Shaina Rudman

HW4 Tutorial 

GIthub: https://github.com/ShainaRudman/HONRPHYS

-After encountering some issues in the past with pushing files to github I was able to fix this problem, the error was that I was trying to push it using the https:// address for the github repository but since we are using SSH keys it should be different (discussion: gitpush failure)

-commit and push to github using the following:

echo # HONRPHYS >> file name here.cpp 

ls

git init

git add file name here.cpp 

git status

git commit -m "test command"

git remote set-url origin git@github.com:ShainaRudman/HONRPHYS.git

git push -u origin master

-we learned to compile and execute using these 2 commands: 

g++ 

./a.out

-When typing the code in the sections for variables with numbers I am getting an error at the end saying that the parenthesis are mismatched 

-Update: If anyone has this problem in the future there are some forums that reccomend enabling TCPkeepalives and then changing the seconds between keepalives from 0 to 5. It worked when I changed it to 3. Hopefully this doesn't cause any unintended problems.

-The final part of the tutorial was difficult in my opinion- inner and outer loops using the "while" and "for" loops. First I put in the original while code (newloop.cpp) to understand the purpose of the 2 loops and what they outputted. I tried several different methods of coding translate the "while" code to "for" format- what i didn't realize is that for each "for" is an inner/outter loop and it can't all be put in one line with both variables within the for() at once (newloop2.cpp)

HW 5 Tutorial 

-logic statements fulfill a statement if a condition is met 

-the tutorial code just says n is 10

-can also say what will happen if the condition is not met 

if (x > 0)   cout << "x is positive"; else if (x < 0)   cout << "x is negative"; else   cout << "x is 0";

Pointers:

-each peice of data has a computer "address" where it is located

-can check where something is located using the & operator

cout << “The memory address of i is “ << &i;

cout << “The data stored at memory address “ << &i << “ is “ << i;

-a pointer is a variable that stores this address

to declare a point (use int)

int* p = &i;

cout << “The value of p is “ << p;

cout << “We say that p ‘points at’ the memory location referenced by address “ << &i;

-can use the * operator to tell the program to open up this location

cout << “The data stored at memory address “ << p << “ is “ << *p;

-pointer tell a location and are a value??

remember: 

chmod +x to make a .tcsh script executable 

then ./scriptname.tcsh to execute it

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

Other Notes

-Spaces and lines don't affect the code they are just for organization

-writing endl; makes it so the result is printed on another line

-cin used so that user can input something

-void means the function does not return a value

-if function does not return a value then the return statement is not neccessary

-in c++ programs cannot be written within one another (called nesting)

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

HW 6 Tutorial

-root program 

exit using: .quit

-store multiple values in arrays 

-begins with 0 index

-combing code from different files

this written at the top of the main file:

     #include "nameoffile.cpp"

open the input file using the following 2 lines:

    ifstream infile;

    infile.open("nameoftxtfile.txt");

close the input file using this ending line:

    infile.close();

after making it so that the information is incorporated in the code you have to inlcude the functions that are defined in the files that you #include "filename.cpp" at the top

Here is the main program:

/* MAIN */ #include <iostream> #include <fstream> #include"dotprod.cpp" #include"scalarmult.cpp" using namespace std;  int main(){   double vector1[3];   double vector2[3];   double scalar[1];   ifstream infile;  infile.open("vectors.txt");     infile>>vector1[0]>>vector1[1]>>vector1[2];    cout<<" Vector 1 is ("<<vector1[0]<<","<<vector1[1]<<","<<vector1[2]<<")"<<endl;    infile>>vector2[0]>>vector2[1]>>vector2[2];    cout<<" Vector 2 is ("<<vector2[0]<<","<<vector2[1]<<","<<vector2[2]<<")"<<endl;    infile>>scalar[0];    cout<<" Scalar is " <<scalar[0]<<endl;

 infile.close();

 dot_prod(vector1,vector2);

 scalar_mult(scalar, vector1);

 scalar_mult(scalar, vector2);

   return 0;

