# pvtu : number of pieces and relationship to parallel visualisation in Paraview

**URL:** https://discourse.vtk.org/t/pvtu-number-of-pieces-and-relationship-to-parallel-visualisation-in-paraview/6522
**Category:** Support
**Created:** [August 31, 2021, 9:51pm UTC](https://discourse.vtk.org/t/pvtu-number-of-pieces-and-relationship-to-parallel-visualisation-in-paraview/6522 "2021-08-31T21:51:11Z")
**Posts on this page:** 1
**Page:** 1

<div class="post-metadata">

### Author: ![Nicholas\_Yue](https://discourse.vtk.org/user_avatar/discourse.vtk.org/nicholas_yue/32/3751_2.png) [@Nicholas\_Yue](https://discourse.vtk.org/u/Nicholas_Yue)
#### Post date: [August 31, 2021, 9:51pm UTC](https://discourse.vtk.org/t/pvtu-number-of-pieces-and-relationship-to-parallel-visualisation-in-paraview/6522/1 "2021-08-31T21:51:11Z")

</div>

I am learning about vtu/pvtu

I am testing out deal.ii which save data as vtu/pvtu, only one piece is being written out per snapshot.

I plan to test out visualizing the data set on a 4 node cluster (each with GPU) and was wondering if I should invest the time to investigate writing out 4 piece for each \*.pvtu to benefit my cluster setup. Does the number of pieces have any performance improvement value if it matches my node count?

Cheers