hanrach / p2d_fast_solver Goto Github PK

Hi@hanrach,
I study to use the code to compute the heat during charge and discharge cycle. but the code do not provide example to consult with. I try to modified the code as below (naive/p2d_main_fn.py)

I can obtain the Ohm heat and rxn heat, but can not arrive at reversible heat.

It is appreciated if you help me to solve this problem.

Hi,

As you may know, jax.ops.index* have been deprecated and we must use "ndarray.at[idx]" like indexing.
Do you have any plans to update this rule?

@cbm755 @hanrach Hi, I modified the code (p2d_main_fn.py) as I understood, but the result is wrong, can you help me to improve the code for heat generation and SOC. I post the heat generation as below:

I think the recommended imports are:

import stuff_from_python_stdlibs
import more_stuff_from_python_stdlibs
<blank line>
import packages_like_scipy
import and_numpy
<blank line>
import local_stuff

Not urgent but we should reorder some of the files to follow this. I'm a little hesitant in case the order matters with some of the jax stuff...

I'm going to track dependencies of what I need to get this running again. This should be used to create a virtual env for reproducibility. See also #7

So far:

Ubuntu 20.04
Python 3.8
sudo apt install libsuitesparse-dev swig
pip install scikit-umfpack==0.3.2
pip install jax==0.2.20 jaxlib==0.1.71

Newer Jax would probably require some effort. I could not get old 0.2.20 jax to work on Python 3.10:

Traceback (most recent call last):
  File "/home/cbm/work/People/Rachel_Han/p2d_fast_solver.git/examples/main_reorder.py", line 1, in <module>
    from jax.config import config
    from jax._src.numpy import lax_numpy
  File "/home/cbm/.local/lib/python3.10/site-packages/jax/_src/numpy/lax_numpy.py", line 53, in <module>
    from jax._src.ops import scatter
  File "/home/cbm/.local/lib/python3.10/site-packages/jax/_src/ops/scatter.py", line 31, in <module>
    from typing import EllipsisType
ImportError: cannot import name 'EllipsisType' from 'typing' (/usr/lib64/python3.10/typing.py)

It would be helpful to show the output of a successful run.

For example in 8b1fa99#commitcomment-85160488 someone asks about overly large Ohm heat values.

@jsfeng-fudan asked:

The Qrev() function prototype is Qrev(j,Tc,cM,cMp,cmax). I dont know the meaning of cM and cMp, I can usig the umat_pe to replace them.
can you help me to solve this problem?

I didn't install numba, and can run the main codes but in utils/banded_matrix.py I see:

import numba as nb