}

Here is the dotproduct function file 

/* DOTPROD.CPP */

#include <iostream>

using namespace std;

double dot_prod(double v1[3],double v2[3]) {

  double dotdot;

  dotdot = v1[0]*v2[0]+v1[1]*v2[1]+v1[2]*v2[2];

  cout<<" The dot product is "<<dotdot<<endl;

  return 0;

}

Here is the scalarmult function file 

/* Scalarmult.cpp */

#include <iostream>

using namespace std;

double scalar_mult(double s1[1], double v1[3]) {

  cout<<"the scalar multiple is ("<<s1[0]*v1[0]<<","<<s1[0]*v1[1]<<","<<s1[0]*v1[2]<<")"<<endl;

  return 0;

}

-the input file is just a text file with 3 lines for the 2 vectors and the scalar 

-getting a lot of errors for random program

-can use srand and rand functions from the <stdlib.h> library

-it's possible to make up your own variables called classes

whenever you log in)

• cd to the CMSSW_5_3_30

• cmsenv

• cd to the testRoot subdirectory

We are ready to start the root executable. 

• root –nw (note: if you forgot the instructions about X11 or Xming, you will get an error “DISPLAY not set”)

exit root by typing 

• .quit (note the  dot  .  before word quit )

Other notes:

-to execute a file on root it must end with .root, resolutions.c will not work

-execute by doing

.x filename.root

Fireworks notes:

-use cmsShow to open program

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

Homework 7

TH1F *histo1 = new TH1F("histo1","s",50,0.,50.)

-"s" is the title

-this creates 50 bins that go from a minimum of 0 and a max of 50

use arrow -> since histo1 is a pointer for line:

histo1->Draw("")

for (int i=0;i<500;i++) 

-500 is the number of entries 

TRandom r;

    gRandom->SetSeed();

this is the random element

creation of a pointer with the name histo1, pointing to a variable called TH1F (custom variable made by root to store histogram)

-all classes have data members

-There are also 3D histograms

-all classes have a creator

-Fill data members

Root also provides many math libraries such as 

• • TRandom r;

  • gRandom->SetSeed();

  • float p1=r.Gaus(10,2);

helpful root information: http://www-root.fnal.gov/root/GettingStarted/GS.htm

to save the root macro from a historgram from an outside source go to save as and save it as a .root or .C file

example root code (http://www.desy.de/~blist/summerstudents/root-intro-07-transp.pdf):

int main() {

// initialization here

//parameters

//event files

//book histograms

for( int i=0; i<events; i++){

//load event numbers

//event properties

}

}

//draw histogram

//write out historgram file

//write out info like number ofevents etc.

return 0;

}

"E0", "E1" error bars

the gRandom variable is a variable that points to a random number generator instance of the type TRandom3. Such a variable is useful to access in every point of a program the same random number generator, in order to achieve a good quality of the random sequence.

gStyle->SetOptTitle(0);     // suppress title box

The following methods are provided to generate random numbers distributed according to some basic distributions:

• Exp(tau)

• Integer(imax)

• Gaus(mean,sigma)

• Rndm()

• Uniform(x1)

• Landau(mpv,sigma)

• Poisson(mean)

• Binomial(ntot,prob)

https://root.cern.ch/doc/master/classTRandom.html (info for random number gen)

Other helpful root links:

http://www.desy.de/~blist/summerstudents/root-intro-07-transp.pdf

https://root.cern.ch/root/html518/TGraph.html

https://root.cern.ch/working-macros

http://www.nevis.columbia.edu/~seligman/root-class/RootClass2011.pdf

https://root.cern.ch/doc/master/classTF1.html

http://www-ekp.physik.uni-karlsruhe.de/~quast/Skripte/diving_into_ROOT.pdf

When using a mac/xquartz, login using: (my computer is temporarily broken)

ssh -X -Y rudman@hepcms.umd.edu

use TF1 to plot function

TF1 *f1 = new TF1("name","function", xmin, xmax);

f1->Draw();

another class in tutorial: TPad

    -a pad is contained in a canvas and can contain other pads

    -can use pad to create multipe things at once

cd() (what does this do?)

use TFormula to define a function that can then be later references

https://root.cern.ch/doc/master/classTColor.html (this link has the colors that correspond with each number)

