Parallel computing and An installation problem (MpiRun） about underworld2 HOT 4 CLOSED

Hi, julesghub
Sorry, there was a problem with the previous file.
LOG_FILE_1.txt

However, I think I have located the issue.
Please ensure Petsc is using version 3.19.4 and the Setuptools version is less than 65.
GCC 13 results in errors during compilation.
（Although, I don't know which error is preventing pip from installing.）

Hopefully, this will help anyone experiencing the same problem

Cheers

from underworld2.

I have recently configured a new environment that includes Python 3.10.13 and Petsc 3.19.4. However, I encountered an issue while trying to install it using the recommended method, which is the command pip install -v git+https://github.com/underworldcode/underworld2. I have attached the installation log for reference. Errors are different from the previous ones.
LOG_FILE_1.log

from underworld2.

Hi @tyszwh,
Yes a newer version of petsc is required, compared to the previous attempt you made. 3.16 is too old for the current version of UW2.

I see this in the error log.

2023-11-30T19:57:39,393   Running command git clone --filter=blob:none https://github.com/underworldcode/underworld2 /tmp/pip-req-build-1ltcspyg
2023-11-30T19:57:39,396   Cloning into '/tmp/pip-req-build-1ltcspyg'...
2023-11-30T19:57:43,693 ERROR: Operation cancelled by user

Your download from git appears to have been cancelled. Any ideas why?

from underworld2.

Thanks for the points @tyszwh. Glad you got to the bottom of it. I'll close the ticket for now.

PS: The compile error in the log file is

2023-12-10T23:36:53,280   /home/tysz/miniconda3/envs/geo3/include/petscstring.h:688:249: error: expected ')' before 'PRIxPTR'
2023-12-10T23:36:53,280     688 |   PetscAssert(!(((al > bl) && (al - bl) < n) || (bl - al) < n), PETSC_COMM_SELF, PETSC_ERR_ARG_INCOMP, "Memory regions overlap: either use PetscMemmove()\nor make sure your copy regions and lengths are correct.\nLength (bytes) %zu first address %" PRIxPTR " second address %" PRIxPTR, n, al, bl);

I guessing this is a gcc version related issue.

Feel free to reopen the ticket if you like.

from underworld2.