neural-rs is a spiking neural network library for Rust (similar to Brian2 or Nest).
I haven't used it but its supposed to be better for a number of reasons: http://www.scholarpedia.org/article/Adaptive_exponential_integrate-and-fire_model
If nothing else I'm hoping the model won't require any special casing for various firing patterns (see: src/izhikevich/neuron.rs#L86).
Its much faster than my Go implementation by about an order of magnitude but that doesn't mean much.
The slowest part is the synapse update phase because, well, there are usually many more synapses... Maybe time to look at parallelizing on the GPU: https://github.com/luqmana/rust-opencl.