Graphs' plotting options

Graphs can be drawn with the following options:

for example 

f1->Draw("ALP");

this will make it so each data point is connected by a line beacause of the L

When creating formulas or using TF1 must have x in all equations, even if the function is only one formula plus another, you need to add +0*x to include x

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

HW 9

2 types of Monte Carlo we are looking at: event generator and detector simulation

what does root.exe do? 

an “event” corresponds to a collision that was selected by a trigger, the data that was in the detectors at the time of the “trigger” was read out and stored together

This is the graph the appeared after doing .x toyMC.C

-writing code to analyze data in output.root

TFile (another class)

ls lists whats in the file, just like in normal linux

.h and .c files present after MakeClass("MyClass");

.L pulls .C file information into memory 

this is the result after m.Loop()

Note the "+" sign at the end of the loading command would compile the code and would make the command run faster. It will also check your code for some obvious mistakes.

result after el1_eta_>Draw();

What is an MC???

What is a .h file

 chain->Add(<file>.root) function allows us to chain together multiple files so that we are able to perform analysis on multiple files using one function

Not entirely clear on what chaining specifically does 

Resulting graphs

step 3

.L ana.C+

ana t

t.Loop()

ls

Z_ee->Draw()

step 4

It is much faster to do root -l then just root, this way that screen doesn't pop up

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

Pythia Tutorial

http://home.thep.lu.se/~torbjorn/pythia82html/Welcome.html

this website is important

go to cmssw area

git clone https://github.com/saraheno/honrpythia.git

cd honrpythia

source PYTHIA_setup.csh

make pythiaTree

./pythiaTree >& output.txt

here area the files created by donig this 

emacs pythiaTree.cc to alter it 

change this line 

  for (int iEvent = 0; iEvent < 10000; ++iEvent) {

so that it is 10000

alter the code so that the mass is stored in the tree 

declare mass as a float 

Float_t mass[10000]

then add as branch 

 t1->Branch("mass",mass,"mass/F");

how to compile after making changes 

did make pythiaTree and got many errors saying everything is undefined

emacs output.txt to run 

when i open it it says 1032 not 10000

to compile tried this 

source PYTHIA_setup.csh

make pythiaTree

./pythiaTree >& output.txt

its taking a while to load... 

after opening the output file there are only 1032 events although it says the events are created 

remove the lines and left pythia pythia but then all the graphs were empty 

after changed lines got nothing and got error in pythia

to create pT histograms 

declare double pT1 and pT2 = 0 

 create histograms 

 TH1F *massSignal = new TH1F("massSignal", "Invariant mass of two highest pT electrons for signal events", 100, 0, 2000); 

 TH1F *massDY = new TH1F("massBackground", "Invariant mass of two highest pT electrons for background events", 100, 0 ,2000);

add the if statement to det highest and 2nd highest 

(see code)

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

Meeting #1 with Jimmy (2/16/2016)

-downloaded cygwin terminal 

-atlassian copy laptop

-add path to cygwin terminal to what copies to computer (http://www.howtogeek.com/howto/41382/how-to-use-linux-commands-in-windows-with-cygwin/)

-Vlasov equations

    -linear theory

-magnetic recconection

-instability

    -tearing mode instability

Meeting #2 with Jimmy (2/23/2016)

-definition of plasma

statistical (N>>1, many particles when talk about plasma)

mobile(free to move, average KE>>PE, interparticle, particle to particle not plasma as a whole mean field)

charged(can be affected by EM forces F=q(E+vxB) )

-plasma collection behavior vs. individial plasma behavior 

T(temp)=units of energy bc multiplied by k (J/k)

-Debye Length (scale that e field is screened out by plasma)

electron has an e field associated with it (can solve by gausses law) E(r)=e/(4piepsilonnot*r^2)

impact of having tons of particle? E e/ particle experiences? 

rewrite gausses law in differential form 

only consider particles standing still 

(insert math from written notes)

system has come to an equilibrium, particles come into eq w themselves

electric force falls off faster in plasma bc "screen" eachothers charge 

-Plasma Parameter (dimensionless)

