pub enum CopyRoute {
DeviceIndexed,
WholeLayerPageable,
}Expand description
How a staged copy for one layer may be carried out.
Variants§
DeviceIndexed
The normal path: gather the missing expert rows from the registered bank into whichever slots the LRU picked.
WholeLayerPageable
A synchronous pageable copy of the whole layer into slots
[0, num_experts), position == expert id. Never captured in
a CUDA graph: prefill is not captured, and decode never reaches
this branch because an unpinned layer routes to the CPU
executor.
Trait Implementations§
impl Copy for CopyRoute
impl Eq for CopyRoute
impl StructuralPartialEq for CopyRoute
Auto Trait Implementations§
impl Freeze for CopyRoute
impl RefUnwindSafe for CopyRoute
impl Send for CopyRoute
impl Sync for CopyRoute
impl Unpin for CopyRoute
impl UnsafeUnpin for CopyRoute
impl UnwindSafe for CopyRoute
Blanket Implementations§
Source§impl<T> BorrowMut<T> for Twhere
T: ?Sized,
impl<T> BorrowMut<T> for Twhere
T: ?Sized,
Source§fn borrow_mut(&mut self) -> &mut T
fn borrow_mut(&mut self) -> &mut T
Mutably borrows from an owned value. Read more
Source§impl<T> CloneToUninit for Twhere
T: Clone,
impl<T> CloneToUninit for Twhere
T: Clone,
Source§impl<T> IntoEither for T
impl<T> IntoEither for T
Source§fn into_either(self, into_left: bool) -> Either<Self, Self> ⓘ
fn into_either(self, into_left: bool) -> Either<Self, Self> ⓘ
Converts
self into a Left variant of Either<Self, Self>
if into_left is true.
Converts self into a Right variant of Either<Self, Self>
otherwise. Read moreSource§fn into_either_with<F>(self, into_left: F) -> Either<Self, Self> ⓘ
fn into_either_with<F>(self, into_left: F) -> Either<Self, Self> ⓘ
Converts
self into a Left variant of Either<Self, Self>
if into_left(&self) returns true.
Converts self into a Right variant of Either<Self, Self>
otherwise. Read more