As servers may become inaccessible this content is hosted from three servers. If you find access problems, try another server.
To subscribe to the low volume mailing list, send an email to opticks+subscribe@groups.io
| https://groups.io/g/opticks | forum/mailing list archive |
| email:opticks+subscribe@groups.io | subscribe to mailing list |
Note that the Opticks documentation and examples are lagging behind Opticks developments. So I suggest only a cursory look at the above documentation.
Any examples using G4Opticks.hh are from the old workflow and should not be followed. The G4CX package is the top level package in the new workflow that replaces G4OK.
A scratch document with some notes that are uptodate (as of December 2022) is available at:
I do not recommend the use of NVIDIA OptiX < 7 currently, as all Opticks development effort is currently on using the all new NVIDIA OptiX 7 API.
Note that the NVIDIA OptiX 6->7 transition is like an entirely different project, practically all Opticks code has been changed or needs to change to accomodate this API transition.
The Opticks "reference" NVIDIA OptiX version is currently 7.0.0 so I recommend you use the first column of versions in the below.
The recommended version sets to use are driven by the OptiX version release notes. They specify the development CUDA version used to build each OptiX version and the minimum NVIDIA driver version.:
OptiX *7.0.0* 7.5.0 7.6.0 NVIDIA Driver 435.21+ 515+ 520+ CUDA 10.1 11.7 11.8
Beyond that the CUDA version then constrains the versions of gcc and c++ dialect that can be used:
CUDA 10.1 11.7 11.8 nvcc c++ c++03,11,14 c++03,11,14,17 ? gcc 8.3.0 11.2 ?
Integration of JUNO simulation framework with Opticks: GPU accelerated optical propagation via NVIDIA OptiX
CHEP 2021 Proceedings : Simon Blyth, EPJ Web of Conferences 251, 03009 (2021)
Meeting the challenge of JUNO simulation with Opticks: GPU optical photon acceleration via NVIDIA OptiX
CHEP 2019 Plenary Proceedings : Simon Blyth, EPJ Web of Conferences 245, 11003 (2020)
Opticks : GPU Optical Photon Simulation for Particle Physics using NVIDIA OptiX
CHEP 2018 Proceedings : Simon Blyth, EPJ Web Conf. 214 02027 (2019)
http://iopscience.iop.org/article/10.1088/1742-6596/898/4/042001
Blyth, Simon C, 2017 J. Phys.: Conf. Ser. 898 042001
Opticks : GPU Optical Photon Simulation for Particle Physics using NVIDIA OptiX
CHEP 2017 Proceedings : Published by IOP Science
https://geant4-data.web.cern.ch/ReleaseNotes/ReleaseNotes.11.0.html
Geant 11.0 release notes, new example:
CaTS - New application implementing a flexible and extendable framework for the simulation of calorimeter and tracking detectors. It also demonstrates how to use Opticks for the creation and propagation of optical photons.
https://www.youtube.com/watch?v=CBpOha4RzIs
Longer cut: https://www.youtube.com/watch?v=QzH6y0pKXk4
Repository with some examples of CUDA and Thrust, including an extended "reconstruction" example integrating a Minuit2 controlled fit with NLL evaluated on the GPU using thrust::transform_reduce.
Learning NumPy is highly recommended for more efficient data preparation for upload to the GPU and analysis of results.
env/presentation/opticks_20240418_ihep_epd_seminar_story_of_opticks.html
The Story of Opticks, applying NVIDIA OptiX GPU ray tracing to Optical Photon Simulation
IHEP EPD Seminar, Beijing, April 18, 2024
The story behind the development of Opticks will be detailed, covering the ups and downs of cutting edge development and some of the hard learned lessons of experience that can benefit anyone working on innovative new techniques.