(insert math from written notes)

cgs units used so can use c 

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

Week 3: Problem

see powerpoint for solution- we are still working on part B2 and B3

-how to determine number of Euler timesteps use?

ongoing powerpoint can be found here: https://docs.google.com/presentation/d/1b_Mey4jTyaituSIRs-DHtyqb-YjgOXZnlGjkTHhAcpw/edit?usp=sharing

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

Meeting #3

-the problem

-concept of point operations

    -get number from 3 equations

-expense comes from number of particles in a plasma

    -one way: decrease the number of particles per debye cube

        problem: introd. counting noise statistics- which is bad

            if want to look at distribution, changes shape of distribution

    -other way: Vlasov Equations (insert equation here)

        -can derive by considering large number of particles and debye shielding

        -Lioville's Theorem (ch. 6 of nicholson)- in xv phase space, if have volume, evolve under equations, then volume conserved

            -assuming this gives a conservation equation (insert equation with weird D)

        -this equation still easier than so many from the problem

        -thinking more about distribution

-going to track the distribution of particles

    -where in real and velocity space

-collision operator in code- so codes don't have collision C[f], just incl operator

-resistive tearing mode relies on having a collision operator

3 more parameters!

-parameters are relatable

1. Plasma frequency

(insert math from notes on paper)

these 2 parameters- when introduce a magnetic field!!!

2. Gyro frequency (Cyclotron frequency)

-velocity vector assumes harmonic motion

3. Gyro Radius (Larmor radius)

likewise

-relate to thermal speed & cyclo. freq.

also: thermal velocity!

used these instructions to download cygwin c++ compiler: stuck at part 8,4

https://netbeans.org/community/releases/72#cygwin

use set up to add packages to cygwin incl. c++ comiler

website with input code

http://ammar-hakim.org/sj//cpp-setup-instructions.html

.lua files 

finite difference, volume, element 

numerical methods resources: 

http://www.cfm.brown.edu/people/jansh/resources/Publications/Lectures/RMMC08-I.pdf

https://en.wikipedia.org/wiki/Finite_volume_method

https://en.wikipedia.org/wiki/Finite_difference_method

https://en.wikipedia.org/wiki/Finite_element_method

http://www.math.uci.edu/~chenlong/226/FVM.pdf

http://crd.lbl.gov/assets/pubs_presos/AMCS/ANAG/A137.pdf

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

3/25 meeting

-log into xsede

-go to putty 

-log in using xsede log in 

-go to amar haakim bit bucket and clone gekyll all 

-same thing for gekyll

-go to gekyll folder 

-mkgkeall.sh and then it goes crazy 

cd $WORK 

-work directory 

run code in scratch directory 

cd $SCRATCH

work director where .exe file will be 

output in scratch 

in gekyll all, clone gekyll 

gekyll directory where all the code is that solves everything 

./mkgkeall.sh giant shell script 

if close your computer to restart installation process do this: 

go to work 

go to gekyll all 

do ./mkgkeall command 

then it will restard

right click to paste

host name:

login.xsede.org

then user name and passwork from xsede

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

results after ./mkgkeall command:

this website:

https://portal.tacc.utexas.edu/password-reset/-/password/request-reset

put tacc number and then sent password request so that able to log in directly

tg833855

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

4/12 meeting 

-never change first line 

-first line tells what the job is called (probably ddont have to change) 

- -n with number 

*request between 16 and 256 processors*

*dont ask for a prime number* 

#SBATCH -t is the time 2:00:00

runs job faster if ask for less 

#SBATCH --mail-user=jjuno@umd.edu (change this to your email) 

long line 

project directory 

change end of long line 

organize this in some fashion 

directory weibel 

mkdir

make a new folder for each simulationn

copy lewis script and this file into that folder 

-dont have to change processors or time requested 

most important thing to change is email and end of long line, so that its your input file 

other file

speeds written in terms of speed of light 

dont chane temperature mass charge

vary k0 to establish a spectrum of growth rates, between 0.01 and 10 

a0sin(kx) (a0 called pertub)

give sinusoidal pertubation 

dont change LX (size of box) 

