Monte Carlo eXtreme (MCX) - GPU-accelerated photon transport simulator
C Pascal Cuda Matlab C++ Makefile Other
Clone or download
Fetching latest commit…
Cannot retrieve the latest commit at this time.

README.txt

---------------------------------------------------------------------
                   Monte Carlo eXtreme  (MCX)
                          CUDA Edition
---------------------------------------------------------------------

Author:  Qianqian Fang <q.fang at neu.edu>
License: GNU General Public License version 3 (GPLv3)
Version: 1.0 Final (v2018, Dark Matter - Final)
Website: http://mcx.space

---------------------------------------------------------------------

Table of Content:

I.    Introduction
II.   Requirement and Installation
III.  Running Simulations
IV.   Using JSON-formatted input files
V.    Using JSON-formatted shape description files
VI.   Using MCXLAB in MATLAB and Octave
VII.  Using MCX Studio GUI
VIII. Interpreting the Outputs
IX.   Best practices guide
X.    Acknowledgement
XI.   Reference

---------------------------------------------------------------------

I.  Introduction

Monte Carlo eXtreme (MCX) is a fast photon transport simulation 
software for 3D heterogeneous turbid media. By taking advantage of 
the massively parallel threads and extremely low memory latency in a 
modern graphics processing unit (GPU), this program is able to perform Monte 
Carlo (MC) simulations at a blazing speed, typically hundreds to
a thousand times faster than a fully optimized CPU-based MC 
implementation.

The algorithm of this software is detailed in the References [1,2]. 
A short summary of the main features includes:

*. 3D heterogeneous media represented by voxelated array
*. support over a dozen source forms, including wide-field and pattern illuminations
*. boundary reflection support
*. time-resolved photon transport simulations
*. saving photon partial path lengths and trajectories
*. optimized random number generators
*. build-in flux/fluence normalization to output Green's functions
*. user adjustable voxel resolution
*. improved accuracy with atomic operations
*. cross-platform graphical user interface
*. native Matlab/Octave support for high usability
*. flexible JSON interface for future extensions
*. multi-GPU support

This software can be used on Windows, Linux and Mac OS. 
MCX is written in CUDA and can be used with NVIDIA hardware
with the native NVIDIA drivers, or used with the open-source
GPU Ocelot libraries for CPUs and AMD GPUs. An OpenCL implementation 
of MCX, i.e. MCXCL, was announced on July, 2012. It supports
NVIDIA/AMD/Intel hardware out-of-box. If your hardware does
not support CUDA, please download MCXCL from the below URL:

  https://github.com/fangq/mcxcl

---------------------------------------------------------------------------
II. Requirement and Installation

Please read this section carefully. The majority of failures 
using MCX were found related to incorrect installation of CUDA 
library and NVIDIA driver. 

Please browse http://mcx.space/#documentation for step-by-step
instructions.

For MCX-CUDA, the requirements for using this software include

*. a CUDA capable NVIDIA graphics card
*. pre-installed NVIDIA graphics driver

You must use a CUDA capable NVIDIA graphics card in order to use
MCX. A list of CUDA capable cards can be found at [2]. The oldest 
graphics card that MCX supports is the Fermi series (circa 2010).
Using the latest NVIDIA card is expected to produce the best
speed. You must have a fermi (GTX 4xx) or newer 
(5xx/6xx/7xx/9xx/10xx series) graphics card. The default release 
of MCX supports atomic operations and photon detection within a 
single binary. In the below webpage, we summarized the speed differences
between different generations of NVIDIA GPUs

http://mcx.space/gpubench/

For simulations with large volumes, sufficient graphics memory 
is also required to perform the simulation. The minimum amount of 
graphics memory required for a MC simulation is Nx*Ny*Nz*Ng
bytes for the input tissue data plus Nx*Ny*Nz*Ng*4 bytes for 
the output flux/fluence data - where Nx,Ny,Nz are the dimensions of the 
tissue volume, Ng is the number of concurrent time gates, 4 is 
the size of a single-precision floating-point number.
MCX does not require double-precision support in your hardware.

To install MCX, you need to download the binary executable compiled for your 
computer architecture (32 or 64bit) and platform, extract the package 
and run the executable under the <mcx root>/bin directory. 

For Windows users, you must make sure you have installed the appropriate
NVIDIA driver for your GPU. You should also configure your OS to run
CUDA simulations. This requires you to open the mcx/setup/win64 folder
using your file explorer, right-click on the "apply_timeout_registry_fix.bat"
file and select "Run as administrator". After confirmation, you should 
see a windows command window with message 

  Patching your registry
  Done
  Press any key to continue ...

You must reboot your Windows computer to make this setting effective.
The above patch modifies your driver settings so that you can run MCX 
simulations for longer than a few seconds. Otherwise, when running MCX
for over a few seconds, you will get a CUDA error: "unspecified error".

