Search papers, labs, and topics across Lattice.
1
0
2
HSMA-TRSM achieves over 2x speedup on key GPU platforms by optimizing shared memory usage for complex triangular solves, revealing untapped performance potential in BLAS operations.