env/presentation/opticks_20240227_zhejiang_seminar.html
Opticks : Optical Photon Simulation via GPU Ray Tracing from NVIDIA® OptiX™
Zhejiang University Seminar, Hangzhou, 27 Feb 2024
env/presentation/opticks_20240224_offline_software_review.html
JUNOSW + Opticks : Status and Plan
JUNO Collaboration Meeting, 24 Feb 2024, Kaiping
env/presentation/opticks_20240215_geant4_forum.html
Slides describing G4OpBoundaryProcess customizations for the JUNO PMT Optical Model (some of which may be presented by Cécile Jollet at the below meeting)
Geant4 Technical Forum, 15 Feb 2024
env/presentation/opticks_20231219_using_junosw_plus_opticks_release.html
Advert for first JUNOSW+Opticks relase
AFG Simulation Meeting, Dec 19, 2023
env/presentation/opticks_20231211_profile.html
Profiling + Status
Simulation Meeting, IHEP, Beijing, China, Dec 11, 2023
env/presentation/standalone_20230930_cpp_test_debug_ana_with_numpy.html
Standalone C++ software testing, debugging and analysis with python packages: NumPy, matplotlib, pyvista, ...
Hohhot School, Inner Mongolia, China, Sept 30, 2023
env/presentation/opticks_20230907_release.html
JUNOSW + Opticks : Release Planning
JOC Meeting, IHEP, Beijing, Sept 7, 2023
env/presentation/opticks_20230726_kaiping_software_review.html
JUNO+Opticks : Progress Review
JUNO Summer Meeting, Kaiping, Software Review, July 26, 2023
env/presentation/opticks_20230611_qingdao_sdu_workshop.html
Opticks : GPU Optical Photon Simulation via NVIDIA® OptiX™ 7, NVIDIA® CUDA™ (11 June 2023)
Workshop on HEP Computing and Software, SDU, Qingdao Campus, Remote Presentation
env/presentation/opticks_20230525_MR180_timestamp_analysis.html
Opticks + JUNO : MR180 Timestamp Analysis (25 May 2023)
env/presentation/opticks_20230508_chep.html
Opticks : GPU Optical Photon Simulation via NVIDIA® OptiX™ 7, NVIDIA® CUDA™ (8 May 2023)
Presented by Dr Tao Lin, IHEP, CAS
Opticks + JUNO : More junoPMTOpticalModel issues and validation of CustomG4OpBoundaryProcess fix (28 April 2023)
env/presentation/opticks_20230206_JUNO_PMT_Geometry_and_Optical_Model_Progress.html
Opticks + JUNO : PMT Geometry and Optical Model Progress (6 February 2023)
Summary of last 6 months, Abstract:
PMT Geometry fixes and several bugs found with the FastSim based junoPMTOpticalModel implementation are presented, including incorrect polarization and propagation time. A proposed fix using standard Geant4 together with customized G4OpBoundaryProcess is described. The fix has the advantage of allowing a natural 2-volume PMT geometry which can also be used on GPU.
junoPMTOpticalModel FastSim issues, and proposed fix using a CustomG4OpBoundaryProcess (20 December 2022)
- fixes would facilitate bringing the TMM functionality to GPU within Opticks
env/presentation/opticks_20221117_mask_debug_and_tmm.html
JUNO + Opticks geometry fixes (Mask, MaskTail, MaskVirtual) (17 November 2022)
- Validations reveals Opticks and JUNO bugs
- rare close to ellipsoid "apex" rays missed : fixed with zcut+safety
- nmskSolidMaskTail uncoincidence : still spurious intersects
- thin cylinder (hz 0.15 mm) "lip" issue : mis-translation + precision loss
- arises from near axial rays onto very thin cylinders
- fixed by reimplementing CSG_CYLINDER more simply (less flops)
- nmskSolidMaskVirtual : fixed : CSG_CONE quadratic precision loss, close to apex issue
- hama : MaskTail crosses PMT body
- junoPMTOpticalModel TMM calc on GPU
env/presentation/opticks_20220718_towards_production_use_juno_collab_meeting.html
JUNO Progress Towards Production : Completed OptiX 7 Simulation re-implementation (July 18, 2022)
- Validation : Opticks-Geant4 A-B testing
- Opticks Packages : Many Removed, Many Added
- Simulation : Why so much had to be re-implemented ?