Please see the below link for details

http://mcx.space/wiki/index.cgi?Doc/FAQ#I_am_getting_a_kernel_launch_timed_out_error_what_is_that

For Linux and MacOS users, you need to add the following lines to your
shell initialization scripts. First, use "echo $SHELL" command to 
identify your shell type. For csh/tcsh, add the following lines 
to your ~/.cshrc file:

  if ("`uname -p`" =~ "*_64" ) then
	  setenv LD_LIBRARY_PATH "/usr/local/cuda/lib64:$LD_LIBRARY_PATH"
  else
	  setenv LD_LIBRARY_PATH "/usr/local/cuda/lib:$LD_LIBRARY_PATH"
  endif
  setenv PATH "/usr/local/cuda/bin:$PATH"

and for bash/sh users, add 

  if [[ "`uname -p`" =~ .*_64 ]]; then
	  export LD_LIBRARY_PATH="/usr/local/cuda/lib64:$LD_LIBRARY_PATH"
  else
	  export LD_LIBRARY_PATH="/usr/local/cuda/lib:$LD_LIBRARY_PATH"
  fi
  export PATH="/usr/local/cuda/bin:$PATH"

to your ~/.bash_profile (for Ubuntu users, this file is ~/.bashrc).

If the path "/usr/local/cuda/lib*" does not exist on your system or
the CUDA library is not installed under this directory, you should
substitute the actual path under which libcudart.* presents.


III.Running Simulations

To run a simulation, the minimum input is a configuration (text) file,
and a volume file (a binary file with each byte representing a medium 
index). Typing the name of the executable without any parameters, 
will print the help information and a list of supported parameters, 
such as the following:

<pre>###############################################################################
#                      Monte Carlo eXtreme (MCX) -- CUDA                      #
#          Copyright (c) 2009-2018 Qianqian Fang <q.fang at neu.edu>          #
#                             http://mcx.space/                               #
#                                                                             #
# Computational Optics & Translational Imaging (COTI) Lab- http://fanglab.org #
#            Department of Bioengineering, Northeastern University            #
###############################################################################
#    The MCX Project is funded by the NIH/NIGMS under grant R01-GM114365      #
###############################################################################
$Rev::7850c3 $ Last $Date::2018-07-21 12:28:44 -04$ by $Author::Qianqian Fang $
###############################################################################

usage: mcx <param1> <param2> ...
where possible parameters include (the first value in [*|*] is the default)

== Required option ==
 -f config     (--input)       read an input file in .json or .inp format

== MC options ==

 -n [0|int]    (--photon)      total photon number (exponential form accepted)
                               max accepted value:9.2234e+18 on 64bit systems
 -r [1|+/-int] (--repeat)      if positive, repeat by r times,total= #photon*r
                               if negative, divide #photon into r subsets
 -b [1|0]      (--reflect)     1 to reflect photons at ext. boundary;0 to exit
 -B [0|1]      (--reflectin)   1 to reflect photons at int. boundary; 0 do not
 -u [1.|float] (--unitinmm)    defines the length unit for the grid edge
 -U [1|0]      (--normalize)   1 to normalize flux to unitary; 0 save raw
 -E [0|int|mch](--seed)        set random-number-generator seed, -1 to generate
                               if an mch file is followed, MCX "replays" 
                               the detected photon; the replay mode can be used
                               to calculate the mua/mus Jacobian matrices
 -z [0|1]      (--srcfrom0)    1 volume origin is [0 0 0]; 0: origin at [1 1 1]
 -R [-2|float] (--skipradius)  -2: use atomics for the entire domain (default)
                                0: vanilla MCX, no atomic operations
                               >0: radius in which use shared-memory atomics
                               -1: use crop0/crop1 to determine atomic zone
 -k [1|0]      (--voidtime)    when src is outside, 1 enables timer inside void
 -Y [0|int]    (--replaydet)   replay only the detected photons from a given 
                               detector (det ID starts from 1), used with -E 
			       if 0, replay all detectors and sum all Jacobians
			       if -1, replay all detectors and save separately
 -P '{...}'    (--shapes)      a JSON string for additional shapes in the grid
 -V [0|1]      (--specular)    1 source located in the background,0 inside mesh
 -N [10^7|int] (--reseed)      number of scattering events before reseeding RNG
 -e [0.|float] (--minenergy)   minimum energy level to terminate a photon
 -g [1|int]    (--gategroup)   number of time gates per run
 -a [0|1]      (--array)       1 for C array (row-major); 0 for Matlab array

== GPU options ==
 -L            (--listgpu)     print GPU information only
 -t [16384|int](--thread)      total thread number
 -T [64|int]   (--blocksize)   thread number per block
 -A [0|int]    (--autopilot)   auto thread config:1 dedicated GPU;2 non-dedica.
 -G [0|int]    (--gpu)         specify which GPU to use, list GPU by -L; 0 auto
      or
 -G '1101'     (--gpu)         using multiple devices (1 enable, 0 disable)
 -W '50,30,20' (--workload)    workload for active devices; normalized by sum
 -I            (--printgpu)    print GPU information and run program

