# Effect of changing processor number **URL:** \ **Category:** Unified Model\ **Tags:** ARCHER2, PUMA2\ **Created:** [5 October 2026 12:40 UTC](https://cms-helpdesk.ncas.ac.uk/t/effect-of-changing-processor-number/2189 "2026-10-05T12:40:35Z")\ **Posts on this page:** 2\ **Page:** 1
**Author:** ![timmyf](https://avatars.discourse-cdn.com/v4/letter/t/da6949/32.png) [@timmyf](https://cms-helpdesk.ncas.ac.uk/u/timmyf)\ **Post date:** [5 October 2026 12:40 UTC](https://cms-helpdesk.ncas.ac.uk/t/effect-of-changing-processor-number/2189/1 "2026-10-05T12:40:35Z")
Hi, When I was testing one my code modifications in UM12.0 in Archer2-puma2 for my research related work, I tried changing to the below processor numbers in the rose-suite.conf MAIN\_ATM\_PROCX=26 MAIN\_ATM\_PROCY=24 compared to the original values of MAIN\_ATM\_PROCX=24 MAIN\_ATM\_PROCY=21. I have kept all other entries related to the processor numbers intact (OMP\_ATM, PE\_ATM, etc.). This was to test if my code modifications has any dependence on the processor number configurations. But I did find some small changes in the model output. I was wondering if this is an issue with my code modifications, or you really expect the results to look slightly different when I make the processor number changes as above? Cheers, Timmy ---
**Author:** ![mdalvi](https://avatars.discourse-cdn.com/v4/letter/m/ad7895/32.png) [@mdalvi](https://cms-helpdesk.ncas.ac.uk/u/mdalvi)\ **Post date:** [5 October 2026 13:07 UTC](https://cms-helpdesk.ncas.ac.uk/t/effect-of-changing-processor-number/2189/2 "2026-10-05T13:07:57Z")
Hi, This behaviour depends on the model schemes. If this is an UKESM type configuration with full chemistry then changing answers with PE setup is a known feature of the solver used. On the other hand a GA/GC model without chemistry is expected to maintain the results for different PE setups. If this is not seen a bug has likely been introduced from a code change. -- Mohit