Error while running Flipping routine in qw5q_platinum

[L4VM4T]

Chip: qw5q_platinum
Instruments: QM

I am running the Flipping routine:

  - id: flipping B1
    operation: flipping
    parameters:
      delta_amplitude: 0.0015
      nflips_max: 50
      nflips_step: 1
      nshots: 256
      relaxation_time: 100_000
      unrolling: true
    update: true

and I am getting the following error:

[hay-k]

From my side I can confirm that on qw11q the same runcard runs fine.

[stavros11]

Thanks for reporting this. I tried using my environment and I get the same error after a few successful executions. I will investigate, but in the meantime I noticed that turning unrolling: False works. It may be a bit slower (for these parameters it will take about 7 minutes), but it will at least give you some results until we fix the issue. I would also suggest increasing nshots to at least 1000, because I am not sure if 256 are sufficient for the statistical error.

From my side I can confirm that on qw11q the same runcard runs fine.

Thanks for checking, this is good to know. It is weird because the program generation in the driver is identical for both these instruments, so I would only assume there is some internal difference that is causing the error.

[L4VM4T]

Thanks @stavros11 and @hay-k !

[L4VM4T]

Again, this time I was running a Flipping signal experiment. It worked for several runs, but on the last one, I got another error message related to QM.
This is the Runcard:

I am also attaching the log file:
_qw5q_platinum_latest1.txt

[alecandido]

The compilation timed out part seems a strong indication that the unrolling is a bit too extreme: it may fit the memory of the device, but for some reason the internal QM compiler is taking forever to process it.

Until we dig out more details, if you want to keep using the unrolling, maybe you could try to lower the QM bounds, to make smaller batches.