trnsparse 0.4.1

Sparse matrix operations for AWS Trainium via NKI

trnsparse

Sparse matrix operations for AWS Trainium via NKI.

CSR/COO formats, SpMV, SpMM, and integral screening for sparse scientific computing on Trainium. Part of the trnsci scientific computing suite (github.com/trnsci).

Current phase

trnsparse follows the trnsci 5-phase roadmap. Active work is tracked in phase-labeled GitHub issues:

• Phase 1 — correctness ✅ v0.2.0: NKI SpMM validated on trn1 via densify-then-GEMM; first torch.autograd.Function-wrapped NKI kernel in the suite (see trnsci/trnsci#3). Benchmarks in docs/benchmarks.md.
• Phase 3 — perf: nnz-bucketing SpMM, streaming large-sparse, NEFF cache reuse.
• Phase 4 — multi-chip: sharded sparse matrices across chips.
• Phase 5 — generation: trn2 DMA bandwidth exploitation.

(No Phase 2 for trnsparse — the precision story is inherited from trnblas.)

Suite-wide tracker: trnsci/trnsci#1.

Install

pip install trnsparse

Usage

import torch
import trnsparse

# Dense → sparse
A = torch.randn(100, 100)
A[torch.abs(A) < 1.0] = 0.0
csr = trnsparse.from_dense(A)

# SpMV: y = A @ x
y = trnsparse.spmv(csr, x, alpha=2.0)

# SpMM: C = A @ B
C = trnsparse.spmm(csr, B)

# Integral screening
Q = trnsparse.schwarz_bounds(diagonal_integrals)
mask = trnsparse.screen_quartets(Q, threshold=1e-10)
stats = trnsparse.sparsity_stats(Q)

Operations

Operation	Description
spmv	Sparse × dense vector
spmm	Sparse × dense matrix
spmv_symmetric	Symmetric SpMV (half storage)
sparse_add	C = αA + βB
sparse_scale	B = αA
sparse_transpose	A^T
schwarz_bounds	Schwarz screening bounds
screen_quartets	Shell quartet significance mask
density_screen	Density-weighted screening

License

Apache 2.0 — Copyright 2026 Scott Friedman