[FEAT.] BrainCog examples的跨平台部署 Cross-platform deployment of BrainCog example

[yuxuan-z19]

我正在对类脑应用进行 负载建模 (workload characterization)，选用 BrainCog 提供的 examples 作为应用集。

目前我通过A800的profile作为硬件资源的验证，但是仅在GPGPU平台上仿真不能很好地代表实际地类脑硬件。

我注意到 BrainCog团队 去年发布了 基于systolic array的加速器 Firefly v1, v2，想请问该如何 自动部署 BrainCog examples？同时，我注意到 SpikingJelly 工作支持 lynxi HP300芯片 和 Loihi转换部署，请问 BrainCog 是否有后续支持此类部署的意愿？

为了研究严谨性，我不希望手动迁移到目标硬件上，这样得出的结果不够客观。

---

I am conducting workload characterization for brain-inspired applications and have selected BrainCog's examples as my application suite.

I am currently using profiling results from the Ascend A800 to validate hardware resource usage. However, simulation solely on GPGPU platforms cannot fully reflect the behavior of actual neuromorphic hardware.

I noticed that the BrainCog team released Firefly v1 and v2, systolic array-based accelerators, last year. May I ask how one could automatically deploy BrainCog examples onto these platforms?

Additionally, I noticed that SpikingJelly supports deployment on hardware like the Lynxi HP300 chip and even conversion for Loihi deployment. Does the BrainCog team have any plans to support such deployments in the future?

For the sake of research rigor, I would prefer not to manually migrate code to the target hardware, as such results may lack objectivity.