== Output options ==
 -s sessionid  (--session)     a string to label all output file names
 -d [1|0]      (--savedet)     1 to save photon info at detectors; 0 not save
 -x [0|1]      (--saveexit)    1 to save photon exit positions and directions
                               setting -x to 1 also implies setting '-d' to 1
 -X [0|1]      (--saveref)     1 to save diffuse reflectance at the air-voxels
                               right outside of the domain; if non-zero voxels
			       appear at the boundary, pad 0s before using -X
 -m [0|1]      (--momentum)    1 to save photon momentum transfer,0 not to save
 -q [0|1]      (--saveseed)    1 to save photon RNG seed for replay; 0 not save
 -M [0|1]      (--dumpmask)    1 to dump detector volume masks; 0 do not save
 -H [1000000] (--maxdetphoton) max number of detected photons
 -S [1|0]      (--save2pt)     1 to save the flux field; 0 do not save
 -F [mc2|...] (--outputformat) fluence data output format:
                               mc2 - MCX mc2 format (binary 32bit float)
                               nii - Nifti format
                               hdr - Analyze 7.5 hdr/img format
 -O [X|XFEJP]  (--outputtype)  X - output flux, F - fluence, E - energy deposit
                               J - Jacobian (replay mode),   P - scattering
                               event counts at each voxel (replay mode only)

== User IO options ==
 -h            (--help)        print this message
 -v            (--version)     print MCX revision number
 -l            (--log)         print messages to a log file instead
 -i 	       (--interactive) interactive mode

== Debug options ==
 -D [0|int]    (--debug)       print debug information (you can use an integer
  or                           or a string by combining the following flags)
 -D [''|RMP]                   1 R  debug RNG
                               2 M  store photon trajectory info
                               4 P  print progress bar
      combine multiple items by using a string, or add selected numbers together

== Additional options ==
 --root         [''|string]    full path to the folder storing the input files
 --gscatter     [1e9|int]      after a photon completes the specified number of
                               scattering events, mcx then ignores anisotropy g
                               and only performs isotropic scattering for speed
 --maxvoidstep  [1000|int]     maximum distance (in voxel unit) of a photon that
                               can travel before entering the domain, if 
                               launched outside (i.e. a widefield source)
 --maxjumpdebug [1000000|int]  when trajectory is requested (i.e. -D M),
                               use this parameter to set the maximum positions
                               stored (default: 1e6)
 --faststep [0|1]              1-use fast 1mm stepping, [0]-precise ray-tracing

== Example ==
example: (autopilot mode)
       mcx -A -n 1e7 -f input.inp -G 1 -D P
or (manual mode)
       mcx -t 16384 -T 64 -n 1e7 -f input.inp -s test -r 2 -g 10 -d 1 -b 1 -G 1
or (use multiple devices - 1st,2nd and 4th GPUs - together with equal load)
       mcx -A -n 1e7 -f input.inp -G 1101 -W 10,10,10
or (use inline domain definition)
       mcx -f input.json -P '{"Shapes":[{"ZLayers":[[1,10,1],[11,30,2],[31,60,3]]}]}'
</pre>

the 2nd command above will launch 16384 GPU threads (-t) with every 64 threads
a block (-T); a total of 1e7 photons will be simulated by the first GPU (-G 1) 
with two equally divided runs (-r); the media/source configuration will be 
read from input.inp (-f) and the output will be labeled with the session 
id "test" (-s); the simulation will run 10 concurrent time gates (-g). 
Photons passing through the defined detector positions will be saved for 
later rescaling (-d); refractive index mismatch is considered at media 
boundaries (-b).

Historically, MCX supports an extended version of the input file format 
used by tMCimg. The difference is that MCX allows comments in the input file.
A typical MCX input file looks like this:

1000000              # total photon, use -n to overwrite in the command line
29012392             # RNG seed, negative to generate
30.0 30.0 0.0 1      # source position (in grid unit), the last num (optional) sets srcfrom0 (-z)
0 0 1                # initial directional vector
0.e+00 1.e-09 1.e-10 # time-gates(s): start, end, step
semi60x60x60.bin     # volume ('unsigned char' binary format)
1 60 1 60            # x voxel size in mm (isotropic only), dim, start/end indices
1 60 1 60            # y voxel size, must be same as x, dim, start/end indices 
1 60 1 60            # y voxel size, must be same as x, dim, start/end indices
1                    # num of media
1.010101 0.01 0.005 1.37  # scat. mus (1/mm), g, mua (1/mm), n
4       1.0          # detector number and default radius (in grid unit)
30.0  20.0  0.0  2.0 # detector 1 position (real numbers in grid unit) and individual radius (optional)
30.0  40.0  0.0      # ..., if individual radius is ignored, MCX will use the default radius
20.0  30.0  0.0      #
40.0  30.0  0.0      # 
pencil               # source type (optional)
0 0 0 0              # parameters (4 floats) for the selected source
0 0 0 0              # additional source parameters

Note that the scattering coefficient mus=musp/(1-g).

