Mambalaya delivers 4.9x prefill and 1.9x generation speedups on Mamba layers over prior accelerators by systematically fusing inter-Einsum operations.
Timeloop: A systematic approach to DNN accelerator evaluation
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
fields
cs.AR 1years
2026 1verdicts
UNVERDICTED 1representative citing papers
citing papers explorer
-
Mambalaya: Einsum-Based Fusion Optimizations on State-Space Models
Mambalaya delivers 4.9x prefill and 1.9x generation speedups on Mamba layers over prior accelerators by systematically fusing inter-Einsum operations.