- Opticks Packages : Many more to be removed
- Systematic Approach to Random Alignment of Two Simulations
- NEXT : Update JUNO-Opticks integration to use new Opticks
env/presentation/opticks_20220329_progress_towards_production.html
JUNO Progress Towards Production : QSim WIP (March 29, 2022)
- January : Updating to JUNO trunk geometry
- February : Generalize Opticks CSG to handle "list-nodes" : n-ary (not binary) CSG
- March : Geometry Optimization + Migrating Sim from Old to New Workflow
- Big Picture of Multi Year Transition to new workflow (OptiX 7 compatible)
- Standalone Geant4 Tests : Cerenkov, RayLeigh, Boundary
- Boundary matched QSimTest.sh and G4OpBoundaryProcessTest.sh
- Dynamic Prim Selection :in memory CSGCopy with selection => construct any combination of Prim
- enables per-prim measurement : to locate source of slowdowns
- Dynamic Prim Selection Speed Scan : One "G4VSolid" Geometries
env/presentation/opticks_20220307_fixed_global_leaf_placement_issue.html
JUNO Geometry Updates and Fixes (March 7, 2022)
- Investigate XJfixtureConstruction Complex CSG Solid
- Reveals incompatibility of : CSG intersect alg and tree balancing
- develop CSG list-nodes as solution for complex CSG without balancing
- Algorithm for Ray Intersection onto Shape defined by CSG Tre
- Generalized Opticks CSG into three levels : tree < node < leaf (to support list-nodes without recursion)
- CSG_CONTIGUOUS Union : n-ary (not bin-ary) CSG intersection
- CSG_OVERLAP Intersection : CSG intersection
- More closely suiting intersect algorithm to geometry => better performance
- transform bug from incomplete tree/node/leaf split
- OTracerTest debug comparing rasterized OpenGL render with raytrace render
env/presentation/opticks_20220227_LHCbRich_UK_GPU_HACKATHON.html
LHCb RICH + NVIDIA UK GPU HACKATHON (February 28, 2022)
- Instructions for RICH Simplified rendering with Opticks/CSGOptiX and NVIDIA OptiX 7
env/presentation/opticks_20220118_juno_collaboration_meeting.html
JUNO+Opticks Photons : Validation and Deployment Plan (January 18, 2022)
- Geometry/Translation Issues Summary
- PMT Mask Modelling Fix
- Fastener : interfering subtraction-of-subtraction issue from daughter cavity
- Opticks CSG is surface-centric : more fragile but inherently faster that Geant4 volume-centric implementation
- 2D Geant4 Geometry Slicing
- Cutdown PMT breaks Opticks translation, ZSolid fix : actually cut the tree
- Render Speed Check After Geometry fixes : smallish time range
- History Matching Check
- Chi2 points to PE_PA solid discrepant => XJfixtureConstruction
- XJfixtureConstruction : Complex CSG causing coincidence problems
- Investigate XJfixtureConstruction Positions : geometry overlaps, mid-chimney
- XJanchorConstruction : spurious line associated wit big sphere subtraction
- SJReceiverConstruction : disjoint
env/presentation/opticks_20220115_innovation_in_hep_workshop_hongkong.html
IAS Program on High Energy Physics 2022 : Mini-workshop on Experiment/Detector: Innovation in HEP Detectors and Computing, January 13-14, 2022
Opticks : Innovation in Optical Photon Simulation via state-of-the-art GPU Ray Tracing from NVIDIA® OptiX™ (January 14, 2022)
- Movies ≈ monte carlo optical photon simulations
- Fundamental "Rendering Equation" of Computer Graphics (Kajiya 1986), Neumann series solution