The volume file (semi60x60x60.bin in the above example),
can be read in two ways by MCX: row-major[3] or column-major
depending on the value of the user parameter "-a". If the volume file
was saved using matlab or fortran, the byte order is column-major,
and you should use "-a 0" or leave it out of the command line. 
If it was saved using the fwrite() in C, the order is row-major, 
and you can either use "-a 1".

You may replace the binary volume file by a JSON-formatted shape file.
Please refer to Section V for details.

The time gate parameter is specified by three numbers:
start time, end time and time step size (in seconds). In 
the above example, the configuration specifies a total time 
window of [0 1] ns, with a 0.1 ns resolution. That means the 
total number of time gates is 10. 

MCX provides an advanced option, -g, to run simulations when 
the GPU memory is limited. It specifies how many time gates to simulate 
concurrently. Users may want to limit that number to less than 
the total number specified in the input file - and by default 
it runs one gate at a time in a single simulation. But if there's 
enough memory based on the memory requirement in Section II, you can 
simulate all 10 time gates (from the above example) concurrently by using 
"-g 10" in which case you have to make sure the video card has at least  
60*60*60*10*5=10MB of free memory.   If you do not include the -g, 
MCX will assume you want to simulate just 1 time gate at a time.. 
If you specify a time-gate number greater than the total number in the 
input file, (e.g, "-g 20") MCX will stop when the 10 time-gates are 
completed. If you use the autopilot mode (-A), then the time-gates
are automatically estimated for you.


---------------------------------------------------------------------------
IV. Using JSON-formatted input files

Starting from version 0.7.9, MCX accepts a JSON-formatted input file in
addition to the conventional tMCimg-like input format. JSON 
(JavaScript Object Notation) is a portable, human-readable and 
"fat-free" text format to represent complex and hierarchical data.
Using the JSON format makes a input file self-explanatory, extensible
and easy-to-interface with other applications (like MATLAB).

A sample JSON input file can be found under the examples/quicktest
folder. The same file, qtest.json, is also shown below:

 {
    "Help": {
      "[en]": {
        "Domain::VolumeFile": "file full path to the volume description file, can be a binary or JSON file",
        "Domain::Dim": "dimension of the data array stored in the volume file",
        "Domain::OriginType": "similar to --srcfrom0, 1 if the origin is [0 0 0], 0 if it is [1.0,1.0,1.0]",
	"Domain::LengthUnit": "define the voxel length in mm, similar to --unitinmm",
        "Domain::Media": "the first medium is always assigned to voxels with a value of 0 or outside of
                         the volume, the second row is for medium type 1, and so on. mua and mus must 
                         be in 1/mm unit",
        "Session::Photons": "if -n is not specified in the command line, this defines the total photon number",
        "Session::ID": "if -s is not specified in the command line, this defines the output file name stub",
        "Forward::T0": "the start time of the simulation, in seconds",
        "Forward::T1": "the end time of the simulation, in seconds",
        "Forward::Dt": "the width of each time window, in seconds",
        "Optode::Source::Pos": "the grid position of the source, can be non-integers, in grid unit",
        "Optode::Detector::Pos": "the grid position of a detector, can be non-integers, in grid unit",
        "Optode::Source::Dir": "the unitary directional vector of the photon at launch",
        "Optode::Source::Type": "source types, must be one of the following: 
                   pencil,isotropic,cone,gaussian,planar,pattern,fourier,arcsine,disk,fourierx,fourierx2d,
		   zgaussian,line,slit,pencilarray,pattern3d",
        "Optode::Source::Param1": "source parameters, 4 floating-point numbers",
        "Optode::Source::Param2": "additional source parameters, 4 floating-point numbers"
      }
    },
    "Domain": {
	"VolumeFile": "semi60x60x60.bin",
        "Dim":    [60,60,60],
        "OriginType": 1,
	"LengthUnit": 1,
        "Media": [
             {"mua": 0.00, "mus": 0.0, "g": 1.00, "n": 1.0},
             {"mua": 0.005,"mus": 1.0, "g": 0.01, "n": 1.0}
        ]
    },
    "Session": {
	"Photons":  1000000,
	"RNGSeed":  29012392,
	"ID":       "qtest"
    },
    "Forward": {
	"T0": 0.0e+00,
	"T1": 5.0e-09,
	"Dt": 5.0e-09
    },
    "Optode": {
	"Source": {
	    "Pos": [29.0, 29.0, 0.0],
	    "Dir": [0.0, 0.0, 1.0],
	    "Type": "pencil",
	    "Param1": [0.0, 0.0, 0.0, 0.0],
	    "Param2": [0.0, 0.0, 0.0, 0.0]
	},
	"Detector": [
	    {
		"Pos": [29.0,  19.0,  0.0],
		"R": 1.0
	    },
            {
                "Pos": [29.0,  39.0,  0.0],
                "R": 1.0
            },
            {
                "Pos": [19.0,  29.0,  0.0],
                "R": 1.0
            },
            {
                "Pos": [39.0,  29.0,  0.0],
                "R": 1.0
            }
	]
    }
 }

