<div dir="ltr">Dear Matt Budde,<div>Thank for your details comments and the direction tutorial to the source code. I will take a look and try to see what I can do. 😉</div><div>Thanks</div><div>Khieu<br><div class="gmail_extra"><br><div class="gmail_quote">On Thu, Feb 1, 2018 at 8:29 PM, Matt Budde <span dir="ltr"><<a href="mailto:matt.d.budde@gmail.com" target="_blank">matt.d.budde@gmail.com</a>></span> wrote:<br><blockquote class="gmail_quote" style="margin:0 0 0 .8ex;border-left:1px #ccc solid;padding-left:1ex"><div dir="ltr">If the
<span style="color:rgb(34,34,34);font-family:arial,sans-serif;font-size:small;font-style:normal;font-variant-ligatures:normal;font-variant-caps:normal;font-weight:400;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;background-color:rgb(255,255,255);text-decoration-style:initial;text-decoration-color:initial;float:none;display:inline"><span> </span>"comment closed surface"</span> is present in the ply file header, camino checks the locations of each walker at each step to determine if it is still intra/extra as you'd like it to be, which is computationally intensive and slow. This check is in the code somewhere around line 580 in DiffusionSimulation.java<div><br></div><div><div> <span style="white-space:pre-wrap"> </span>if(p==0.0){</div><div> <span style="white-space:pre-wrap"> </span>if(substrate.intracellular(<wbr>walker[i])!=isIntra[i]){ <span style="white-space:pre-wrap"> </span></div><div> <span style="white-space:pre-wrap"> </span>logger.severe("walker pos: "+walker[i].r[0]+" "+walker[i].r[1]+" "+walker[i].r[2]);</div><div> <span style="white-space:pre-wrap"> </span></div><div> <span style="white-space:pre-wrap"> </span>throw new LoggedException("t= "+t+" i= "+i+" has crossed. isIntra="+isIntra[i]);</div><div> <span style="white-space:pre-wrap"> </span>}</div><div> <span style="white-space:pre-wrap"> </span>}</div></div><div><br></div><div>The substrate.intracellular function first checks to see if this is a convexHull mesh (has the "closed surface"). If not, it simply returns false.</div><div><br></div><div>There's a couple of workarounds that depending on your programming experience could be used to avoid those step:</div><div><br></div><div>1) comment out the lines above to avoid checking at each timestep and recompile. Obviously, if the walker jumps between intra/extra spaces, you won't catch it. Be sure to inspect your trajectories using matlab to plot/visualize the walkers and make sure you don't get strange behavior. It should still initialize the walkers to the intra (which is slow to begin the simulation), but doesn't check at each subsequent step.</div><div><br></div><div>2) set your permeability to something very very small (1e-300)? which will also avoid this if statement.
<span style="color:rgb(34,34,34);font-family:arial,sans-serif;font-size:small;font-style:normal;font-variant-ligatures:normal;font-variant-caps:normal;font-weight:400;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;background-color:rgb(255,255,255);text-decoration-style:initial;text-decoration-color:initial;float:none;display:inline">However, because the permeability is non-zero, it may affect the results. </span>Obviously check this against the non-mesh version to see how it performs, I haven't tested it, but conceivably it could work. </div><div><br></div><div>3) if you remove the "comment closed surface" in the ply file header, it will also avoid checking at each step. In this case you can't set the initial positions to intra/extracellular. If you are only interested in "-initial uniform", that is not a problem. You could initialize the runs with uniform setting and edit the trajectory files in matlab to only save the intracellular walkers. This approach is least recommended, but exists as a possibility. You could at least give it a try to see what kind of speed improvement you get.</div><div><br></div><div>Matt (Hall) may have some better insight to this or can tell me if I've mis-stated, but option 1 has worked for me in the past to
<span style="color:rgb(34,34,34);font-family:arial,sans-serif;font-size:small;font-style:normal;font-variant-ligatures:normal;font-variant-caps:normal;font-weight:400;letter-spacing:normal;text-align:start;text-indent:0px;text-transform:none;white-space:normal;word-spacing:0px;background-color:rgb(255,255,255);text-decoration-style:initial;text-decoration-color:initial;float:none;display:inline">considerably</span>
speed up the mesh simulations.</div><div><br></div><div>-matt (budde)</div><div><br></div><div><br></div></div></blockquote></div>
</div></div></div>