Skip to main content

parallelization - ParallelTable and Precision


I'm using ParallelTable[] to calculate a function over a range of my parameters , (ω,ℓ). This seems to be working well (in terms of speed increase) except for some strange Precision issues.


I set the $MinPrecision=40 at the start of the notebook as well as my working precision, precision goal and accuracy goal for some computations later in the notebook as


 wp=$MinPrecision;

 ac=$MinPrecision-8;
 pg=wp/2;

The rest of my code basically defines some helper functions before going on to the main function (that I feed into ParallelTable). This main function uses the helper functions along with NDSolve and NIntegrate to do some computations, and it's in these that I feed the Working Precision->wp ,AccuracyGoal->ac, PrecisionGoal->pg.


I have used DistributeDefinitions on all my variables, helper functions, and the main function, and even on $MinPrecision, also all my initial conditions for NDSolve have N[...,wp] wrapped around them, but still when I use ParallelTable[...] I get errors:


 Precision::precsm. Requested precision 39.153977328439204` is smaller than $MinPrecision.

I don't get any such error when I simply use Do[...] or Table[..] on my main function, and indeed the results differ at the 33rd decimal place, which is the decimal the above number is precision too. I have no idea what this number is unfortunately, it doesn't look like any of my outputs or inputs.


I just don't understand how this could not crop up with the non-parallelized forms, but crop up with parallelized?


Minimal Example



This is much simpler than my code, but I think it still captures it and the problem:


Definitions:


M = 1;
$MinPrecision = 40;
    wp = $MinPrecision;
ac = $MinPrecision - 8;
pg = wp/2;
rinf = 15000;

The main function to be parallelized:



dGenBessE[\[Omega]_?NumericQ, l_?IntegerQ] := 
Block[{\[CapitalPhi]out, init, dinit},

init = 0.00006630728036817007679447778124486601253323 + 
6.913102762021489976135937610105907096265*10^-6 I;
dinit = -6.958226432502243329110910813935678705519*10^-7 + 
6.631148430876236565520382557577187147081*10^-6 I;

  \[CapitalPhi]out[\[Omega], l] = \[CapitalPhi] /. 
  Block[{$MaxExtraPrecision = 100}, 

      NDSolve[{\[CapitalPhi]''[r] + (2 (r - M))/(
           r (r - 2 M)) \[CapitalPhi]'[
            r] + ((\[Omega]^2 r^2)/(r - 2 M)^2 - (l (l + 1))/(
             r (r - 2 M))) \[CapitalPhi][r] == 
         0, \[CapitalPhi][rinf] == 
         N[init, wp], \[CapitalPhi]'[rinf] == 
         N[dinit, wp]}, \[CapitalPhi], {r, 30, 40}, 
       WorkingPrecision -> wp, AccuracyGoal -> ac, 
       PrecisionGoal -> pg, MaxSteps -> \[Infinity]]][[1]];
     Print["For \[Omega]=", \[Omega], " and l=", l, ": "];

     Print["Kernel ID: ", $KernelID];
 Print["Precision of init: ", Precision[init]];
 Print["Precision of dinit: ", Precision[dinit]];
 Print["\[CapitalPhi]out at 35=" , 
N[\[CapitalPhi]out[\[Omega], l][35], wp] ];
 Print["Precision of \[CapitalPhi]out: ", 
 Precision[\[CapitalPhi]out[\[Omega], l][35]]];
]

Do Paralleization prereqs:



LaunchKernels[4]
DistributeDefinitions[M, rinf, wp, ac, pg, $MinPrecision,dGenBessE];

Attempt to run it with Paralleize for two different values:


Parallelize[{dGenBessE[1/10, 0], dGenBessE[1/10, 1]}] // AbsoluteTiming

Result


For me this leads to a


Precision::precsm: Requested precision 38.95475956978393` is smaller than   $MinPrecision. Using $MinPrecision instead.

Answer




I think your problem arises because the value of $MinPrecision is not distributed correctly (If I remember correctly none of the variables in the System context are distributed automatically).