A JSON input file requiers several root objects, namely "Domain", "Session", 
"Forward" and "Optode". Other root sections, like "Help", will be ignored. 
Each object is a data structure providing information
indicated by its name. Each object can contain various sub-fields. 
The orders of the fields in the same level are flexible. For each field, 
you can always find the equivalent fields in the *.inp input files. 
For example, The "VolumeFile" field under the "Domain" object 
is the same as Line#6 in qtest.inp; the "RNGSeed" under "Session" is
the same as Line#2; the "Optode.Source.Pos" is the same as the 
triplet in Line#3; the "Forward.T0" is the same as the first number 
in Line#5, etc.

An MCX JSON input file must be a valid JSON text file. You can validate
your input file by running a JSON validator, for example http://jsonlint.com/
You should always use "" to quote a "name" and separate parallel
items by ",".

MCX accepts an alternative form of JSON input, but using it is not 
recommended. In the alternative format, you can use 
 "rootobj_name.field_name": value 
to represent any parameter directly in the root level. For example

 {
    "Domain.VolumeFile": "semi60x60x60.json",
    "Session.Photons": 10000000,
    ...
 }

You can even mix the alternative format with the standard format. 
If any input parameter has values in both formats in a single input 
file, the standard-formatted value has higher priority.

To invoke the JSON-formatted input file in your simulations, you 
can use the "-f" command line option with MCX, just like using an 
.inp file. For example:

  mcx -A -n 20 -f onecube.json -s onecubejson

The input file must have a ".json" suffix in order for MCX to 
recognize. If the input information is set in both command line,
and input file, the command line value has higher priority
(this is the same for .inp input files). For example, when 
using "-n 20", the value set in "Session"/"Photons" is overwritten 
to 20; when using "-s onecubejson", the "Session"/"ID" value is modified.
If your JSON input file is invalid, MCX will quit and point out
where the format is incorrect.

---------------------------------------------------------------------------
V. Using JSON-formatted shape description files

Starting from v0.7.9, MCX can also use a shape 
description file in the place of the volume file.
Using a shape-description file can save you from making
a binary .bin volume. A shape file uses more descriptive 
syntax and can be easily understood and shared with others.

Samples on how to use the shape files are included under
the example/shapetest folder. 

The sample shape file, shapes.json, is shown below:

 {
  "MCX_Shape_Command_Help":{
     "Shapes::Common Rules": "Shapes is an array object. The Tag field sets the voxel value for each
         region; if Tag is missing, use 0. Tag must be smaller than the maximum media number in the
         input file.Most parameters are in floating-point (FP). If a parameter is a coordinate, it
         assumes the origin is defined at the lowest corner of the first voxel, unless user overwrite
         with an Origin object. The default origin of all shapes is initialized by user's --srcfrom0
         setting: if srcfrom0=1, the lowest corner of the 1st voxel is [0,0,0]; otherwise, it is [1,1,1]",
     "Shapes::Name": "Just for documentation purposes, not parsed in MCX",
     "Shapes::Origin": "A floating-point (FP) triplet, set coordinate origin for the subsequent objects",
     "Shapes::Grid": "Recreate the background grid with the given dimension (Size) and fill-value (Tag)",
     "Shapes::Sphere": "A 3D sphere, centered at C0 with radius R, both have FP values",
     "Shapes::Box": "A 3D box, with lower corner O and edge length Size, both have FP values",
     "Shapes::SubGrid": "A sub-section of the grid, integer O- and Size-triplet, inclusive of both ends",
     "Shapes::XLayers/YLayers/ZLayers": "Layered structures, defined by an array of integer triples:
          [start,end,tag]. Ends are inclusive in MATLAB array indices. XLayers are perpendicular to x-axis, and so on",
     "Shapes::XSlabs/YSlabs/ZSlabs": "Slab structures, consisted of a list of FP pairs [start,end]
          both ends are inclusive in MATLAB array indices, all XSlabs are perpendicular to x-axis, and so on",
     "Shapes::Cylinder": "A finite cylinder, defined by the two ends, C0 and C1, along the axis and a radius R",
     "Shapes::UpperSpace": "A semi-space defined by inequality A*x+B*y+C*z>D, Coef is required, but not Equ"
  },
  "Shapes": [
     {"Name":     "Test"},
     {"Origin":   [0,0,0]},
     {"Grid":     {"Tag":1, "Size":[40,60,50]}},
     {"Sphere":   {"Tag":2, "O":[30,30,30],"R":20}},
     {"Box":      {"Tag":0, "O":[10,10,10],"Size":[10,10,10]}},
     {"Subgrid":  {"Tag":1, "O":[13,13,13],"Size":[5,5,5]}},
     {"UpperSpace":{"Tag":3,"Coef":[1,-1,0,0],"Equ":"A*x+B*y+C*z>D"}},
     {"XSlabs":   {"Tag":4, "Bound":[[5,15],[35,40]]}},
     {"Cylinder": {"Tag":2, "C0": [0.0,0.0,0.0], "C1": [15.0,8.0,10.0], "R": 4.0}},
     {"ZLayers":  [[1,10,1],[11,30,2],[31,50,3]]}
  ]
 }

