I noticed since we are able to run Theory In WSL container, that the docker results show a big gap between elapsed time and CPU-time (cputime much lower than elapsed)
VBox tasks don't show a big difference between elapsed and cpu-time. Sometimes even equal or CPU higher than elapsed.
I wanted to find out whether running Theory in a VirtualBox VM is faster then in a docker container on a Windows system.
To test that you should exactly run the same task in both circumstances.
I did not want to wait > 100 years to get the same job with the same parameters, so I had to manipulate the input file, so that the job is exactly doing the same processing.
A first quick test showed me that VirtualBox was -surprise- much slower. Coïncidence?
So I did a second test under the same host-load.
The outcome of 2 times job ppbar zinclusive 1800 -,-,50,130 - pythia8 8.303 CP1-CR2 100000 642

                   1. docker                               2. VirtualBox
			
Returned          15 Feb 2026, 11:11:24 UTC                15 Feb 2026, 12:37:39 UTC
Computer ID       10772116		                   10772116
Run time          19 min 22 sec		                   53 min 36 sec
CPU time          16 min 36 sec                            52 min 34 sec
Validate state    Valid                                    Valid
Credit            8.93		                           24.69
peak FLOPS        3.32 GFLOPS                              3.32 GFLOPS
App-version       Theory Simulation v302.10 (docker)       Theory Simulation v301.00 (vbox64_theory)
PeakWorkingSet    13.20 MB                                 69.93 MB
Peak swap size    6.63 MB                                  56.40 MB
Peak disk usage   8.31 MB                                  614.12 MB

job: run exitcode 0                                        0
job: diskusage    18052                                    18084
job: logsize      8072 k                                   8192 k
job: time         23m 15.5s                                50m 55.4s
job: cpuusage     1447 s                                   3080 s

Cern-IT want to make this change from Virtualbox to Docker.
Docker have less Overhead against Virtualbox.

It would be good to get more help for this Transfer.
But, Docker is not for free to use it.

Two other jobs where I tested the cpu- and elapsed time difference when using VBox rather than docker for the same job:

pp bbbar 7000 20 - pythia8 8.313 default 100000 644

vbox:   Run time  1 hours 18 min  2 sec        CPU time  1 hours 16 min 34 sec
docker: Run time          45 min  3 sec        CPU time          38 min 16 sec

pp zinclusive 8000 -,-,100,170 - pythia8 8.306 tune-AU2m 100000 650

vbox:   Run time  5 hours  7 min 54 sec        CPU time	4 hours 57 min 50 sec
docker: Run time  1 hours 54 min 10 sec        CPU time 1 hours 40 min 49 sec