- Monte carlo path tracing : limited by ray tracing
- Optical Simulation : Computer Graphics vs Physics
- SIGGRAPH 2018 : NVIDIA Quadro RTX : leap in ray tracing performance
env/presentation/opticks_20211223_pre_xmas.html
JUNO Opticks : OptiX 7 Ray Trace times >100x faster following geometry fixes (December 23, 2021)
- ZSolid::ApplyZCutTree applied to both NNVT and Hama PMTs
- fixed base_steel G4Polycone with multiple R-inner
- Opticks-Offline CMake Integration
- subtraction of huge acrylic sphere led to huge Fastener bounding box fixed by positivizing CSG tree and not including bbox of complemented
- render timings suggests no bad bottlenecks remaining
- LHCb RICH SphereWithPhiSegment : CSG_PHICUT unbounded expt
- NEXT : mask tail cutting across PMT bulb, integrate CSGOptiX into Opticks
env/presentation/opticks_20211117.html
JUNO Opticks : PMT Z-cut issue : (November 17, 2021)
- Hamamatsu PMT Solid breaking Opticks : CSG tree height 8 : TOO DEEP
- SOLUTION ACTUALLY CUT THE CSG TREE : using ZSolid.hh
- ZSolid::ApplyZCutTree using "placement new" trick
- MAJOR SIMPLIFICATION tree cut from (node:15, prim:8, height:7) -> (node:7, prim:4, height:3)
env/presentation/opticks_autumn_20211019.html
Opticks Autumn : (October 19, 2021)
- Cerenkov energy sampling via ICDF lookups ?
- Monte Carlo "Rejection Sampling" VS "Inverse Transform Sampling"
- Opticks Updates for Geant4 11.beta (1100)
- Planar 2d ray tracing to create CSG geometry cross-sections (later known as "simtrace")
- AdditionAcrylicConstruction : pointless CSG subtraction of holes
- CSG sub sub bug ? Subtracted subtraction yielding spurious intersects
- inherent CSG fragility regarding coincident faces
- DO NOT subtract CSG holes for daughters : not necessary, doubles expense, causes bugs
env/presentation/juno_opticks_cerenkov_20210902.html
JUNO Opticks : Summer Progress (Sep 2, 2021)
- "s2" Cerenkov integration (HMM: does s2 impl cope with float better ?)
- PMTSimParamSvc::get_pmt_ce bug
- New Opticks Package for NVIDIA OptiX 7.0 API : CSG, CSGOptiX, QUDARap, (transitional: CSG_GGeo)
- CPU/GPU counterpart header pattern eg QSim.hh/qsim.h
- NEXT : Looking for single precision Cerenkov generation solution
env/presentation/lz_opticks_optix7_20210727.html
Notes on OptiX 6->7 migration : "Qudarap" pure CUDA photon generation (July 27, 2021, Virtual, H7)
- Cerenkov Matching : requires double precision rejection sampling, attempts to understand why
- nascent qudarap : qctx.h(later becomes qsim.h)
- NEXT : gs seeding photons, integrate CSGOptiX with qudarap
env/presentation/juno_opticks_20210712.html
JUNO Opticks/Geant4 Optical Photon Simulation Matching (July 12, 2021)
- Tools for Optical Photon Simulation Matching
- JUNO Offline DsG4Scintillation : Reemission Bookkeeping
- Long list of fixed JUNO-Opticks material property/geometry issues
- Scintillation Wavelength chi2 Matching : Multi-resolution GPU texture
- JUNO Offline trunk G4Cerenkov_modified : BUG FIX
- NEXT : Slow Fasteners CSG, Unify wit OptiX 7 "branch"
env/presentation/opticks_vchep_2021_may19.html
Integration of JUNO simulation framework with Opticks (May 19, 2021) vCHEP
- New "Foundry" Model : Shared CPU/GPU Geometry Context
- First OptiX 7 Renders : slow geometry found
env/presentation/lz_opticks_optix7_20210518.html