A shape file must contain a "Shapes" object in the root level.
Other root-level fields are ignored. The "Shapes" object is a
JSON array, with each element representing a 3D object or 
setting. The object-class commands include "Grid", "Sphere",
"Box" etc. Each of these object include a number of sub-fields
to specify the parameters of the object. For example, the 
"Sphere" object has 3 subfields, "O", "R" and "Tag". Field "O" 
has a value of 1x3 array, representing the center of the sphere; 
"R" is a scalar for the radius; "Tag" is the voxel values. 
The most useful command is "[XYZ]Layers". It contains a 
series of integer triplets, specifying the starting index, 
ending index and voxel value of a layered structure. If multiple
objects are included, the subsequent objects always overwrite 
the overlapping regions covered by the previous objects.

There are a few ways for you to use shape description records
in your MCX simulations. You can save it to a JSON shape file, and
put the file name in Line#6 of yoru .inp file, or set as the
value for Domain.VolumeFile field in a .json input file. 
In these cases, a shape file must have a suffix of .json.

You can also merge the Shapes section with a .json input file
by simply appending the Shapes section to the root-level object.
You can find an example, jsonshape_allinone.json, under 
examples/shapetest. In this case, you no longer need to define
the "VolumeFile" field in the input.

Another way to use Shapes is to specify it using the -P (or --shapes)
command line flag. For example:

 mcx -f input.json -P '{"Shapes":[{"ZLayers":[[1,10,1],[11,30,2],[31,60,3]]}]}'

This will first initialize a volume based on the settings in the 
input .json file, and then rasterize new objects to the domain and 
overwrite regions that are overlapping.

For both JSON-formatted input and shape files, you can use
the JSONlab toolbox [4] to load and process in MATLAB.


---------------------------------------------------------------------------
VI. Using MCXLAB in MATLAB and Octave

MCXLAB is the native MEX version of MCX for Matlab and GNU Octave. It includes
the entire MCX code in a MEX function which can be called directly inside
Matlab or Octave. The input and output files in MCX are replaced by convenient
in-memory struct variables in MCXLAB, thus, making it much easier to use
and interact. Matlab/Octave also provides convenient plotting and data
analysis functions. With MCXLAB, your analysis can be streamlined and speed-
up without involving disk files.

Please read the mcxlab/README.txt file for more details on how to
install and use MCXLAB.


---------------------------------------------------------------------------
VII. Using MCX Studio GUI

MCX Studio is a graphics user interface (GUI) for MCX. It gives users
a straightforward way to set the command line options and simulation
parameters. It also allows users to create different simulation tasks 
and organize them into a project and save for later use.
MCX Studio can be run on many platforms such as Windows,
GNU Linux and Mac OS.

To use MCX Studio, it is suggested to put the mcxstudio binary
in the same directory as the mcx command; alternatively, you can
also add the path to mcx command to your PATH environment variable.

Once launched, MCX Studio will automatically check if mcx
binary is in the search path, if so, the "GPU" button in the 
toolbar will be enabled. It is suggested to click on this button
once, and see if you can see a list of GPUs and their parameters 
printed in the output field at the bottom part of the window. 
If you are able to see this information, your system is ready
to run MCX simulations. If you get error messages or not able
to see any usable GPU, please check the following:

* are you running MCX Studio/MCX on a computer with a supported card?
* have you installed the CUDA/NVIDIA drivers correctly?
* did you put mcx in the same folder as mcxstudio or add its path to PATH?

If your system has been properly configured, you can now add new simulations 
by clicking the "New" button. MCX Studio will ask you to give a session
ID string for this new simulation. Then you are allowed to adjust the parameters
based on your needs. Once you finish the adjustment, you should click the 
"Verify" button to see if there are missing settings. If everything looks
fine, the "Run" button will be activated. Click on it once will start your
simulation. If you want to abort the current simulation, you can click
the "Stop" button.

