#[repr(u32)]pub enum CuTensorMapDataType {
Show 13 variants
Uint8 = 0,
Uint16 = 1,
Uint32 = 2,
Int32 = 3,
Uint64 = 4,
Int64 = 5,
Float16 = 6,
Float32 = 7,
Float64 = 8,
Bfloat16 = 9,
Float32Ftz = 10,
TF32 = 11,
TF32Ftz = 12,
}Expand description
Element data type for the TMA descriptor.
Corresponds to CUtensorMapDataType in the CUDA Driver API.
Variants§
Uint8 = 0
Unsigned 8-bit integer.
Uint16 = 1
Unsigned 16-bit integer.
Uint32 = 2
Unsigned 32-bit integer.
Int32 = 3
Signed 32-bit integer.
Uint64 = 4
Unsigned 64-bit integer.
Int64 = 5
Signed 64-bit integer.
Float16 = 6
IEEE-754 half-precision float (f16).
Float32 = 7
IEEE-754 single-precision float (f32).
Float64 = 8
IEEE-754 double-precision float (f64).
Bfloat16 = 9
Brain float 16 (bfloat16).
Float32Ftz = 10
f32 with flush-to-zero for subnormals.
TF32 = 11
TensorFloat-32 (TF32).
TF32Ftz = 12
TF32 with flush-to-zero.
Implementations§
Source§impl CuTensorMapDataType
impl CuTensorMapDataType
Sourcepub const fn element_size_bytes(self) -> u32
pub const fn element_size_bytes(self) -> u32
Returns the element size in bytes.
Trait Implementations§
Source§impl Clone for CuTensorMapDataType
impl Clone for CuTensorMapDataType
Source§fn clone(&self) -> CuTensorMapDataType
fn clone(&self) -> CuTensorMapDataType
Returns a duplicate of the value. Read more
1.0.0 · Source§fn clone_from(&mut self, source: &Self)
fn clone_from(&mut self, source: &Self)
Performs copy-assignment from
source. Read moreSource§impl Debug for CuTensorMapDataType
impl Debug for CuTensorMapDataType
Source§impl Hash for CuTensorMapDataType
impl Hash for CuTensorMapDataType
Source§impl PartialEq for CuTensorMapDataType
impl PartialEq for CuTensorMapDataType
impl Copy for CuTensorMapDataType
impl Eq for CuTensorMapDataType
impl StructuralPartialEq for CuTensorMapDataType
Auto Trait Implementations§
impl Freeze for CuTensorMapDataType
impl RefUnwindSafe for CuTensorMapDataType
impl Send for CuTensorMapDataType
impl Sync for CuTensorMapDataType
impl Unpin for CuTensorMapDataType
impl UnsafeUnpin for CuTensorMapDataType
impl UnwindSafe for CuTensorMapDataType
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