Package index • amatrix

All functions

KronMatrix-class: Lazy Kronecker product of two matrices
aMatrix-class: Virtual base class for backend-aware matrices
aTransposeView-class: Lazy transpose view of an adgeMatrix
addmm(): Scaled matrix multiply with optional bias: alpha*(A%*%B) + beta*C
adgCMatrix-class: Sparse column-compressed matrix with backend-dispatch metadata
adgCMatrix(): Create a backend-aware sparse matrix
adgeMatrix-class: Dense general matrix with backend-dispatch metadata
adgeMatrix(): Create a backend-aware dense matrix
adlgCMatrix-class: Sparse logical matrix with backend-dispatch metadata
adlgeMatrix-class: Dense logical matrix with backend-dispatch metadata
amChol-class: Cholesky factorization result
amLU-class: LU factorization result for general square matrices
amSVD-class: Truncated SVD factorization result
am_rowargmax() am_rowargmin() am_colargmax() am_colargmin(): Row and column argmax/argmin
am_ewise_inplace(): In-place elementwise operation on a resident handle
am_qr(): QR decomposition of an amatrix object
am_scatter_mean(): Scatter mean by group labels
am_sweep(): Backend-dispatched sweep
am_sweep_inplace(): In-place broadcast sweep on a resident handle
amatrix_backend_capabilities(): Query the capabilities of a registered backend
amatrix_backend_features(): Query the features of a registered backend
amatrix_backend_health_probe(): Run a canary health probe against a registered backend
amatrix_backend_matrix(): Tabulate dispatch plans across multiple operations
amatrix_backend_names(): List names of all registered backends
amatrix_backend_plan(): Compute the dispatch plan for a single operation
amatrix_backend_precision_modes(): Query the precision modes supported by a registered backend
amatrix_backend_status(): Summarise the status of registered backends
amatrix_benchmark_report(): Report amatrix benchmark status across ops and backends
amatrix_bind_resident(): Bind an amatrix object to resident backend storage
amatrix_cache_max_size() amatrix_set_cache_max_size(): Get or set the model cache maximum size
amatrix_calibrate(): Calibrate GPU dispatch thresholds for this machine
amatrix_calibration_info(): Retrieve the current calibration state
amatrix_compile_product(): Compile a reusable matrix-product plan
amatrix_default_policy(): Get the session-level default dispatch policy
amatrix_default_precision(): Get the session-level default precision mode
amatrix_dispatch_op(): Low-level backend dispatch for a single operation
amatrix_execution_info(): Collect full dispatch information for an aMatrix object
amatrix_explain(): Explain dispatch decisions for an aMatrix operation
amatrix_fallback_log(): Return the amatrix backend fallback log
amatrix_fallback_log_reset(): Clear the amatrix backend fallback log
amatrix_gc(): Free stale GPU residency entries and optionally flush the model cache
amatrix_gpu_status(): GPU backend status: why am I (not) on the GPU?