You can create multiple tasks with MCX Studio by hitting the "New"
button again. The information for all session configurations can
be saved as a project file (with .mcxp extension) by clicking the
"Save" button. You can load a previously saved project file back
to MCX Studio by clicking the "Load" button.


---------------------------------------------------------------------------
VIII. Interpreting the Output

MCX output consists of two parts, the flux volume 
file and messages printed on the screen.

8.1 Output files

An mc2 file contains the fluence-rate distribution from the simulation in 
the given medium. By default, this fluence-rate is a normalized solution 
(as opposed to the raw probability) therefore, one can compare this directly 
to the analytical solutions (i.e. Green's function). The order of storage in the 
mc2 files is the same as the input file: i.e., if the input is row-major, the 
output is row-major, and so on. The dimensions of the file are Nx, Ny, Nz, and Ng
where Ng is the total number of time gates.

By default, MCX produces the '''Green's function''' of the 
'''fluence rate'''  for the given domain and source. Sometime it is also 
known as the time-domain "two-point" function. If you run MCX with the following command

  mcx -f input.inp -s output ....

the fluence-rate data will be saved in a file named "output.dat" under
the current folder. If you run MCX without "-s output", the
output file will be named as "input.inp.dat".

To understand this further, you need to know that a '''fluence-rate (Phi(r,t))''' is
measured by number of particles passing through an infinitesimal 
spherical surface per '''unit time''' at '''a given location''' regardless of directions.
The unit of the MCX output is "W/mm<sup>2 = J/(mm<sup>2</sup>s)", if it is interpreted as the 
"energy fluence-rate" [6], or "1/(mm<sup>2</sup>s)", if the output is interpreted as the 
"particle fluence-rate" [6].

The Green's function of the fluence-rate means that it is produced
by a '''unitary source'''. In simple terms, this represents the 
fraction of particles/energy that arrives a location per second 
under '''the radiation of 1 unit (packet or J) of particle or energy 
at time t=0'''. The Green's function is calculated by a process referred
to as the "normalization" in the MCX code and is detailed in the 
MCX paper [6] (MCX and MMC outputs share the same meanings).

Please be aware that the output flux is calculated at each time-window 
defined in the input file. For example, if you type 

 0.e+00 5.e-09 1e-10  # time-gates(s): start, end, step

in the 5th row in the input file, MCX will produce 50 fluence-rate
snapshots, corresponding to the time-windows at [0 0.1] ns, 
[0.1 0.2]ns ... and [4.9,5.0] ns. To convert the fluence rate
to the fluence for each time-window, you just need to
multiply each solution by the width of the window, 0.1 ns in this case. 
To convert the time-dependent fluence-rate to continuous-wave (CW) 
fluence (fluence in short), you need to integrate the
fluence-rate along the time dimension. Assuming the fluence-rate after 
5 ns is negligible, then the CW fluence is simply sum(flux_i*0.1 ns, i=1,50). 
You can read <tt>mcx/examples/validation/plotsimudata.m</tt>
and <tt>mcx/examples/sphbox/plotresults.m</tt> for examples 
to compare an MCX output with the analytical fluence-rate/fluence solutions.

One can load an mc2 output file into Matlab or Octave using the
loadmc2 function in the <mcx root>/utils folder. 

To get a continuous-wave solution, run a simulation with a sufficiently 
long time window, and sum the flux along the time dimension, for 
example

   mcx=loadmc2('output.mc2',[60 60 60 10],'float');
   cw_mcx=sum(mcx,4);

Note that for time-resolved simulations, the corresponding solution
in the results approximates the flux at the center point
of each time window. For example, if the simulation time window 
setting is [t0,t0+dt,t0+2dt,t0+3dt...,t1], the time points for the 
snapshots stored in the solution file is located at 
[t0+dt/2, t0+3*dt/2, t0+5*dt/2, ... ,t1-dt/2]

A more detailed interpretation of the output data can be found at 
http://mcx.sf.net/cgi-bin/index.cgi?MMC/Doc/FAQ#How_do_I_interpret_MMC_s_output_data

MCX can also output "current density" (J(r,t), unit W/m^2, same as Phi(r,t)) -
referring to the expected number of photons or Joule of energy flowing
through a unit area pointing towards a particular direction per unit time.
The current density can be calculated at the boundary of the domain by two means:

1. using the detected photon partial path output (i.e. the second output of mcxlab.m),
one can compute the total energy E received by a detector, then one can
divide E by the area/aperture of the detector to obtain the J(r) at a detector
(E should be calculated as a function of t by using the time-of-fly of detected
photons, the E(t)/A gives J(r,t); if you integrate all time gates, the total E/A
gives the current I(r), instead of the current density).

2. use -X 1 or --saveref/cfg.issaveref option in mcx to enable the
diffuse reflectance recordings on the boundary. the diffuse reflectance
is represented by the current density J(r) flowing outward from the domain.

