NIF Tour and AI

Hello World!

This looks unnervingly like straight-up propaganda

This week I went on a tour of the National Ignition Facility! Unfortunately we weren’t allowed to take pictures or videos, so all the pictures that I will be showing you will be from the internet. The good news is that there is a wealth of pictures on the internet because both Tron and Star Trek: Into Darkness were filmed there, which makes sense because of the sheer wealth of sciency-machiney stuff in the room. They also gave us a pamphlet about NIF to give to any young kids we knew. It looks quite similar to ones used for military recruiting, because apparently the government pamphlet makers can’t make anything else. It explains things pretty will though!

Luckily our group leader was good friends with the people in charge of giving the tours, so we were allowed to go into the target chamber! Typically only congresspeople get to tour that area. The amount of knobs and buttons was so overwhelming that I am surprised I managed to force myself not to touch anything! At one point I asked the tour leader what would happen if I turned this one knob. It had a metal dog tag with words I didn’t understand on it! He told me that first, he would tackle me to the ground, and second, the laser area chamber would be filled with argon gas. So I didn’t turn the knob, but it was difficult. Inside the chamber is a small object (a few mm in width) called a hohlraum. Inside that hohlraum is whatever object needs to be energized. That object is then heated up and vaporized, and nearly all the equipment near the hohlraum is destroyed. Each experiment has its own diagnostics equipment tailored to whatever the scientists are trying to figure out, which are hopefully not destroyed.

laser chamber

You can probably see that the floor is made of glass. The glass was installed because in Star Trek: Into Darkness, they wanted to see what was going on downstairs. After they were done filming, they kept it. I was a bit surprised to learn that the 192 laser beams that converge on the hohlraum start their life as infra-red, because infra-red is actually pretty weak. But apparently it’s because a stronger beam would break all the lenses it has to go through, so it isn’t strengthened until later. Even then though, it’s only strengthened into ultra-violet light. When it enters the little hohlraum it becomes X-Rays. Those of you familiar with light physics will note that X-Rays are not the most powerful beam of light that humans can create – those are gamma rays. While gamma rays are sometimes created by shooting the laser (the plus-sign thing in the target chamber is for detecting them) that is not the intention of NIF. This demonstrates something very important: NIF is not the highest intensity laser in the world. That is Vulcan Laser in the UK. It’s not most powerful laser in the world. That’s LFEX in Japan. NIF is the most energetic laser in the world. The wavelengths aren’t necessarily the most high intensity, but the total amount of energy that goes to the hohlraum when the beams are shot is the most energy shot by any laser in the world. It’s also the hottest and most high-pressure thing in the solar system, including the core of the sun. Which if you think about it, is very American. We may not have the most powerful or most efficient or even highest intensity laser, but we'll be damned if we can't build a bunch of power banks and just pump as much raw power into our laser as we possible can.

laser chamber

After that we went over to the laser bay, where the 192 beams that are shot into the target chamber begin their life. The laser bays were enormous, which is to be expected since they ultimately have to produce around two million joules. They didn’t have as much complex machinery exposed than the target chamber, but it was amazing to be in the immediate vicinity of so much raw power. It’s actually a bit of a strange feeling — but then again I think windmills are really neat so maybe I’m just a fan of power generators. It was also a bit terrifying to discover that the code I was working with all summer controlled them!

On Thursday, I presented my poster at the poster symposium!

my poster

We each were judged by two different LLNL workers. The first judge seemed to like my poster quite a bit. The second liked my poster, but was very surprised at a few things. Firstly, he was surprised to hear that we hadn’t migrated to Java 11 already. I told him that I was a bit surprised too, but better late than never. Then I told him that I was writing unit tests for the code. Unit tests are basically tests that make sure that the code works the way you expect it. Whenever you alter the code, you run the unit tests to make sure no functionality has changed. There is a school of thought called Test Driven Development that says that you should write unit tests before code, and it’s beginning to be adopted by many companies, even ones such as Google! Even if you don’t follow TDD, you still should write many unit tests. Ideally you want all of your code to have unit tests, but realistically it rarely gets over 95%. The codebase I’m working on has 10% unit test coverage max. I explained our situation to the judge, and he was horrified. He understood it wasn’t my fault, because I was only there for twelve weeks, but he seemed like he wanted to run into NIF and start writing unit tests for all six million lines of code right that second.

I didn’t win, which was to be expected. The winners are typically PhD students working on very glamorous topics and while code updates are necessary, they aren’t particularly glamorous. For example, one of the winners studied methods of tagging molecules so that you can use aerosols to figure out how air travels in a subway system. That way, if there was ever a biological attack on a subway the US would be prepared. Another guy figured out a better way to do do partial differential equations. I went to the awards ceremony to learn about what the symposium winners were studying, and it was all quite interesting!

I also accidentally got a new project. Recently, I’ve been a bit frustrated at what I am working on at the lab. We finished the project much quicker than expected and a lot of the time I’m just studying code so I can code better in the future. So I talked to Mikhail, the leader of our section, and he gave me a new project! Now, I am automating the update process!

This is not the last time we’re going to have to update our Java codebase. It would be much easier if we could have a quicker and easier way to update it. So he introduced me to Java templating engines and source transformations, and I’m going to work on using those to make the updating process easier!

May you be ever victorious in your future endeavors!