Peter Willendrup

Results 264 comments of Peter Willendrup

(It should also be investigated if the example fully works in OpenACC and combined MPI-OpenACC settings)

Workaround found (may require editing filesystem permissions for current user) * In your McStas environment shell, please install a conda-based gcc: * ```conda install m2w64-gcc``` * cd to the location...

Please double-check wrt. the equivalent McStas component code, if I remember correctly I have managed to iron out some fluctuating results on 3.0 / GPU there - potentially related.

Hi @g5t, thanks for both issue and proposed solution. :) I OK with you, I might come by for a "demo" at your desk also, I think there may be...

@g5t now reproduced on vanilla Ubuntu 22.04 Arm. As we discussed, likely related to exact version of Qt/pyqt. Will fix on master and mccode-3 branches toward next set of stable...

![Screenshot 2022-08-19 at 13 19 30](https://user-images.githubusercontent.com/8593355/185608697-c19ec25c-b265-4a8c-a1f8-3d58fceb5b1f.png)

These are ready: ``` EPSD_monitor.comp E_monitor.comp L_monitor.comp PSD_monitor.comp PSD_monitor_4PI.comp PSD_monitor_coh.comp (Monitor_nD.comp already has dynamics but not via DArray's) (PreMonitor_nD.comp goes together with Monitor_nD) (Monitor.comp - is a one-bin monitor ;-)...

This should now be fixed on mccode-3 (plus pulled to master to allow dynamic arrays also there)

McStas: 2.7.1 similarly reports built for macOS 11. (Outdated 3.0 also does not run on earlier releases than 10.15 it seems...) McXtrace: 3.0 similar issue related: https://github.com/McStasMcXtrace/McCode/issues/1229

master: ``` [FAILED] ILL_H5: H5_I = 1.86744e+13 +/- 30175800000, should be 7.46e+13 ``` mccode-3: ``` [FAILED] ILL_H5: H5_I = 1.8665e+13 +/- 30180700000, should be 7.46e+13 ``` Clean factor of 4...