QPU error in remote_job.is_complete

[Jambi]

Hi there,

I have recently gotten this error when running jobs on QPU Ascella:

I’m following the tutorial and created a while loop to wait for the job to finish before retrieving the data, but it seems like it is too fast (??) so the complete_time is not an integer or None.

Is there a fix for this?

I also just started getting this warning today:
[warning] DeprecationWarning: Call to deprecated function (or staticmethod) thresholded_output. (Add detectors as components) – Deprecated since version 0.12
It must be raised by one of the QPU lines below (I assume the min_detected_photons_filter, self.N is 3 in this case), any suggestion on what to do about it? I know it’s only a warning but I’d just like it clean if possible.

qpu_platform = "qpu:ascella"
QPU = pcvl.RemoteProcessor(qpu_platform)
QPU.set_circuit(self.chip)
QPU.with_input(input_state)
QPU.min_detected_photons_filter(self.N)

Thanks
Markus

[Jambi]

FYI, I’m currently doing this as a hack around it. Works, just not pretty:

while True:
    try:
        time.sleep(1)
        sample_count = remote_job.get_results()
        break 
    except:
        pass

[Aubaert]

Hello Markus,

It seems like the cloud is not sending the starting time back when asked to. We will make our code more robust against this. In the meanwhile, you should keep your workaround.

For the deprecation warning, can you update perceval? In your given example, there is no reason in the current version of the code that you get this error. The only way to get it without explicitly using thresholded_output is by using estimate_required_shots or estimate_expected_samples from the RemoteProcessor. This will be fixed on the next version.

Regards,
Aurélien

[Jambi]

Hi Aurélien,

Thanks for looking into the remote job issue I am having.

About the deprecation: I am indeed doing the shot estimation first, looking forward to the fix

Thanks
Markus