# GMRES (krylovkit) info.residual

**URL:** <https://discourse.julialang.org/t/gmres-krylovkit-info-residual/79860>\
**Category:** General Usage\
**Tags:** linearalgebra\
**Created:** [April 22, 2022, 6:12pm UTC](https://discourse.julialang.org/t/gmres-krylovkit-info-residual/79860 "2022-04-22T18:12:54Z")\
**Posts on this page:** 1\
**Page:** 1

<div class="post-metadata">

**Author:** ![Srikumar](https://sea2.discourse-cdn.com/julialang/user_avatar/discourse.julialang.org/srikumar/32/34565_2.png) [@Srikumar](https://discourse.julialang.org/u/Srikumar)\
**Post date:** [April 22, 2022, 6:12pm UTC](https://discourse.julialang.org/t/gmres-krylovkit-info-residual/79860/1 "2022-04-22T18:12:54Z")

</div>

I find that the norm(info.residual) that GMRES prints at each step (inside a single iteration) is different from the actual residual norm ||(b-Ax)|| that one computes with the current approximate x manually. Are these 2 norms (GMRES and manual) supposed to match at the end of every iteration though?