amatrix_materialize_host(): Force materialization of an aMatrix to a host Matrix object
amatrix_memory_stats(): Report GPU residency and model cache usage
amatrix_prepare_operands(): Prepare operands for a repeated matrix product
amatrix_register_backend(): Register a backend with the amatrix dispatch system
amatrix_release_resident(): Release GPU-resident data held by an amatrix object
amatrix_residency_info(): Query GPU residency state of an aMatrix object
amatrix_resident_backend_for(): Choose a residency-capable accelerator backend for a hot path
amatrix_set_default_policy(): Set the session-level default dispatch policy
amatrix_set_default_precision(): Set the session-level default precision mode
amatrix_use_gpu(): Enable GPU acceleration for this session
amatrix_warm(): Warm up GPU backends to eliminate cold-start latency
array_lm(): Fit linear models with array-shaped response
as_adgCMatrix(): Coerce an object to adgCMatrix
as_adgeMatrix(): Coerce an object to adgeMatrix
as_adgeMatrix.resident_handle(): Convert a resident handle back to an adgeMatrix
batch_chol(): Batch Cholesky factorization
batch_crossprod(): Batch crossproduct
batch_solve(): Batch triangular solve
block_lanczos() block_svd(): Block Lanczos SVD via block Krylov iteration
chol(<adgeMatrix>): Cholesky factorization for adgeMatrix
chol(<adgCMatrix>): Cholesky factorization for adgCMatrix
chol_diag(): Extract the diagonal of a Cholesky factor
chol_factor(): Compute the Cholesky factorization of an adgeMatrix
chol_logdet(): Log-determinant from a Cholesky factor
chol_solve(): Solve a linear system using a Cholesky factor
chol_solve_batches(): Solve many right-hand-side batches with one Cholesky factor
as.matrix(<adgeMatrix>) as.matrix(<adgCMatrix>) as.matrix(<aTransposeView>) as.matrix(<amChol>) as.matrix(<KronMatrix>) as.numeric(<adgeMatrix>) as.vector(<adgeMatrix>) as.array(<adgeMatrix>) as.array(<adgCMatrix>): Coerce amatrix objects to base R types
correlation(): Compute a correlation matrix
cov2cor(<adgeMatrix>) cov2cor(<adgCMatrix>): Covariance-to-correlation methods for amatrix objects
covariance(): Backend-dispatched covariance matrix
crossprod(<adgeMatrix>,<ANY>) crossprod(<adgeMatrix>,<missing>) tcrossprod(<adgeMatrix>,<ANY>) tcrossprod(<adgeMatrix>,<missing>): Cross-product methods for adgeMatrix
crossprod(<adgCMatrix>,<missing>) crossprod(<adgCMatrix>,<ANY>) crossprod(<adgCMatrix>,<matrix>) crossprod(<adgCMatrix>,<Matrix>) crossprod(<adgCMatrix>,<dgeMatrix>) crossprod(<adgCMatrix>,<dgCMatrix>) crossprod(<adgCMatrix>,<adgeMatrix>) crossprod(<adgCMatrix>,<adgCMatrix>) tcrossprod(<adgCMatrix>,<missing>) tcrossprod(<adgCMatrix>,<ANY>) tcrossprod(<adgCMatrix>,<matrix>) tcrossprod(<adgCMatrix>,<Matrix>) tcrossprod(<adgCMatrix>,<dgeMatrix>) tcrossprod(<adgCMatrix>,<dgCMatrix>) tcrossprod(<adgCMatrix>,<adgeMatrix>) tcrossprod(<adgCMatrix>,<adgCMatrix>): Cross-product methods for adgCMatrix
crossprod_add_diag(): Cross-product plus diagonal perturbation
crossprod_weighted(): Weighted cross-product X'WX
dist_matrix(): GPU-accelerated pairwise distance matrix
dot(): Inner product of two vectors or matrices
eigen(<adgeMatrix>): Eigendecomposition for adgeMatrix
eigen(<adgCMatrix>): Eigendecomposition for adgCMatrix
eigh(): Symmetric eigendecomposition
ewise(): Element-wise operations
gemm(): Generalised matrix multiply (BLAS DGEMM interface)
irlba(): GPU-accelerated truncated SVD via irlba
irlba_native(): GPU-native truncated SVD via Lanczos bidiagonalization
kernel_matrix(): GPU-accelerated pairwise kernel matrix
