XC.meta doesn't exist in output

[yz3062]

I'm trying to read the run results of the MIT-gcm tutorial example "verification/tutorial_global_oce_biogeo/" (no modification). When I tried to use open_mdsdataset(), it returned the error:
OSError: Couldn't find XC.meta file to infer nx and ny.

Turns out there is not XC.meta file in the output folder. Instead, there are four separate files named "XC.001.001.meta", "XC.001.002.meta", "XC.002.001.meta", and "XC.002.002.meta". I paste below the content of the files. Is there a way to work around the issue and read the dimensions correctly?

"XC.001.001.meta"
simulation = { 'Tutorial Biogeo' }; nDims = [ 2 ]; dimList = [ 128, 1, 64, 64, 1, 32 ]; dataprec = [ 'float32' ]; nrecords = [ 1 ];

"XC.001.002.meta"
simulation = { 'Tutorial Biogeo' }; nDims = [ 2 ]; dimList = [ 128, 1, 64, 64, 33, 64 ]; dataprec = [ 'float32' ]; nrecords = [ 1 ];

"XC.002.001.meta"
simulation = { 'Tutorial Biogeo' }; nDims = [ 2 ]; dimList = [ 128, 65, 128, 64, 1, 32 ]; dataprec = [ 'float32' ]; nrecords = [ 1 ];

"XC.002.002.meta"
simulation = { 'Tutorial Biogeo' }; nDims = [ 2 ]; dimList = [ 128, 65, 128, 64, 33, 64 ]; dataprec = [ 'float32' ]; nrecords = [ 1 ];

[rabernat]

The problem is that your tutorial was run with multi-tile output. Xmitgcm can't read those files (see #28). Try rerunning your simulation with singlecpuio=.TRUE. in data.

[yz3062]

Thanks Ryan. I tried globalFiles=.TRUE., which should be equivalent to your solution, and that solves the problem.