A Persistent-State Dataflow Accelerator for Memory-Bound Linear Attention Decode on FPGA
This work provides a significant performance and energy efficiency improvement for memory-bound linear attention decode in hybrid LLMs, particularly for edge or power-constrained applications.
This paper addresses the memory-bound nature of Gated DeltaNet (GDN) linear attention decode on GPUs by presenting an FPGA accelerator. By persistently storing the 2 MB recurrent state on-chip, the accelerator converts the workload to compute-bound, achieving 63 μs per token, which is 4.5× faster than an NVIDIA H100 PCIe GPU and up to 60× more energy-efficient.
Gated DeltaNet (GDN) is a linear attention mechanism that replaces the growing KV cache with a fixed-size recurrent state. Hybrid LLMs like Qwen3-Next use 75% GDN layers and achieve competitive accuracy to attention-only models. However, at batch-1, GDN decode is memory-bound on GPUs since the full recurrent state must be round-tripped through HBM every token. We show that this bottleneck is architectural, not algorithmic, as all subquadratic sequence models exhibit arithmetic intensities below 1 FLOP/B at decode time, making them more memory-bound than standard Transformers. We present an FPGA accelerator that eliminates this bottleneck by holding the full 2 MB recurrent state persistently in on-chip BRAM, converting the workload from memory-bound to compute-bound. Our design fuses the GDN recurrence into a five-phase pipelined datapath that performs only one read and one write pass over each state matrix per token, exploits Grouped Value Attention for paired-head parallelism, and overlaps preparation, computation, and output storage via dataflow pipelining. We explore four design points on an AMD Alveo U55C using Vitis HLS, varying head-level parallelism from 2 to 16 value-heads per iteration. Our fastest configuration achieves 63 $μ$s per token, 4.5$\times$ faster than the GPU reference on NVIDIA H100 PCIe. Post-implementation power analysis reports 9.96 W on-chip, yielding up to 60$\times$ greater energy efficiency per token decoded.