Struct GpuBLAS 

pub struct GpuBLAS<F: Float + Debug> { /* private fields */ }

GPU-accelerated BLAS operations

Implementations

impl<F: Float + Debug + Clone> GpuBLAS<F>

pub fn new(config: GpuConfig) -> Self

Create new GPU BLAS processor

pub fn dot(&self, x: &Array1<F>, y: &Array1<F>) -> Result<F>

GPU-accelerated vector dot product (BLAS Level 1)

pub fn norm(&self, x: &Array1<F>) -> Result<F>

GPU-accelerated vector norm (BLAS Level 1)

pub fn axpy(&self, alpha: F, x: &Array1<F>, y: &mut Array1<F>) -> Result<()>

GPU-accelerated SAXPY: y = alpha * x + y (BLAS Level 1)

pub fn gemv( &self, alpha: F, a: &Array2<F>, x: &Array1<F>, beta: F, y: &mut Array1<F>, ) -> Result<()>

GPU-accelerated matrix-vector multiplication (BLAS Level 2)

pub fn gemm( &self, alpha: F, a: &Array2<F>, b: &Array2<F>, beta: F, c: &mut Array2<F>, ) -> Result<()>

GPU-accelerated matrix-matrix multiplication (BLAS Level 3)

pub fn transpose(&self, a: &Array2<F>) -> Array2<F>

GPU-accelerated matrix transpose

pub fn batch_gemm( &self, alpha: F, a_batch: &[Array2<F>], b_batch: &[Array2<F>], beta: F, c_batch: &mut [Array2<F>], ) -> Result<()>

GPU-accelerated batch matrix operations

Trait Implementations

impl<F: Debug + Float + Debug> Debug for GpuBLAS<F>

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

Formats the value using the given formatter.