buffer error in MCXLABCL

[Alex Antrobus]

Dear Dr Fang & The community

I'm new to MCX(cl) and am learning a lot!

I am trying to run a version of the below code in MCXLABCL. The initial simulation runs fine, but I have an error when I try to run the replay. I assume it's because I'm not detecting any photons (?!)
But then why is that?!

I'm running on my inbuilt Intel(R) Iris(R) Plus GPU, so nothing fancy.

Thanks in advance :-)

- Alex

------------ CODE USED--------------------

clear all;

clear cfg;

%% Define simulation parameters

% set seed to make the simulation repeatible

cfg.seed=hex2dec('623F9A9E'); 

% Volume model being used.

VOLUME='brain';

cfg.nphoton=1e8;

load colin27_v3.mat

cfg.vol=colin27;

cfg.prop=[  0         0         1.0000    1.0000 % background/air

        0.0190    7.8182    0.8900    1.3700 % scalp

        0.0190    7.8182    0.8900    1.3700 % skull

        0.0040    0.0090    0.8900    1.3700 % csf

        0.0200    9.0000    0.8900    1.3700 % gray matters

        0.0800   40.9000    0.8400    1.3700 % white matters

             0         0    1.0000    1.0000]; % air pockets

cfg.srcpos=[75 14 73]; %Source position

cfg.detpos=[75 18 92 5]; %Detector Position

% time-domain simulation parameters

cfg.tstart=0;% Start

cfg.tend=5e-9;% Stop

cfg.tstep=2e-10;% timestep

cfg.srcdir=[0.0 1.0 0.0, 5.0];% define the source DIRECTION

cfg.issrcfrom0 = 0; % Defines the location of the origin: 0 or 1 (0 here).

cfg.isreflect=1; % enable reflection at exterior boundary

cfg.isrefint=1;  % enable reflection at interior boundary too

%% Define OUTPUT settings 

cfg.issavedet=1; % Record partial pathlength of detected photons

cfg.ismomentum=1; % Save photon momentums

cfg.issaveref=1; % Save diffuse reflectance at air voxels.

cfg.issaveexit=1; % Save photon exit postitions & directions.

cfg.replaydet=-1; % Replay photons from all detectors.

cfg.issaveseed=1;

%% GPU thread configuration

cfg.autopilot=1;% Let MCX decide how many threads are important

cfg.gpuid=1;% 

%% Run Simulation

fprintf('running simulation ...\n');

tic;

%[f2,det2]=mcxlabcl(cfg);

[fluence,detphoton,vol,seed,trajectory] = mcxlabcl(cfg);

toc;

% Everything above here runs fine. The next section throws the "Invalid

% Buffer" error

%% Replay detected photons

newcfg=cfg;

newcfg.seed=seed.data;

newcfg.outputtype='jacobian';

newcfg.detphotons=detphoton.data;

newcfg.replaydet=1;

[flux2, detp2, vol2, seeds2, traj2]=mcxlabcl(newcfg);

jac=sum(flux2.data,4); 

-------------- ERROR MESSAGE --------------------

MCXCL ERROR(61):Invalid buffer size in unit mcx_host.cpp:497

MCXLAB ERROR 61 in unit mcx_host.cpp:497: Invalid buffer size

Error from thread (0): Invalid buffer size

Error using mcxcl

MCXLAB Terminated due to an exception!

Error in mcxlabcl (line 391)

    [varargout{1:max(1,nargout)}]=mcxcl(varargin{1});

Error in colin27_mwe (line 66)

[flux2, detp2, vol2, seeds2, traj2]=mcxlabcl(newcfg);

[Qianqian Fang]

On 3/30/21 11:41 AM, Alex Antrobus wrote:

Dear Dr Fang & The community

I'm new to MCX(cl) and am learning a lot!

I am trying to run a version of the below code in MCXLABCL. The initial simulation runs fine, but I have an error when I try to run the replay. I assume it's because I'm not detecting any photons (?!)
But then why is that?!

hi Alex,

I ran your below script, the issue that you got the invalid buffer error was because your initial (i.e. baseline) simulation returns 0 detected photon, therefore, nothing to be replayed.

the reason that you did not get any detected photon was because your detector-sphere has no intersection with the outer boundary of the non-zero voxels - the only place where photon detection happens. please see my mcxpreview(cfg) screenshot attached. notice the detector sphere (green, at the occipital region) is fully buried under the skin.

for more information about how mcx detector works, please check out the below links

https://github.com/fangq/mcx/issues/102

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

http://groups.google.com/group/mcx-users/browse_thread/thread/e5e0140d7e73e4bf?hl=en

Qianqian

--
You received this message because you are subscribed to the Google Groups "mcx-users" group.
To unsubscribe from this group and stop receiving emails from it, send an email to mcx-users+...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/mcx-users/3a05e6eb-a4b9-4721-8d4b-3229ac733704n%40googlegroups.com.