The problem of `affine.parallel`

[EllisLambda]

Is your feature request related to a problem? Please describe.
Due to the difficulties in using affine.parallel, the outermost loop in the BatchMatMulOptimize of my PR does not utilize affine.parallel. However, this is crucial for the upcoming attempts to parallelize matmul using OpenMP. There's no open-source information available. I've tried writing code following the template, but it always results in errors

Describe the solution you'd like
Rewrite the outermost affine.for loop to affine.parallel

[xlinsist]

Maybe we should follow this topic. Could you provide some code demos to further illustrate the technical background and difficulties?

BTW Here are some suggestions when creating an issue that you may follow:

1. You can mention your PR by directly copying the link(whether it is from this repository or not) like this: [OpOptimization] Add BatchMatMul benchmark and [OpOptimization] Further optimize BatchMatMulBroadcast and add OpenMP tests buddy-benchmark#73.
2. When you are seeking for a solution, try to avoid using ambiguous or unfocused naming of the title, make it spefific but not overly technical.

[EllisLambda]

The problem have been solved. Thanks for the advice!