Notes on OptiX 6->7 migration, CSG, CSG_GGeo, CSGOptiX (May 18, 2021, Virtual, H6)
1st full JUNO ray trace with OptiX 7 CSGOptiX enabled by CSGFoundry model
env/presentation/lz_opticks_optix7_20210504.html
Notes on OptiX 6->7 migration, CSG, CSGOptiX, CSGOptiXGGeo (May 4, 2021, Virtual, H5)
env/presentation/juno_opticks_20210426.html
JUNO-Opticks Notes, slow solids identified (April 26, 2021)
env/presentation/lz_opticks_optix7_20210406.html
Notes on OptiX 6->7 migration, "Foundry" CSG Model (April 4, 2021, Virtual, H4)
env/presentation/lz_opticks_optix7_20210315.html
Notes on OptiX 6->7 migration, (Mar 15, 2021, Virtual, H3)
env/presentation/opticks_detector_geometry_caf_mar2021.html
Compute Accelerator Forum (March 10, 2021, Virtual, 20+10 min)
env/presentation/lz_opticks_optix7_20210225.html
Notes on OptiX 6->7 migration, (Feb 25, 2021, Virtual, H2)
env/presentation/lz_opticks_optix7_20210208.html
Notes on OptiX 6->7 migration, (Feb 8, 2021, Virtual, H1)
env/presentation/opticks_jan2021_juno_sim_review.html (Feb 1, 2021, Virtual)
Review JUNO + Opticks progress + plans, 20 mins
env/presentation/opticks_aug2020_sjtu_neutrino_telescope_workshop.html (August 2020, Virtual)
SJTU Shanghai, Next Generation Neutrino Telescope Workshop
Opticks : GPU photon simulation via NVIDIA® OptiX™ + GPU/Graphics background + Application to neutrino telescope simulations ?
env/presentation/opticks_jul2020_juno.html (6 July 2020, IHEP Virtual)
JUNO Collaboration Meeting, Software Session
Integration of JUNO Offline + Opticks : GPU Optical Simulation with NVIDIA® OptiX™
env/presentation/opticks_may2020_hsf.html (27 May 2020, CERN Virtual)
HSF Simulation Working Group Meeting, 20 minutes presentation
Opticks GPU Optical Simulation with NVIDIA® OptiX™ - Development Experience : Problems and Successes
env/presentation/opticks_gpu_optical_photon_simulation_dec2019_ihep_epd_seminar.html (Dec 2019, Beijing)
IHEP EPD/PIFI Seminar
Opticks : GPU Optical Simulation via NVIDIA® OptiX™ + A Mental Model for Effective Application of GPUs
env/presentation/opticks_gpu_optical_photon_simulation_nov2019_chep.html (Nov 2019, Adelaide)
CHEP 2019 Invited Plenary
Meeting the Challenge of JUNO Simulation with Opticks : GPU Optical Photon Acceleration via NVIDIA® OptiX™
env/presentation/opticks_gpu_optical_photon_simulation_jul2019_ihep.html (July 2019, Beijing)
JUNO Collaboration Meeting
Progress with JUNO + Opticks : GPU Optical Photon Simulation with NVIDIA OptiX
env/presentation/opticks_gpu_optical_photon_simulation_jan2019_sjtu.html (17 Jan 2019, Shanghai)
JUNO Collaboration Meeting
Status of Opticks : GPU Optical Photon Simulation for Particle Physics with NVIDIA OptiX
env/presentation/opticks_gpu_optical_photon_simulation_oct2018_ihep.html (31 Oct 2018, Beijing)
IHEP Seminar
Tools, Techniques and Opticks : GPU Optical Photon Simulation for Particle Physics with NVIDIA OptiX
env/presentation/opticks_gpu_optical_photon_simulation_jul2018_chep.html (10 July 2018, Sofia)
CHEP 2018 Conference
Opticks : GPU Optical Photon Simulation for Particle Physics with NVIDIA OptiX
env/presentation/opticks_gpu_optical_photon_simulation_sep2017_wollongong.html (Sep 2017, Wollongong)
Geant4 Collaboration Meeting, University of Wollongong, Australia
A brief introduction to Opticks, presented in plenary session despite being listed in parallel.
