackedCudaSlice<T>

type Storage = DeviceParamStorage<'a>

fn into_kernel_param_storage(self) -> Self::Storage

impl<T: DeviceRepr> IntoKernelParamStorage for &mut TrackedCudaSlice<T>

type Storage = DeviceParamStorage<'static>

fn into_kernel_param_storage(self) -> Self::Storage