When Spike Sparsity Does Not Translate to Deployed Cost: VS-WNO on Jetson Orin Nano
For researchers deploying spiking neural networks on commodity edge GPUs, this paper shows that algorithmic sparsity does not guarantee efficiency gains due to runtime limitations.
The paper investigates whether spike sparsity in spiking neural operators reduces deployed cost on edge GPUs. On a Jetson Orin Nano, VS-WNO showed high algorithmic sparsity but higher latency (59.6 ms vs 53.2 ms) and energy (228.0 mJ vs 180.7 mJ) than dense WNO, because the runtime does not exploit sparsity.
Spiking neural operators are appealing for neuromorphic edge computing because event-driven substrates can, in principle, translate sparse activity into lower latency and energy. Whether that advantage survives deployment on commodity edge-GPU software stacks, however, remains unclear. We study this question on a Jetson Orin Nano 8 GB using five pretrained variable-spiking wavelet neural operator (VS-WNO) checkpoints and five matched dense wavelet neural operator (WNO) checkpoints on the Darcy rectangular benchmark. On a reference-aligned path, VS-WNO exhibits substantial algorithmic sparsity, with mean spike rates decreasing from 54.26% at the first spiking layer to 18.15% at the fourth. On a deployment-style request path, however, this sparsity does not reduce deployed cost: VS-WNO reaches 59.6 ms latency and 228.0 mJ dynamic energy per inference, whereas dense WNO reaches 53.2 ms and 180.7 mJ, while also achieving slightly lower reference-path error (1.77% versus 1.81%). Nsight Systems indicates that the request path remains launch-dominated and dense rather than sparsity-aware: for VS-WNO, cudaLaunchKernel accounts for 81.6% of CUDA API time within the latency window, and dense convolution kernels account for 53.8% of GPU kernel time; dense WNO shows the same pattern. On this Jetson-class GPU stack, spike sparsity is measurable but does not reduce deployed cost because the runtime does not suppress dense work as spike activity decreases.