A fused exponential-multiplication hardware unit using logarithmic quantization and exponent adjustment reduces FlashAttention accelerator area by about 29% and power by about 18% without visible accuracy loss on GLUE.
A3: Accelerating attention mechanisms in neural net- works with approximation,
1 Pith paper cite this work. Polarity classification is still indexing.
1
Pith paper citing it
citation-role summary
background 1
citation-polarity summary
fields
cs.AR 1years
2025 1verdicts
CONDITIONAL 1roles
background 1polarities
unclear 1representative citing papers
citing papers explorer
-
Low-Cost FlashAttention with Fused Exponential and Multiplication Hardware Operators
A fused exponential-multiplication hardware unit using logarithmic quantization and exponent adjustment reduces FlashAttention accelerator area by about 29% and power by about 18% without visible accuracy loss on GLUE.