kron(): Eager Kronecker product
kron_matrix(): Construct a lazy Kronecker product
kronecker(<adgeMatrix>,<adgeMatrix>) kronecker(<adgeMatrix>,<adgCMatrix>) kronecker(<adgCMatrix>,<adgeMatrix>) kronecker(<adgCMatrix>,<adgCMatrix>) kronecker(<adgeMatrix>,<matrix>) kronecker(<matrix>,<adgeMatrix>) kronecker(<adgCMatrix>,<matrix>) kronecker(<matrix>,<adgCMatrix>): Kronecker product of backend-aware matrices
lm_fit(): Fit a single linear model
lm_loo_cv(): Leave-one-out cross-validation for linear models
lu_factor(): Store a general square matrix for LU-based solving
lu_solve(): Solve a linear system using an LU factor
many_lm(): Fit multiple linear models against a shared design matrix
mat_sqrt() mat_pow() mat_log(): Matrix functions via symmetric eigendecomposition
`%*%`(<adgeMatrix>,<matrix>) `%*%`(<adgeMatrix>,<Matrix>) `%*%`(<adgeMatrix>,<dgeMatrix>) `%*%`(<adgeMatrix>,<dgCMatrix>) `%*%`(<adgeMatrix>,<adgeMatrix>) `%*%`(<adgeMatrix>,<adgCMatrix>) `%*%`(<numeric>,<adgeMatrix>): Matrix multiplication for adgeMatrix
`%*%`(<adgCMatrix>,<ANY>) `%*%`(<adgCMatrix>,<matrix>) `%*%`(<adgCMatrix>,<Matrix>) `%*%`(<adgCMatrix>,<dgeMatrix>) `%*%`(<adgCMatrix>,<dgCMatrix>) `%*%`(<adgCMatrix>,<adgeMatrix>) `%*%`(<adgCMatrix>,<adgCMatrix>): Matrix multiplication for adgCMatrix
matmul(): Matrix multiplication
pairwise_sqdist_argmin(): Nearest-centroid assignment via fused squared-distance computation
pca_coef(): Project and reconstruct data using a truncated SVD
qr_downdate(): QR downdate after removing one row
qr_info(): Inspect an amQR factorization object
quad_form(): Evaluate a quadratic form using a Cholesky factor
resident_handle(): Create a mutable GPU-resident handle
rh_colSums(): Column sums of a GPU-resident handle
rh_rowSums(): Row sums of a GPU-resident handle
ridge_fit(): Fit a single ridge regression model
ridge_path(): Compute a ridge regression solution path
rowSums(<adgeMatrix>) colSums(<adgeMatrix>) rowMeans(<adgeMatrix>) colMeans(<adgeMatrix>): Row and column summary methods for adgeMatrix
rowSums(<adgCMatrix>) colSums(<adgCMatrix>) rowMeans(<adgCMatrix>) colMeans(<adgCMatrix>): Row and column summary methods for adgCMatrix
rowmeans() colmeans(): Row and column means
rowscale() colscale(): Row and column diagonal scaling
rowsums() colsums(): Row and column sums
rsvd(): GPU-native randomized SVD (Halko et al. 2011)
segment_mean(): Segment mean by group labels
segment_sum(): Segment sum by group labels
sinkhorn(): Doubly-stochastic scaling via Sinkhorn-Knopp iterations
solve(<adgeMatrix>,<missing>) solve(<adgeMatrix>,<ANY>): Solve a linear system for adgeMatrix
solve(<adgCMatrix>,<missing>) solve(<adgCMatrix>,<ANY>): Solve a linear system for adgCMatrix
solve_triangular(): Solve a triangular linear system
svd(<adgeMatrix>): Singular value decomposition for adgeMatrix
svd(<adgCMatrix>): Singular value decomposition for adgCMatrix
svd_factor(): Compute a truncated SVD of an aMatrix
svd_project(): Project new data onto SVD left singular vectors
svd_reconstruct(): Reconstruct data from SVD coordinates
sym(): Symmetrise a matrix
tcrossprod_weighted(): Weighted outer cross-product XWX'
trace(): Matrix trace
trace_estim(): Stochastic trace estimator (Hutchinson)
with_amatrix(): Evaluate code with temporary amatrix defaults
wls_fit(): Fit a weighted least squares model
woodbury_logdet(): Log-determinant via the Woodbury matrix determinant lemma
woodbury_solve(): Solve a linear system using the Woodbury matrix identity
xty_weighted(): Weighted cross-product X'Wy