batch_size for cluster calculation

[jaemolihm]

Dear developers,

I'm trying to learn the SSCHA code, and I really appreciate your effort for developing and releasing this program.

I would like to ask a question about the batch_size parameter for a cluster calculation.

In Cluster.py, jobs are submitted until count > self.batch_size, where count is the number of submitted jobs (see code below). Then, in total self.batch_size+1 jobs are submitted.
Is this an expected feature? To make at most self.batch_size jobs run simultaneously, one could change line 1048 of Cluster.py to count >= self.batch_size:.

            count = 0
            # Submit in parallel
            jobs = [false_id[i : i + self.job_number] for i in range(0, len(false_id), self.job_number)]
            
            for job in jobs:
                # Submit only the batch size
                if count > self.batch_size:
                    break
                t = threading.Thread(target = compute_single_jobarray, args=(job, ase_calc, ))
                t.start()
                threads.append(t)
                count += 1

Regards,
Jae-Mo Lihm
Seoul National University

[mesonepigreco]

Hi Jae-Mo Lihm,
You are right. I fix this immediately.
Thanks for the report.
Luckily, it is not a bug that affects the calculation or the results :)
Bests,
Lorenzo