change resolution, time stepping 

NX, resolution in x space (dont go lower than 64 or higher than 512)

nvx, etc is resolution in y space (dont go lower than 8 or higher than 64), if chnage velocity space values change all 3 not 1 at a time 

change one, change other fixed

polynomial order- change

log(string...) all this stuff prints diagnostics (for example, can pick vl vu to be greater than 1 bc then thats greater than c) 

vary phaseDecomp - tells how to do domain decomposition, dividing work amongst processors 

cuts (1,1,1,1) when multiple all together this gives number of processors requestions, so 1 in this case 

beyond that don't change anything 

initial

fields

pertubations in bz

c++  the first entry is entry 0

gkeyllall/gkeyll/scripts/gkeplot.py

moodule list (to check)

then make sure there is a pyton there

after simulation is done

type

-y (name of script)_emEnergy

emEngery is basically integral B^2/2mu0 at every point in time

plots info for every timestep

plot that as a function of time

exponential growth , linear in log space

growth e^gamma t

find what gamma is!!!

--write-history (this writes everything to a textfile)

result will be (name of file)_emEnergy.txt

file will have 2 colums

first is time, second is integral B^2?2mu0

once you run a simulation

pull text file off of stampede

plot text file

write to array time

write to em energy

plot time vs. log of integral B^2/2mu0 then find the slope!

on thursday show a single plot or assortment of plots

on the y axis- the growth rate

on the x axis- some varying parameter (any that vary)

k0 will be dispersion relation (most  interesting)

accumulate what gamma is when vary each paramter

converge to some value for some of parameters

error associated with descritization

weibel 1x3v

weibel 1x2v

retain or don't retain z velocity

pertubation in bz

sbatch to que simulation 

squeue -u tgusername 

gives status of the job

stampede.tacc.xsede.org

use tg username and new password 

mkdir weibel (new directory in gkeyll)

check sbatch _______.pbs to make sure 

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

uploaded files to the cluster by copying from my github

git clone https://github.com/ShainaRudman/HONRPHYS.git

changed name to weibel 

edit this file gkeyll_test_trunk.pbs

this file used to send jobs to the cluster

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

4/19/2016 meeting 

change t=100 to t=50 

plots of growth rates for varying parameters 

format long

matlab

test=testread('nameoffile.txt')

time = test (:,1)

mEnergy = test (:,2)

semilogy(time,mEnergy)

didn't understand that the .lua and .pbs files need to be in the scratch directiory not in work directory 

use cp to copy files to different directories (do this while in the directory of the file you want to copy) 

cp (filename.lua) /path/want/to/copy/to

squeue -u yourusername

scancel jobnumber

input file changed 

needed to change this line in the .lua original file

phaseDecomp = DecompRegionCalc3D.CartProd { cuts = {1,1,1} }

change cuts to 4 4 16 so the product is 256 

do this command in the folder where the .lua and .pbs files are 

/work/04086/tg833855/gkeyllall/gkeyll/scripts/gkeplot.py -y weibel_1X2V_Bz2 --write-history 

this creates a graph and a text file 

this is the graph created for the trial simulation where i didnt change anything, a text file is also created with 2 columns 

need to create a folder for each simulation 

copy original files to this directory 

change the .lua file name 

mv (old name) (new name) 

to remove all log files: 

rm -rf *.log 

I was having problems with creating new folders and renaming so i just kept the same directory and changed the text file name each time to designate the parameter value, and deleting all the .log and .h5 files after each run 

1. change parameter value in .lua file 

2. run simulation using sbatch command 

3. use -y command to get graph and text file

4. save graph 

5.delete all log files and h5 files

6.   rename txt file to have parameter value in name 

7. repeat with more values

**make sure you have X11 enabled** 

next need to pull files from the cluster onto my computer so that i can use matlab to make log graphs on matlab, and then find each of the slopes

10 slopes to find 

To put file in my documents

scp tg833855@login3.stampede:/scratch/04086/tg833855/weibel/weibel_1X2V_Bz2_80.txt UserSHRDocuments

for some reason this command isnt working even though it says it is