Jeff Fifield

I've hit this in the past. I think the solution is to introduce a custom terminator instead of using `cf.br`

This should be fixed by https://github.com/Xilinx/mlir-aie/pull/247

> hawkpoint is running the wheels instead is there a disadvantage to having a bigger pool that can do all the jobs?

I was able to reproduce the issue outside of IREE using [input.mlir](https://gist.github.com/Abhishek-Varma/cd54f0b20c24ce237288a9933139def3#file-input-mlir) and a c++ testbench, but after looking closer I think there is no bug. I think the problem...

@eddierichter-amd the duplicates to resolve w.r.t. vck5000 after merge are ``` programming_examples/basic/passthrough_hardware, programming_examples/basic/passthrough_hardware2 programming_examples/basic/vector_scalar, programming_examples/basic/vector_scalar2 ```

added in https://github.com/Xilinx/mlir-aie/pull/1112

I don't own a mac, so I can't address this issue. If someone fixes it, please send me a pull request.