# Not a number computing L2 and H1 errors
**URL:** https://community.freefem.org/t/not-a-number-computing-l2-and-h1-errors/4040
**Category:** General Discussion
**Created:** [August 14, 2025, 7:20pm UTC](https://community.freefem.org/t/not-a-number-computing-l2-and-h1-errors/4040 "2025-08-14T19:20:02Z")
**Posts on this page:** 4
**Page:** 1
### Author:  [@juanpaca](https://community.freefem.org/u/juanpaca)
#### Post date: [August 14, 2025, 7:20pm UTC](https://community.freefem.org/t/not-a-number-computing-l2-and-h1-errors/4040/1 "2025-08-14T19:20:02Z")
Dear FreeFem community, I hope you are well.
I am implementing a stabilized method for a reaction–diffusion model, which is
\sigma u - \epsilon \Delta u = f in \Omega, u = 0 on \partial \Omega. When I set \epsilon = 10^{-5} and \sigma = 1.0, and compute the L^2(\Omega) error, I get the following:
```auto
L2 error: -nan
H1 semi-norm error: -nan
H1 error: -nan
```
I get the same results for a standard Galerkin. I want to explore the error for \epsilon ranging from 10^{-3} to 10^{-6}. Do you know how to solve this?
---
### Author:  [@prj](https://community.freefem.org/u/prj)
#### Post date: [August 14, 2025, 7:22pm UTC](https://community.freefem.org/t/not-a-number-computing-l2-and-h1-errors/4040/2 "2025-08-14T19:22:55Z")
Your vector has NaN in it, so the norms are NaN.
---
### Author:  [@juanpaca](https://community.freefem.org/u/juanpaca)
#### Post date: [August 14, 2025, 7:24pm UTC](https://community.freefem.org/t/not-a-number-computing-l2-and-h1-errors/4040/3 "2025-08-14T19:24:45Z")
By vector you mean the vector of DoF’s of the solution u\_h?
---
### Author:  [@frederichecht](https://community.freefem.org/u/frederichecht)
#### Post date: [August 25, 2025, 9:34am UTC](https://community.freefem.org/t/not-a-number-computing-l2-and-h1-errors/4040/4 "2025-08-25T09:34:09Z")
Yes, and try with greater ϵ=10^−5