Issue using chipnetwork with some workloads

[SyedAli1232]

In your artifact repository there are two networks in the interchipnetwork folder: 1. chipnetwork and 2. Interchip network. Your artifact used the interchip network by default. When I try to switch to the chipnetwork some workloads like mis hault after a few instructions. I also tried using the pcie network to connect the chiplets and the same issue arises for a few workloads.
Can you please help me identify the issue here?

[PratheekB]

Hello,

One potential issue I can think of right now is the possibility of deadlocking due to different kinds of packets.
We implemented a form of virtual channel with the interchip network to avoid such deadlocks.
I recommend checking newer versions of MGPUSim (v3), which have much better support for virtual channels in their network stack.