MCX Studio: скачать старые версии

MCX/MCX-CL/MMC v2026 (3.0) - a milestone release
- [Feature] adjoint and frequency-domain (RF) forward simulations and Jacobians in MCX, MCX-CL and MMC
- [Feature] AMD GPU support in MCX through ROCm/HIP
- [Feature] MCX Studio 2, a new graphical interface with a built-in 3-D viewer
- [Feature] implicit MMC on OpenCL and CUDA, all source types in the OptiX-based RT-MMC
- [Feature] exact fixed-point output sums (make fixed), -F bnii binary JSON output
- [Feature] a fixed seed now simulates the same photons on every OS
- [Feature] Redbird and Iso2Mesh bundled with the MATLAB toolboxes
- [Bug fix] long-standing OpenCL GPU memory leak, replay normalization, and many fixes found by unit tests and fuzzing

New Features
- [Feature] initial i18n support for international languages in mcx binary
- [Feature] mcxstudio support for French, German, Spanish, Italian, Japanese, Korean, Portuguese and Chinese
- [Feature] choosing language in windows installer
- [Feature] porting additional matlab functions to Python
- [Feature] multiple source replay
- [Bug fix] fixing mcxstudio bug to render column major output data

New Features
- [Feature] initial i18n support for international languages in mcx binary
- [Feature] mcxstudio support for French, German, Spanish, Italian, Japanese, Korean, Portuguese and Chinese
- [Feature] choosing language in windows installer
- [Feature] porting additional matlab functions to Python
- [Feature] multiple source replay
- [Bug fix] fixing mcxstudio bug to render column major output data

New Features
- [Feature] New flag -N/--net to browse simulations from NeuroJSON.io
- [Feature] Read JSON input file from stdin (standard input) using pipe
- [Feature] New demo script demo_mcxlab_replay_traj.m to plot trajectories in replay
- [Critical Bug fix] Fix incorrect partial path lengths for long-path photons, (fangq/mcx\#222)
- [Bug fix] Further updates to the handling of low absorption medium (fangq/mcx\#164)
- [Bug fix] A bug fix related to multi-source simulation to properly return the source ID (fangq/mcx\#217)

New Features
- [Feature] User-defined photon launch angle distribution (fangq/mcx\#13)
- [Feature] Simulate multiple sources in a single session (fangq/mcx\#163)
- [Feature] Per-voxel mua/mus/g/n format support (fangq/mcx\#203)
- [Critical bug fix] Fix double-multiplication of pattern launched weight, (fangq/mcx\#212)
- [Bug fix] Apply (fangq/mcx\#41) like 2x-float-buffer for dref accumulation, fix (fangq/mcx\#195)

New Features
- Introduced Split voxel MC (SVMC) to accurately model curved boundaries
- GPU polarized light modeling (Stokes) with 900x speedup
- RF replay to build frequency-domain Jacobians for amplitude and phase
- Web-based MCX Cloud platform including web-client and server scripts
- pymcx - an mcxlab-like Python module for running MCX simulations in Python
- Added Debian/Ubuntu packages for easy installation
- Added a unified command line interface, photon, to call mcx/mcxcl/mmc
- Fine-tuned Windows installer
- Extensively developed Github Action for automated building and packaging of mcx
- Adopted standardized NeuroJSON JNIfTI and JData formats to ease data exchange
- New source types include hyperboloid and ring (annulus,annulus sector)