# Umpfack errors with periodic boundary and slightly large mesh in 3D problem

**URL:** https://community.freefem.org/t/umpfack-errors-with-periodic-boundary-and-slightly-large-mesh-in-3d-problem/295
**Category:** General Discussion
**Created:** [March 8, 2020, 12:44am UTC](https://community.freefem.org/t/umpfack-errors-with-periodic-boundary-and-slightly-large-mesh-in-3d-problem/295 "2020-03-08T00:44:55Z")
**Posts on this page:** 4
**Page:** 1

<div class="post-metadata">

### Author: ![caprilo](https://avatars.discourse-cdn.com/v4/letter/c/67e7ee/32.png) [@caprilo](https://community.freefem.org/u/caprilo)
#### Post date: [March 8, 2020, 12:44am UTC](https://community.freefem.org/t/umpfack-errors-with-periodic-boundary-and-slightly-large-mesh-in-3d-problem/295/1 "2020-03-08T00:44:55Z")

</div>

I’m solving a 3D heat problem. This is a minimal working version of the code:

```
load "msh3";	

func T0 = x < .5;

int p = 10;

mesh3 dominio = cube(p, p, p);

fespace Vh(dominio, P2,
	periodic = [[1, x, z], [3, x, z], [2, y, z], [4, y, z], [5, x, y], [6, x, y] ]
);

Vh T = T0, v, Tant;

real dt = 0.1;

problem termico(T, v)
    = int3d(dominio)(
          T * v / dt
        + (
              dx(T) * dx(v)
            + dy(T) * dy(v)
            + dz(T) * dz(v)
        )
    )
    + int3d(dominio)(
        - Tant * v / dt    		
    )
;

for(real t = 0; t < 30; t += 5){
	Tant = T;
	termico; 
	plot(T);
}

```

This code runs OK. But if I raise the number of points on the mesh (for example: `p = 20` instead of `p = 10`), then the following error occurs when solving each iteration:

```
Error umpfack umfpack_di_numeric status -1
Error umfpack_di_solve status -3

```

If I remove the periodic condition, then no error occurs. What might be happening?

I am using version v4.5-18-gcf799e13, installed from the deb package available in [freefem.org](http://freefem.org) and Debian 10.

---

<div class="post-metadata">

### Author: ![prj](https://avatars.discourse-cdn.com/v4/letter/p/ecae2f/32.png) [@prj](https://community.freefem.org/u/prj)
#### Post date: [March 8, 2020, 11:11am UTC](https://community.freefem.org/t/umpfack-errors-with-periodic-boundary-and-slightly-large-mesh-in-3d-problem/295/2 "2020-03-08T11:11:18Z")

</div>

Use another solver, like MUMPS, or even UMFPACK64.

---

<div class="post-metadata">

### Author: ![caprilo](https://avatars.discourse-cdn.com/v4/letter/c/67e7ee/32.png) [@caprilo](https://community.freefem.org/u/caprilo)
#### Post date: [March 11, 2020, 5:06pm UTC](https://community.freefem.org/t/umpfack-errors-with-periodic-boundary-and-slightly-large-mesh-in-3d-problem/295/3 "2020-03-11T17:06:30Z")

</div>

Indeed, using UMFPACK64 solves the problem. Thanks!

---

<div class="post-metadata">

### Author: ![Qianrui](https://avatars.discourse-cdn.com/v4/letter/q/9fc348/32.png) [@Qianrui](https://community.freefem.org/u/Qianrui)
#### Post date: [January 7, 2022, 8:01am UTC](https://community.freefem.org/t/umpfack-errors-with-periodic-boundary-and-slightly-large-mesh-in-3d-problem/295/4 "2022-01-07T08:01:10Z")

</div>

> [@prj](#):
>
> UMFPACK64.

solver UMFPACK64，May it be used to solve PDE in complex fields?
