Assigning memory usage to sorters

Hello,

I am trying to run several spike sorters using spikeinterface. I assign the number of jobs and memory using
global_job_kwargs = dict(n_jobs=12, total_memory='180G')
si.set_global_job_kwargs(**global_job_kwargs)

When I try to run spikesorting using spykingcircus2, using
sorting_spykingcircus2 = ss.run_sorter(sorter_name="spykingcircus2", recording=rec_w,
output_folder="2021-11-29/spkinterface/spykingcircus2",
filtering={},
apply_preprocessing = False,
job_kwargs = {'n_jobs': 12, 'total_memory':'180G'},
detection = {'peak_sign': 'neg', 'detect_threshold': 6},
selection = {"n_peaks_per_channel": 10000,
"min_n_peaks": 20000},
)
I get the error: numpy.core._exceptions._ArrayMemoryError: Unable to allocate 392. GiB for an array with shape (5, 231, 91107520) and data type float32

I tried including matching = {"method": "circus-omp-svd"} within the sorter arguments, but the error persists.
Is there a way to restrict the memory used by spykingcircus or adjust the chunk size? I am using spikeinterface version: 0.99.1.

Start at the si.set_global_job_kwargs and ss.run_sorter entry points, then trace the spykingcircus2 sorter handling for job_kwargs, total_memory, and chunk sizing. Reproduce the reported allocation error with the shown parameters and determine whether the memory limit is honored. Done means the memory behavior is explained and a verified fix or documented limitation is covered by the relevant tests.