Opticks : GPU Optical Photon Simulation for Particle Physics with NVIDIA OptiX
env/presentation/opticks_gpu_optical_photon_simulation_sep2017_jinan.html (Sep 2017, Jinan)
Shandong University, Physics Dept, Jinan, China
Opticks : GPU Optical Photon Simulation for Particle Physics with NVIDIA OptiX
env/presentation/opticks_gpu_optical_photon_simulation_jul2017_ihep.html (July 2017, Beijing)
JUNO Collaboration Meeting, IHEP, Beijing, China
Opticks : GPU Optical Photon Simulation for Particle Physics with NVIDIA OptiX
env/presentation/opticks_gpu_optical_photon_simulation_jan2017_psroc.html (17 Jan 2017, Taipei)
PSROC Annual Meeting, Tamkang University, TKU, New Taipei City, Taiwan
Opticks : GPU Optical Photon Simulation for Particle Physics with NVIDIA OptiX
env/presentation/opticks_gpu_optical_photon_simulation_nov2016_llr.html (2 Dec 2016, Paris)
JUNO Workshop, LLR, Ecole Polytechnique
Opticks : GPU Optical Photon Simulation for Particle Physics with NVIDIA OptiX
env/presentation/opticks_gpu_optical_photon_simulation_oct2016_chep.html (11 Oct 2016, CHEP Conference, San Francisco)
CHEP 2016 Conference
Opticks : GPU Optical Photon Simulation for Particle Physics with NVIDIA OptiX
env/presentation/opticks_gpu_optical_photon_simulation_jul2016_weihai.html (19 July 2016, Weihai, China)
Shandong University Particle Physics Summer School
Opticks : Optical Photon Simulation for Particle Physics with NVIDIA OptiX
env/presentation/opticks_gpu_optical_photon_simulation_may2016_lecospa.html (16 May 2016, NTU, Taipei)
Seminar, Leung Center for Cosmology and Particle Astrophysics (LeCosPA) of the National Taiwan University (NTU)
Opticks : Optical Photon Simulation for Particle Physics with NVIDIA OptiX
env/presentation/opticks_gpu_optical_photon_simulation_april2016_gtc.html (April 2016, GPU Technology Conference, San Jose)
Opticks : Optical Photon Simulation for Particle Physics with NVIDIA OptiX
http://on-demand.gputechconf.com/gtc/2016/video/S6320.html GTC Talk Video (25 mins)
env/presentation/opticks_gpu_optical_photon_simulation_march2016.html (March 2016, DYB Meeting, Beijing)
Opticks : GPU Optical Photon Simulation
Includes single PMT validation
env/presentation/opticks_gpu_optical_photon_simulation.html (Jan 2016, JUNO Meeting, Xiamen)
Opticks : GPU Optical Photon Simulation
env/presentation/opticks_gpu_optical_photon_simulation_psroc.html (Jan 2016, PSROC Meeting, Kaoshiung)
Opticks : GPU Optical Photon Simulation
Same ground as Xiamen talk, but aiming at more diverse audience
env/presentation/optical_photon_simulation_with_nvidia_optix.html (July 2015, JUNO Meeting, Beijing)
Optical Photon Simulation with NVIDIA OptiX
Why not Chroma ? Intro to OptiX. Developments needed to adopt it.
env/presentation/gpu_optical_photon_simulation.html (Jan 2015, JUNO Meeting, Guangzhou)
200x Faster Optical Photon Propagation with NuWa + Chroma ?
Getting Chroma operational with Dayabay Geometry, Runtime bridge development
env/presentation/gpu_accelerated_geant4_simulation.html (Jan 2015, DYB Meeting, Hong Kong)
GPU Accelerated Geant4 Simulation with G4DAE and Chroma
Similar to Guangzhou talk, covers GPU generation of DYB Cerenkov and Scintillation photons
env/presentation/g4dae_geometry_exporter.html (Sept 2014, Geant4 Meeting, Okinawa)
G4DAE : Export Geant4 Geometry to COLLADA/DAE XML files
cd ~/simoncblyth.bitbucket.io git push git push github
tun # open tunnel N # connect to workstation ssh W # connect to web server cd simoncblyth.bitbucket.io git pull