Struct OneToManyMatmul

pub struct OneToManyMatmul<MP: MatmulPrecision, GMM: GlobalMatmul<MP>, S: SpanMatmul, C: CubeDispatch> { /* private fields */ }

Executes matrix multiplication at the batch level, assigning each cube to handle multiple global matmuls.

The algorithm supports any number of cubes, looping as needed to process all data.

Trait Implementations

impl<MP: MatmulPrecision, GMM: GlobalMatmul<MP>, S: SpanMatmul, C: CubeDispatch> BatchMatmul<MP> for OneToManyMatmul<MP, GMM, S, C>

type Config = Config<<GMM as GlobalMatmul<MP>>::Config, C>

fn execute( lhs: VirtualTensor<MP::EI>, rhs: VirtualTensor<MP::EI>, out: VirtualTensor<MP::EO, ReadWrite>, _size_k: u32, quantization: CubeOption<Quantization<MP>>, config: Self::Config, )

Performs batchwise matrix multiplication over tensors.

fn __expand_execute( context: &mut Scope, lhs: <VirtualTensor<MP::EI> as CubeType>::ExpandType, rhs: <VirtualTensor<MP::EI> as CubeType>::ExpandType, out: <VirtualTensor<MP::EO, ReadWrite> as CubeType>::ExpandType, _size_k: <u32 as CubeType>::ExpandType, quantization: <CubeOption<Quantization<MP>> as CubeType>::ExpandType, config: Self::Config, ) -> <() as CubeType>::ExpandType