I compiled on LUMI with both GNU and Cray, and added some performance figures similar to how it was done for Snellius.
To compile with Cray, I had to change one thing related to I/O in the source -- as far as I know the behaviour is the same as the old version but someone with more Fortran experience should check.
At runtime the Cray version still crashed, the check here failed, even though the values of Morton_no and nleafs are in fact equal. To be able to run the test before losing my temporary acess to LUMI, I simply commented out lines 99 to 101, but we should properly fix it later.
I compiled on LUMI with both GNU and Cray, and added some performance figures similar to how it was done for Snellius.
To compile with Cray, I had to change one thing related to I/O in the source -- as far as I know the behaviour is the same as the old version but someone with more Fortran experience should check.
At runtime the Cray version still crashed, the check here failed, even though the values of
Morton_no
andnleafs
are in fact equal. To be able to run the test before losing my temporary acess to LUMI, I simply commented out lines 99 to 101, but we should properly fix it later.