So we have to do this by hand


ParallelEvaluate[$MinPrecision = 40]

Parallelize[{dGenBessE[1/10, 0], dGenBessE[1/10, 1]}] // AbsoluteTiming

Should now work without problems.


Also, presonally I would write the last example as


ParallelMap[dGenBessE[1/10, #]&,{0,1}] // AbsoluteTiming


or


Parallelize@Scan[dGenBessE[1/10, #] &, {0, 1}]

depending on whether you need to return values or not.


Comments

Post a Comment

Popular posts from this blog

functions - Get leading series expansion term?

Given a function f[x] , I would like to have a function leadingSeries that returns just the leading term in the series around x=0 . For example: leadingSeries[(1/x + 2)/(4 + 1/x^2 + x)] x and leadingSeries[(1/x + 2 + (1 - 1/x^3)/4)/(4 + x)] -(1/(16 x^3)) Is there such a function in Mathematica? Or maybe one can implement it efficiently? EDIT I finally went with the following implementation, based on Carl Woll 's answer: lds[ex_,x_]:=( (ex/.x->(x+O[x]^2))/.SeriesData[U_,Z_,L_List,Mi_,Ma_,De_]:>SeriesData[U,Z,{L[[1]]},Mi,Mi+1,De]//Quiet//Normal) The advantage is, that this one also properly works with functions whose leading term is a constant: lds[Exp[x],x] 1 Answer Update 1 Updated to eliminate SeriesData and to not return additional terms Perhaps you could use: leadingSeries[expr_, x_] := Normal[expr /. x->(x+O[x]^2) /. a_List :> Take[a, 1]] Then for your examples: leadingSeries[(1/x + 2)/(4 + 1/x^2 + x), x] leadingSeries[Exp[x], x] leadingSeries[(1/x + 2 + (1 - 1/x...
Post a Comment
Read more

mathematical optimization - Minimizing using indices, error: Part::pkspec1: The expression cannot be used as a part specification

I want to use Minimize where the variables to minimize are indices pointing into an array. Here a MWE that hopefully shows what my problem is. vars = u@# & /@ Range[3]; cons = Flatten@ { Table[(u[j] != #) & /@ vars[[j + 1 ;; -1]], {j, 1, 3 - 1}], 1 vec1 = {1, 2, 3}; vec2 = {1, 2, 3}; Minimize[{Total@((vec1[[#]] - vec2[[u[#]]])^2 & /@ Range[1, 3]), cons}, vars, Integers] The error I get: Part::pkspec1: The expression u[1] cannot be used as a part specification. >> Answer Ok, it seems that one can get around Mathematica trying to evaluate vec2[[u[1]]] too early by using the function Indexed[vec2,u[1]] . The working MWE would then look like the following: vars = u@# & /@ Range[3]; cons = Flatten@{ Table[(u[j] != #) & /@ vars[[j + 1 ;; -1]], {j, 1, 3 - 1}], 1 vec1 = {1, 2, 3}; vec2 = {1, 2, 3}; NMinimize[ {Total@((vec1[[#]] - Indexed[vec2, u[#]])^2 & /@ R...
Post a Comment
Read more

plotting - Plot 4D data with color as 4th dimension

I have a list of 4D data (x position, y position, amplitude, wavelength). I want to plot x, y, and amplitude on a 3D plot and have the color of the points correspond to the wavelength. I have seen many examples using functions to define color but my wavelength cannot be expressed by an analytic function. Is there a simple way to do this? Answer Here a another possible way to visualize 4D data: data = Flatten[Table[{x, y, x^2 + y^2, Sin[x - y]}, {x, -Pi, Pi,Pi/10}, {y,-Pi,Pi, Pi/10}], 1]; You can use the function Point along with VertexColors . Now the points are places using the first three elements and the color is determined by the fourth. In this case I used Hue, but you can use whatever you prefer. Graphics3D[ Point[data[[All, 1 ;; 3]], VertexColors -> Hue /@ data[[All, 4]]], Axes -> True, BoxRatios -> {1, 1, 1/GoldenRatio}]
Post a Comment
Read more