Incorrect filling of computation time into the msg_size of workloads in AIOB

[dageita]

When running benchmark suites ./aicb.py, a csv workload will be generated.
The code for filling the generated computation time into the workload is as follows, which appears to be filled in the msg_size field of workload:

def Comp_with_aiob(workload, compute_cache):
    for item in workload.workload:
        if item.comm_type == CommType.computation:
            for key in compute_cache:
                key_temp = key.split("_")[0]
                if key_temp in item.stage:
                    item.msg_size = compute_cache[key]
                    break
    return workload

However, compute_cache is the extracted operator time, not msg_size, which seems contradictory:

workload = Comp_with_aiob(workload, compute_cache)

[Huoyuan100861]

When running benchmark suites ./aicb.py, a csv workload will be generated. The code for filling the generated computation time into the workload is as follows, which appears to be filled in the msg_size field of workload:

def Comp_with_aiob(workload, compute_cache):
    for item in workload.workload:
        if item.comm_type == CommType.computation:
            for key in compute_cache:
                key_temp = key.split("_")[0]
                if key_temp in item.stage:
                    item.msg_size = compute_cache[key]
                    break
    return workload

However, compute_cache is the extracted operator time, not msg_size, which seems contradictory:

workload = Comp_with_aiob(workload, compute_cache)

You've noticed a subtle detail indeed. There is a naming ambiguity where compute_cache actually refers to the operator execution time, not msg_size, which might be confusing.
Fortunately, this hasn't currently affected the overall functionality. We are reviewing the pull request you submitted.