cullanhowlett / l-picola Goto Github PK

I do not know whether the momentum corrections do make a difference or not, but the line https://github.com/CullanHowlett/l-picola/blob/master/src/main.c#L810 is obviously wrong: sumx is used for all three dimensions, where sumx/sumy/sumz should be used instead.

Hi @cullanhowlett,

I ran L-Picola was able to compile and run L-Picola for some simulations following all the procedures from the documentation. However, for the parameters bellow I found this problem:

gsl: qag.c:248: ERROR: roundoff error prevents tolerance from being achieved

I used these parameters for configuration file (LH0004.txt):

NumFilesWrittenInParallel 1
UseCOLA 1
Buffer 2
Nmesh 512
Nsample 512
Box 1000000.0
Init_Redshift 9
SphereMode 0
WhichSpectrum 1
WhichTransfer 0
StepDist 0
DeltaA 0
nLPT -2.5
UnitLength_in_cm 3.085678e21
UnitMass_in_g 1.989e43
UnitVelocity_in_cm_per_s 1e5
InputSpectrum_UnitLength_in_cm 3.085678e24
OutputRedshiftFile output_redshifts.txt
Omega 0.1433
OmegaBaryon 0.06347
OmegaLambda 0.8567
HubbleParam 0.6127
Sigma8 0.7699
PrimordialIndex 1.1501
FileBase LH0004
OutputDir LH0004
Seed 4
FileWithInputSpectrum Pk_mm_z=0.000.txt
FileWithInputTransfer Pk_mm_z=0.000.txt

with Pk_mm_z=0.000.txt and output_redshifts.txt attached.

LH0004.txt
Pk_mm_z=0.000.txt
output_redshifts.txt

I tested on two computers (on a cluster and on my laptop). The logs are here:
slurm-5492306.txt
log.txt

Also, I compiled L-PICOLA with these flags (and changed some paths properly):

MEMORY_MODE
PARTICLE_ID
GADGET_STYLE
TIMING

and I disabled ONLY_ZA because I wanted 2LPT.

Let me know if you need more information about this. Thank you!