Hi,
I have applied 39 different Matminer featurizers to a dataset consisting of 25126 entries from Materials Project (bandage > 0.1 and ICSD-entry base query). During this process, I found a high memory allocation for the following MPIDs:
“mp-555563”, #PH6C2S2NCl2O4 #DOI: 10.17188/1268877
“mp-583476”, #Nb7S2I19 #DOI: 10.17188/1277059
“mp-600205”, #H10C5SeS2N3Cl #DOI: -
“mp-600217”, #H80C40Se8S16Br8N24 #DOI: -
“mp-1195290”, #Ga3Si5P10H36C12N4Cl11 #DOI: -
“mp-1196358”, #P4H120Pt8C40I8N4Cl8 #DOI: -
“mp-1196439”, #Sn8P4H128C44N12Cl8O4 #DOI: -
“mp-1198652”, #Te4H72C36S24N12Cl4 #DOI: -
“mp-1198926”, #Re8H96C24S24N48Cl48 #DOI: -
“mp-1199490”, #Mn4H64C16S16N32Cl8 #DOI: -
“mp-1199686”, #Mo4P16H152C52N16Cl16 #DOI: -
“mp-1203403”, #C121S2Cl20 #DOI: -
“mp-1204279”, #Si16Te8H176Pd8C64Cl16 #DOI: -
“mp-1204629” #P16H216C80N32Cl8 #DOI: -
It is unknown to me what the cause of the problem is, but since I am running out of memory with 16GB available and doing featurization of one entry at the time, I believe that there is either a memory leak or a bug in the pymatgen structure object.
Best regards,
Oliver