Embedded Files
2/15
Reviewed and relearned Lorentz transformations (especially of energy)
Did basic particle + Feynman diagrams for production of top + anti-top in hadron collisions, i.e., underlying process involves partons (gluons, quarks)
In initial state
Proved that a particle of Mass A decays into two particles of Masses B and C where
(Mass A > Mass B + Mass C)
2/22
Found the Energy of the bottom quark in the rest frame of the Top Quark as function of Masses of Top Quark, Bottom Quark and the W boson using simple energy-momenta conservation (Radicals need to be shared between two sides and then squared)
now to calculate the energy of bottom quark in the lab frame we also assumed the bottom quark is massless and energy of the top quark has a constant boost with the angle distribution being flat over cos(θ). We found that this would give the energy of the bottom quark and Emin and an Emax which is a rectangle (a flat distribution between those two values). Rectangle widens as the boost increases with the Emax going to infinity faster than Emin can go to 1.
2/29
We now assumed the bottom quark had mass to the energy and the momentum became -
After we found that when plotted on the linear scale, the graph is a rectangle we couldn’t really find a peak of the distribution. So we plotted it on a scale where the x axis was the log of energies and then we could see that the peak lies at the Er (mass of the bottom quark) and it the Geometric mean of the Emin and Emax values.
We also mathematically proved that the lower endpoint has to be below Er for the plots to make some sense and give us reliable answers.
3/7
Since we had assumed that the boost is constant in the previous steps, now to get the full actual picture we assume that the boost is non constant and has an arbitrary distribution depending on the parent particle. This was relatively simple to visualize because you see it as stacking up of rectangles of different widths and we get out of it the ever popular Gaussian distribution (for a massless bottom) and an ever so slightly tail-heavy distribution (for a massive bottom)
This rest mass is always inside between the Emin and Emax. So the peak of this distribution is at the Erest.
Now we also conceptualized that at a critical boost the Emin would exceed the Er and cause the peak to shift and cause the theory to break down. We calculated that this boost would have to be 0.9999976 (5 9’s). And the energy required for the top quark to have this boost would be 168Tev so for now this is not a problem (hopefully it should be one in the future!)
3/21
Until now, we had just assumed and relied on intuition that the peak would definitely lie at the rest mass of the bottom quark. Now we have to mathematically prove this. We had to integrate over a function of the Energy and a function of gamma with the lower limit being some other function of gamma. We proved this using the derivative and the second derivative tests. (Very long and drawn out proof – please refer to latex file titles “AgashePeakProof” for detailed step by step explanation – took us a lot of time)
We were also asked to use Pythia to generate events and confirm our theory so we also began to look at that.
3/28
Pythia would give us little to no top quark events on a regular HardQCD scatter so we turned that off and put in Top:all on (pythia.readString("Top:all = on");
) to get only top quark processes.
Now I wrote a couple if statements to detect for bottom quarks and find their energies.
Basically we were looping over the event and if we found a bottom quark, we put it into a histogram.
We got nothing(??...?)
<code>
for (int i = 0; i < pythia.event.size(); ++i) { // loop over all particles in the event
if (pythia.event[i].id() == 5){ //if it is bottom quark
top14->Fill( pythia.event[i].e() ); //add energy of bottom to the graph
</code>
4/4
After meeting with professor. Eno, I realized two important things
1. Check for the absolute value of the quark id (fabs(6))
2. We not only need random bottom quarks but we also need to find bottom quarks whose mothers were top quarks and then to verify that make sure that the event only had one bottom quark and then add it to the histogram.
<code>
for (int i = 0; i < pythia.event.size(); ++i) { // loop over all particles in the event
if (pythia.event[i].id() == fabs(6)){ //if it is top quark
top14->Fill( pythia.event[i].e() ); //add energy of bottom to the graph
nBottom=nBottom; //count
APID[nBottom-1]=pythia.event[i].id();
}
}
t1->Fill();
}
</code>
4/11
We initially thought that this was Pythia not decaying top into bottom but we realized we were looking for the wrong bottom quarks. After figuring it out and running the code for 10000 events at 14 TeV we got a pretty decent graph.
So professor agashe wanted to refine the plot more so he asked us to do it for 100000 events. I tried and it gave me an putty idle timeout. Professor Jabeen suggested since the events were just randomly generated - we could do 10000 events 10 times and overlay the graphs.
Phillip worked on figuring out how to latex all our math to make it look legible and neat.
4/18
This week I had fun with Pythia and made tons of graphs at hypothetically high energies, super low energies, and anywhere in between. (Also looked for dark matter (for dark fun)). Plus learned cool tricks in pythia to overlay graphs and change colors of the histograms.
Now we both dove into Tex and as phillip wrote out the Tex files I compiled them into a paper and started adding more stuff to it.
Final Result -
energy is 1TeV (Blue)
en is 14 TeV (Red)
ener is 100 TeV (Black)
Questions to ponder -
1. Could this work for BSM particles?
2. Why is there such drastic difference between 1TeV and 14TeV whereas very little between 14TeV and 100TeV even though the factor of 10 remains the same?
Cluster Maintenance -
Madgraph -
Figured out what was wrong with Madgraph causing the cluster to overflow and use of up all resources.
The main problem we encounter while using madgraph on large computing clusters is that madgraph tends to use up all the available cores to it and hence causing overflows if any additional jobs come the cluster’s way when it running madgraph.
To solve this problem, after you generate a process, we need to modify the ‘emacs me5_configuration.txt’ file in the following way,
./bin/mg5_aMC
import model heft
generate p p > h, h > a a //generates higgs using heft
output higgs_data //outputs to higgs_data folder
quit //madgraph quit
cd higgs_data //go into higgs data folder
cd Cards
emacs me5_configuration.txt ;
nb_core =1 ;change from # nb_core = None so we don't use up all available cores in the server
Grid-0-3 Config
1. Check if the files on the machine to cleaned to make a gridnode has any files on them. Check if those files have a backup (most likely in /data/users/root/oldGN)
2. Confirm if we have all the files in the backup, use the find command to search for files by name
3. Before doing anything, figure out the purpose of the node – a GridFTP node or a SE – then puppetize the config files
4. Write (or use if already exists) a gridnode.yaml files with the appropriate packages (more on the packages later - Margarita is working(?) on gridnode.yaml file)
5. Check Firewall rules, security protocols
6. Find an appropriate partition table for the machine (Use the omreport system summary to find out the type of machine) on Foreman
7. Double confirm the backup
8. Kickstart and have FreeIPMI on it to have sensor monitoring, system event monitoring, power control features
9. Finally, install gridFTP (have GUMS authorization and gridFTP files)
Questions to ponder over
Will this gridnode need a Public IP?
Can we have more than one Bestman? (Probably not but still not sure)
Commands –
Omreport system summary - Use it to view a comprehensive summary of software and hardware components currently installed on the system.
Organizing Hardware Data - Me and Tal worked on making the hardware page look more legible by organizing the clutter of information about the various hardware used by the Cluster.
Attended all meetings via Skype or in Person and got to learn the very basics of sysadmining.
https://drive.google.com/drive/u/0/folders/0B5_rRMVBy6zabHZwb0ZDTmhoQnc
Page updated
Report abuse