A New Age of JVM Garbage Collectors

Welcome to this blog again; throughout the past blogs we have discussed a lot of topics focused on Lisp, Clojure and computer science in general, but today is the turn of one of the greatest commodities in the contemporaneous area, garbage collection, this feature literally changed and simplified the programming process. So, in this blog I'm going to summaries and give my humble opinion on the topics discussed in Alexander Yakushev talk, A new age of JVM garbage collectors.

So, for anyone not related to programming or familiarize with garbage collection, let me start with an example:

You have two strings, "Hello" and "World",  these strings are saved in an address in memory, meaning they are taking a space in RAM; now you want to concatenate this strings, so you end up with a string containing "Hello World". Now, you don't need the previews strings, but they are still taking space in the finite and precious memory, and if this space is not free, it will be wasted, because you will never be able to use it again. So this is where the garbage collector comes in place and free that unused variables and allowed you to not bother at all and focus on the higher level stuff, cool stuff.

In older programming languages, like C, GC was not a thing, and developers have to manually free the unused space, so as you can imagine, this was a huge part in developing any program, because, if you forget to free some memory, you will have memory leaks. So, be thankful for GC.

Now that we understand the power of GC, lets try to understand how they work:

Well, they're multiple algorithms that implement GC, Mark-Sweep-Compact, Evacuation, G1 and Shenandoah. Each one of them do the same thing, collect all the unused variables and free the memory their using. We aren't going to discuss the algorithms itself, just understand the principles that they all share:

  • Check all variables.
  • Free the ones without a reference.
  • Rearrange the heap.
The difference between one another are the way they implement this three points, but what we are going to discuss is Shenandoah algorithm; this is the newer one and it brings a fascinating stuff to the table. 

Shenandoah does all the GC stuff with pauses between 1 and 50 ms, meanwhile the other ones pause the application 1 to 3 sec in average. This is a real step forward the right direction, and we have to understand that GC algorithm are all about trade-offs. 

I'm really curious where are GC going on the future. 




Comentarios

Publicar un comentario

Entradas más populares de este blog

The Secret History of Women in Coding

  All my life I have been wondering why does man discriminate women so much, like, it is because men have more physical strength? And that make us man thought that women were inferior? In my point of view that’s dumb And I hate that that mentality is still in our days. Reading this article opened my eyes to the exceptional work that women have been doing for Computer sciences all this year, amazing that in most cases man got the credit for their achievement, for example the Eniac demonstration they don’t mention the women that bring that to life. Also, the mentality that programming was a woman’s work was because it was seemed as a secretarial work, and that not man enough, so in the 90s 35% of programmers were women. Now is 13 to 14%, and I hope that the stereotype that engineering is for man don’t continue anymore. The contributions of women are amazing, to the “break point” for debugging a program, the first compiler and Cobol language. Well, if I don’t do programming, was els
Leer más

Rich Hickey on Clojure

This week was the turn of   Episode 158: Rich Hickey on Clojure  of the software engineering radio, that invited the creator of Clojure programming language, Rich Hickey, that give an explanation of the functionality of Clojure and other technical stuff about it. One of the most interesting question that was asked, is ¿Why is LIST not a mainstream language? And the answer was that List was not design to be a mainstream language, but to be used by SUPER USERS, like scientist, people that are trying to solve really hard problems, and the point was to give them the power to do so. List became really popular but lost it when AI boom disappeared.  Other thing that he mention that was fascinating was that programming is a social thing, most of the time when people outside the software engineering path think about the stereotype of the programmers they believed that we are antisocial and love the dark in our room, but the truth is that the field strongly encourage you to work side by side wit
Leer más

Dick Gabriel on Lisp

  Esta semana toco escuchar un podcast de 2008 sobre programación funcional, inteligencia artificial, el lenguaje de programación LISP y poesía, un poco de todo. El podcast de “Software Engineering Radio” llamado “Dick Gabrien on LISP”, nos pone a escuchar a (vaya la redundancia) Dick Gabriel, un científico de la computación, muy conocido por su trabajo en la programación funcional, principalmente en LISP y por su poesía; Bueno, quizá por la poesía no tanto. Dick Gabriel busco desarrollar el lenguaje funcional como una forma en la que pueda ayudar a la implementación más sencilla para la inteligencia artificial, lo cual tuvo su auge en los 50s y 60s, pero fue bajando la popularidad de LISP debido a que hay otros paradigmas de programación que un sentido empresarial, sale más barato contratar personal que conozca de Java o C. La popularidad de LISP llevo a que se crearan modificaciones específicas para el uso de cada comunidad particular, por lo que se decidió crear el CLOS (Common
Leer más