The current density has, as mentioned, the same unit as fluence rate,
but the difference is that J(r,t) is a vector, and Phi(r,t) is a scalar. Both measuring
the energy flow across a small area (the are has direction in the case of J) per unit
time.

You can find more rigorous definitions of these quantities in Lihong Wang's
Biomedical Optics book, Chapter 5.

8.2 Console print messages

Timing information is printed on the screen (stdout). The 
clock starts (at time T0) right before the initialization data is copied 
from CPU to GPU. For each simulation, the elapsed time from T0
is printed (in ms). Also the accumulated elapsed time is printed for 
all memory transaction from GPU to CPU.

When a user specifies "-D P" in the command line, or set cfg.debuglevel='P',
MCX or MCXLAB prints a progress bar showing the percentage of completition.

---------------------------------------------------------------------------
IX. Best practices guide

To maximize MCX's performance on your hardware, you should follow the
best practices guide listed below:

=== Use dedicated GPUs ===
A dedicated GPU is a GPU that is not connected to a monitor. If you use
a non-dedicated GPU, any kernel (GPU function) can not run more than a
few seconds. This greatly limits the efficiency of MCX. To set up a 
dedicated GPU, it is suggested to install two graphics cards on your 
computer, one is set up for displays, the other one is used for GPU 
computation only. If you have a dual-GPU card, you can also connect 
one GPU to a single monitor, and use the other GPU for computation
(selected by -G in mcx). If you have to use a non-dedicated GPU, you
can either use the pure command-line mode (for Linux, you need to 
stop X server), or use the "-r" flag to divide the total simulation 
into a set of simulations with less photons, so that each simulation 
only lasts a few seconds.

=== Launch as many threads as possible ===
It has been shown that MCX's speed is related to the thread number (-t).
Generally, the more threads, the better speed, until all GPU resources
are fully occupied. For higher-end GPUs, a thread number over 10,000 
is recommended. Please use the autopilot mode, "-A", to let MCX determine
the "optimal" thread number when you are not sure what to use.

---------------------------------------------------------------------------
X. Acknowledgement

=== cJSON library by Dave Gamble ===

  Files included: mcx/src/cJSON folder

  Copyright (c) 2009 Dave Gamble

  Permission is hereby granted, free of charge, to any person obtaining a copy
  of this software and associated documentation files (the "Software"), to deal
  in the Software without restriction, including without limitation the rights
  to use, copy, modify, merge, publish, distribute, sublicense, and/or sell
  copies of the Software, and to permit persons to whom the Software is
  furnished to do so, subject to the following conditions:

  The above copyright notice and this permission notice shall be included in
  all copies or substantial portions of the Software.

  THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND, EXPRESS OR
  IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF MERCHANTABILITY,
  FITNESS FOR A PARTICULAR PURPOSE AND NONINFRINGEMENT. IN NO EVENT SHALL THE
  AUTHORS OR COPYRIGHT HOLDERS BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER
  LIABILITY, WHETHER IN AN ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM,
  OUT OF OR IN CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN
  THE SOFTWARE.
  
=== myslicer toolbox by Anders Brun ===

  Files included: mcx/utils/{islicer.m, slice3i.m, image3i.m}
  
  Copyright (c) 2009 Anders Brun, anders@cb.uu.se

  Redistribution and use in source and binary forms, with or without 
  modification, are permitted provided that the following conditions are 
  met:

  * Redistributions of source code must retain the above copyright 
  notice, this list of conditions and the following disclaimer. 
  * Redistributions in binary form must reproduce the above copyright 
  notice, this list of conditions and the following disclaimer in 
  the documentation and/or other materials provided with the distribution 
  * Neither the name of the Centre for Image Analysis nor the names 
  of its contributors may be used to endorse or promote products derived 
  from this software without specific prior written permission.

  THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" 
  AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 
  IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 
  ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE 
  LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 
  CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 
  SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 
  INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 
  CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 
  ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 
  POSSIBILITY OF SUCH DAMAGE.

---------------------------------------------------------------------------
XI. Reference

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

[2] Leiming Yu, Fanny Nina-Paravecino, David Kaeli, Qianqian Fang, \
"Scalable and massively parallel Monte Carlo photon transport simulations \
for heterogeneous computing platforms," J. Biomed. Opt. 23(1), 010504 (2018).

If you use MCX in your research, the author of this software would like
you to cite the above papers in your related publications.

Links: 

[1] http://developer.nvidia.com/cuda-downloads
[2] http://www.nvidia.com/object/cuda_gpus.html
[3] http://en.wikipedia.org/wiki/Row-major_order
[4] http://iso2mesh.sourceforge.net/cgi-bin/index.cgi?jsonlab
[5] http://science.jrank.org/pages/60024/particle-fluence.html
[6] http://www.opticsinfobase.org/oe/abstract.cfm?uri=oe-17-22-20178