@tico Have a look at our proposal : https://discuss.tvm.ai/t/rfc-byoc-an-extended-graph-partitioning-flow/6028. I think the "optimal" graph partitioning/mapping could be easily plugged in as the "De-conflict" step that we are proposing (Although, we are currently proposing a generic simple greedy one). Moreover, I think finding the "optimal" is non-trivial because it will depend on lots of parameters of the backend. (e.g. supported layout, data movement costs, etc). Thus, our idea is to keep it open, so that a user could plug in a complex de-conflict pass to figure out the final disjoint set of subgraphs.
--- [Visit Topic](https://discuss.tvm.ai/t/relay-improved-graph-partitioning-algorithm/5830/16) to respond. You are receiving this because you enabled mailing list mode. To unsubscribe from these emails, [click here](https://discuss.tvm.ai/email/unsubscribe/2109e99f631a56ce717fdc7d24ac8c0b3ea27f7d28c2398cba3a5acef304133e).
