BEGIN:VCALENDAR PRODID:-//Microsoft Corporation//Outlook 16.0 MIMEDIR//EN VERSION:2.0 METHOD:REQUEST X-MS-OLK-FORCEINSPECTOROPEN:TRUE BEGIN:VTIMEZONE TZID:GMT Standard Time BEGIN:STANDARD DTSTART:16011028T020000 RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=10 TZOFFSETFROM:+0100 TZOFFSETTO:-0000 END:STANDARD BEGIN:DAYLIGHT DTSTART:16010325T010000 RRULE:FREQ=YEARLY;BYDAY=-1SU;BYMONTH=3 TZOFFSETFROM:-0000 TZOFFSETTO:+0100 END:DAYLIGHT END:VTIMEZONE BEGIN:VEVENT ATTENDEE;CN="ESC F51 Meeting Room (max 8) RESTRICTED";CUTYPE=RESOURCE;ROLE= NON-PARTICIPANT;RSVP=TRUE:mailto:ESCF51MEETINGROOM@stfc.ac.uk ATTENDEE;CN=tomographic.imaging@gmail.com;RSVP=TRUE:mailto:tomographic.imag ing@gmail.com CLASS:PUBLIC CREATED:20260724T085619Z DESCRIPTION:3-4pm BST:\n\n gVXR Release v2.1.0 - Prof. Franck Vidal\n\n \n\nCheck: https://ccpi.ac.uk/events/ccpi-show-tell-and-user-su pport-drop-in/ for the upcoming schedule.\n\n \n\n4-4:30pm BST user suppor t drop in\n\n \n\n \n\nFranck’s talk will be covering what’s new in th e latest release of gVXR:\n\nThis release builds on 2.0.10 with improvemen ts in performance\, usability\, and simulation fidelity.\n\n🔧 Key Upd ates\n\n⚡ Mixed-Precision Support\n\n* Introduction and refinement of mi xed‑precision computation\n* Enables faster simulations by combining hal f (16 bits) and single (32 bits) precision on GPU where appropriate\n* Mai ntains numerical accuracy while improving GPU performance\n\n✅ Enhanced Python Integration\n\n* Improved PyPI packaging and cross-platform distrib ution\n* New digital twin framework to simplify setting up scans for speci fic devices\n* Integration with Numpy to speed up the computeXRayImage and makeIsoSurface functions\n\n🧪 Feature Refinements\n\n* Flux can now be set as a number of photons per cm2 at one metre\, seamlessly applying t he inverse square law when the source-detector distance is changes\n* Avai lability of an OpenMP backend on CPU when a GPU is not available\n* Comput ation of the peak signal-noise ratio (PSNR) when noise is enabled\n* Gener ation of iso-surfaces using the Marching Cubes algorithm to create meshes from 2D images or 3D volumes\n* Integration of XrayPhysics to generate X-r ay tube spectra without any Python dependency\n* Possibility to automatica lly maximise the magnification\n* More consistent behaviour across API fun ctions\n* Continued maturation of the simulation engine\n\n🐞 Bug Fixe s\n\n* Resolution of issues from previous 2.0.x releases\n* Improved cross -platform compatibility\n* Fixes for edge cases affecting reproducibility and stability\n\n👉 Full changelog:\nhttps://sourceforge.net/p/gvirtua lxray/gitroot/ci/master/tree/CHANGELOG\n\n \n\n \n\n DTEND;TZID="GMT Standard Time":20260728T163000 DTSTAMP:20260724T085351Z DTSTART;TZID="GMT Standard Time":20260728T150000 LAST-MODIFIED:20260724T085619Z LOCATION:ESC F51 Meeting Room (max 8) RESTRICTED\; https://teams.microsoft. com/l/meetup-join/19%3ameeting_MWQ5YjRmODItN2Y5YS00N2JiLWJhOTMtZDMwODQxOTM zYmEx%40thread.v2/0?context=%7b%22Tid%22%3a%223f66361c-a87e-4158-8f61-99e8 2db3cac8%22%2c%22Oid%22%3a%2276e00762-ea1c ORGANIZER;CN="Murgatroyd, Laura (STFC,RAL,SC)":mailto:laura.murgatroyd@stfc .ac.uk PRIORITY:5 SEQUENCE:0 SUMMARY;LANGUAGE=en-gb:CCPi Show & Tell: gVXR Release v2.1.0 - Prof. Franck Vidal\, followed by User Support Drop-in at 4pm TRANSP:OPAQUE UID:040000008200E00074C5B7101A82E0080000000030075EAB511BDD01000000000000000 010000000D5E199652F6FFA4B832881C0BDC3600A X-ALT-DESC;FMTTYPE=text/html:

3-4pm BST:< o:p>

gVXR Release v2.1.0 - Prof. Franck Vidal

\;

Check: https://ccpi.ac .uk/events/ccpi-show-tell-and-user-support-drop-in/ \;for the upcoming schedule.

\;

4-4:30pm BST user support drop in

\;

\;

Franck’s talk will be covering what’s new in the latest release of gVXR:

This release builds o n 2.0.10 with improvements in performance\, usability\, and simulat ion fidelity.

🔧\; Key Updates

< b>⚡ Mixed-Precision Support

Introduction and refinement of mixed‑precision com putation
Enables fa ster simulations by combining half (16 bits) and single (32 bits) precisio n on GPU where appropriate
Maintains numerical accuracy while improving GPU performance

✅ Enhanced Python Integration

Improved PyPI packaging and cross- platform distribution
N ew digital twin framework to simplify setting up scans for specific device s
Integration with Nump y to speed up the computeXRayImage and makeIsoSurface functions

29514\; Feature Refinements

Flux can now be set as a number of pho tons per cm² at one metre\, seamlessly applying the inverse squ are law when the source-detector distance is changes
Availability of an OpenMP backend on CPU whe n a GPU is not available
Computation of the peak signal-noise ratio (PSNR) when noise is enabled< o:p>
Generation of iso-surfa ces using the Marching Cubes algorithm to create meshes from 2D images or 3D volumes
Integration of XrayPhysics to generate X-ray tube spectra without any Python dependenc y
Possibility to automa tically maximise the magnification
More consistent behaviour across API functions
Continued maturation of the simulation engine

🐞\; Bug Fixes

Resolution of issues from prev ious 2.0.x releases
Imp roved cross-platform compatibility
Fixes for edge cases affecting reproducibility and stability

👉\; Full changelog:
https://sourceforge.net/p/ gvirtualxray/gitroot/ci/master/tree/CHANGELOG

\;

\;

X-MICROSOFT-CDO-BUSYSTATUS:BUSY X-MICROSOFT-CDO-IMPORTANCE:1 X-MICROSOFT-DISALLOW-COUNTER:FALSE X-MS-OLK-APPTSEQTIME:20260724T085351Z X-MS-OLK-AUTOFILLLOCATION:FALSE X-MS-OLK-CONFTYPE:0 BEGIN:VALARM TRIGGER:-PT15M ACTION:DISPLAY DESCRIPTION:Reminder END:VALARM END:VEVENT END:VCALENDAR