Release Notes for Monte Carlo eXtreme v2017.7

Code name: Dark Matter, released on July 7, 2017

Click this link to download MCX v2017.7

Acknowledgement: This software release is made possible with the funding support from NIH/NIGMS under the grant number R01-GM114365.

1. Introduction
2. About this release
3. What's new compared to v2017.3
4. System requirements
5. Reference

1. Introduction

Monte Carlo eXtreme, or MCX, is a fast Monte Carlo software package for photon transport simulation in 3D heterogeneous media. It uses Graphics Processing Units (GPU) for massively parallel simulations and offers hundreds times acceleration compared to a traditional single-threaded CPU-based simulation.

This release fully supports NVIDIA Pascal GPUs. The speed comparisons between different generations of NVIDIA GPUs can be found at

http://mcx.space/gpubench/

2. About this release

MCX v2017.7 (code named "Dark Matter") is a Release Candidate for MCX v1.0. It is a result of numerous improvements developed over the past year, and contains a list of new features and bug fixes. If you are using a previous release, you are strongly recommended to upgrade immediately.

In this new release, the most notable update is the significantly improved MCXStudio - a cross-platform graphical user interface. The new MCXStudio became the single unified GUI for all 3 developed MC simulation packages (MCX/MMC/MCXCL), and is packaged with numerous features - including visual domain designer, 3D rendering of the results, built-in media, detector and source designer, remote execution and many more. A list of video tutorials on how to MCXStudio to design complex simulations can be found at http://mcx.sourceforge.net/cgi-bin/index.cgi?Demo/MCXStudio

In addition, the GPU code was heavily optimized to achieve improved speed and robustness. The new MCX supports all NVIDIA GPU architectures from Fermi to Volta. A number of bugs were fixed, including the crash of MCXLAB when using multiple GPUs, and errors previously on Pascal GPUs, and the incorrect progress bar feature on Windows.

Please visit our website (http://mcx.space) for more detailed documentation, demos and tutorials.

3. What's new compared to v2017.3

Compared to the previous release (version v2017.3, released in Apr. 2017), MCX v2017.7 gains the following new features and bug fixes:

  1. !!key!! much improved MCXStudio: simulation designer,volume designer,more options,copy and paste sessions
  2. !!key!! MCXStudio supports using remote GPUs
  3. !!key!! MCXStudio supports MCX, MMC and MCXCL
  4. !!key!! windows registry fix, fix memory leak
  5. !!key!! user specified scattering events to disable anisotropy g to speedup computation
  6. !!key!! save mc2 file under rootpath if --root is defined
  7. !!key!! add RPM and DEB packaging scripts, add root-level makefile
  8. automatically build mcxstudio and include the binary in the mcx nightly build
  9. new source type: "pencilarray" for an array of pencil beams
  10. save mc2 file under rootpath if --root is defined

In the meantime, a number of critical bugs were fixed:

  1. fix windows compilation bug for Pascal cards
  2. fix matlab crash when using multiple GPUs
  3. fix windows progress bar issue

Pre-compiled MCX are provided for Windows (64 bit), Linux (64bit) and Mac OS (64bit). In the case of MCXLAB, mex files for both Matlab and Octave on these platforms are provided. All binaries have been tested on Fermi/Kepler and Maxwell GPUs.

The provided binaries require a Fermi (Compute Capability 2.0) or newer GPU. If you have an older GPU (CC 1.0 or 1.1), you will have to recompile mcx using "make fast".

The detailed change logs can be found in the ChangeLog and Github commit history pages.

4. System requirements

To install MCX version v2017.7, you need

  • a CUDA-enabled graphics card made by NVIDIA, a full list of supported cards can be found here
  • a computer running GNU/Linux, Windows or Mac OS

The CUDA toolkit is no longer required in this release, however, if you run into CUDA errors, please download the latest CUDA driver, you can download from here

In this release, all precompiled binaries, including both mcx executables and mcxlab mex files, have built-in CUDA run-time libraries via static linking. Therefore, downloading/setting CUDA toolkit and the run-time librarie files (cudart.dll/libcudart.so/libcudart.dylib) are no longer needed.

However, if you run into CUDA errors, please first try to update your NVIDIA graphics driver to the latest version

http://www.nvidia.com/Download/index.aspx?lang=en-us

If the latest graphics driver still can not solve the problem, please download the "developer driver" for your GPU. You may download the developer driver as part of the CUDA Toolkit installation package.

https://developer.nvidia.com/cuda-downloads

To use MCXLAB v2017.7 in GNU Octave, you must install the following:

  • GNU Octave
  • libblas, libgfortran and libhdf5

Be aware, if you have a Maxwell GPU (GTX 980Ti and 980) and plan to run MCX on it, please first test the benchmark script "run_benchmark1.sh" or "run_benchmark1.bat" under the mcx/example folder. You are expected to see ~29,000 photon/ms for 980Ti and 20,000 photon/ms for 980. If your simulation speed is around 1,200 to 1,500 photon/ms, that means you are impacted by a bug in the CUDA driver. Please recompile MCX using CUDA 7.0 or 6.5, update your NVIDIA driver to version 375 or later. For details, please see

https://devtalk.nvidia.com/default/topic/925630/cuda-programming-and-performance/cuda-7-5-on-maxwell-980ti-drops-performance-by-10x-versus-cuda-7-0-and-6-5/1

5. Reference

Qianqian Fang and David A. Boas, "Monte Carlo Simulation of Photon Migration in 3D Turbid Media Accelerated by Graphics Processing Units," Opt. Express, vol. 17, issue 22, pp. 20178-20190 (2009)

Powered by Habitat