Benchmarking a Go AI in Ruby: CRuby vs. Rubinius vs. JRuby vs. Truffle/Graal

The world of Artificial Intelligences is often full of performance questions. How fast can I compute a value? How far can I look ahead in a tree? How many nodes can I traverse?

In Monte Carlo Tree Search one of the most defining questions is “How many simulations can I run per second?”. If you want to learn more about Monte Carlo Tree Search and its application to the board game Go I recommend you the video and slides of my talk about that topic from Rubyconf 2015.

Implementing my own AI – rubykon – in ruby of course isn’t going to get me the fastest implementation ever. It forces you to really do less and therefore make nice performance optimization, though. This isn’t about that either. Here I want to take a look at another question: “How fast can Ruby go?” Ruby is a language with surprisingly many well maintained implementations. Most prominently CRuby, Rubinius, JRuby and the newcomer JRuby + Truffle. How do they perform in this task?

The project

Rubykon is a relatively small project – right now the lib directory has less than 1200 lines of code (which includes a small benchmarking library… more on that later). It has no external runtime dependencies – not even the standard library. So it is very minimalistic and also tuned for performance.


The benchmarks were run pre the 0.3.0 rubykon version on the 8th of November (sorry writeups always take longer than you think!) with the following concrete ruby versions (versions slightly abbreviated in the rest of the post):

  • CRuby 1.9.3p551
  • CRuby 2.2.3p173
  • Rubinius 2.5.8
  • JRuby 1.7.22
  • JRuby
  • JRuby run in server mode and with invoke dynamic enabled (denoted as + id)
  • JRuby + Truffle Graal with master from 2015-11-08 and commit hash fd2c179, running on graalvm-jdk1.8.0

You can find the raw data (performance numbers, concrete version outputs, benchmark results for different board sizes and historic benchmark results) in this file.

This was run on my pretty dated desktop PC (i7 870):

tobi@tobi-desktop ~ $ uname -a
Linux tobi-desktop 3.16.0-38-generic #52~14.04.1-Ubuntu SMP Fri May 8 09:43:57 UTC 2015 x86_64 x86_64 x86_64 GNU/Linux
tobi@tobi-desktop ~ $ java -version
openjdk version "1.8.0_45-internal"
OpenJDK Runtime Environment (build 1.8.0_45-internal-b14)
OpenJDK 64-Bit Server VM (build 25.45-b02, mixed mode)
tobi@tobi-desktop ~ $ lscpu
Architecture:          x86_64
CPU op-mode(s):        32-bit, 64-bit
Byte Order:            Little Endian
CPU(s):                8
On-line CPU(s) list:   0-7
Thread(s) per core:    2
Core(s) per socket:    4
Socket(s):             1
NUMA node(s):          1
Vendor ID:             GenuineIntel
CPU family:            6
Model:                 30
Stepping:              5
CPU MHz:               1200.000
BogoMIPS:              5887.87
Virtualization:        VT-x
L1d cache:             32K
L1i cache:             32K
L2 cache:              256K
L3 cache:              8192K
NUMA node0 CPU(s):     0-7

First benchmark: Simulation + Scoring on 19×19

This benchmark uses benchmark-ips to see how many playouts (simulation + scoring) can be done per second. This is basically the “evaluation function” of the Monte Carlo Method. Here we start with an empty board and then play random valid moves until there are no valid moves anymore and then we score the game. The performance of a MCTS AI is hugely dependent on how fast that can happen.

Benchmarks were run with a warmup time of 60 seconds and a run time of 30 seconds. The small black bars in the graph denote standard deviation. Results:

Full 19×19 playout, iterations per second (higher is better)
Ruby Version iterations per second standard deviation
CRuby 1.9.3p551 44.952 8.90%
CRuby 2.2.3p173 55.403 7.20%
Rubinius 2.5.8 40.911 4.90%
JRuby 1.7.22 63.456 15.80%
JRuby 73.479 6.80%
JRuby + invoke dynamic 121.265 14.00%
JRuby + Truffle 192.42 14.00%

JRuby + Truffle runs on a slightly modified version of benchmark-ips. This is done because it is a highly optimizing and speculative runtime that leads to bad results after warmup. This is explained here.

Second benchmark: Full UCT Monte Carlo Tree Search with 1000 playouts

This benchmark does a full Monte Carlo Tree Search, meaning choosing a node to investigate, doing a full simulation and scoring there and then propagating the results back in the tree before starting over again. As the performance is mostly dependent on the playouts the graph looks a lot like the one above.

This uses benchmark-avg, which I wrote myself and (for now) still lives in the rubykon repository. Why a new benchmarking library? In short: I needed something for more “macro” benchmarks that gives nice output like benchmark-ips. Also, I wanted a benchmarking tool that plays nice with Truffle – which means doing warmup and run of a benchmark directly after one another, as detailed in this issue.

This uses a warmup time of 3 minutes and a run time of 2 minutes. Along with the iterations per minute, we have another graph depicting average run time.

MCTS on 19×19 with 1000 playouts, iterations per minute (higher is better)
MCTS on 19×19 with 1000 playouts, average run time (lower is better)
Ruby Version iterations per minute average time (s) standard deviation
CRuby 1.9.3p551 1.61 37.26 2.23%
CRuby 2.2.3p173 2.72 22.09 1.05%
Rubinius 2.5.8 2.1 28.52 2.59%
JRuby 1.7.22 3.94 15.23 1.61%
JRuby 3.7 16.23 2.48%
JRuby + invoke dynamic 7.02 8.55 1.92%
JRuby + Truffle 9.49 6.32 8.33%

Results here pretty much mirror the previous benchmark, although standard deviation is smaller throughout which might be because more non random code execution is involved.

Otherwise the relative performance of the different implementations is more or less the same, with the notable exception of JRuby 1.7 performing better than 9.0 (without invoke dynamic). That could be an oddity, but it is also well within the margin of error for the first benchmark.

For the discussion below I’ll refer to this benchmark, as it ran on the same code for all implementations and has a lower standard deviation overall.


The most striking observation certainly is JRuby + Truffle/Graal sits atop in the benchmarks with a good margin. It’s not that surprising when you look at previous work done here suggesting speedups of 9x to 45x as compared to CRuby. Here the speedup relative to CRuby is “just” 3.5 which teaches us to always run your own benchmarks.

It is also worth noting that Truffle first was unexpectedly very slow (10 times slower than 1.9) so I opened an issue and reported that somewhat surprising lack in performance. Then Chris Season was quick to fix it and along the way he kept an amazing log of things he did to diagnose and make it faster. If you ever wanted to take a peek into the mind of a Ruby implementer – go ahead and read it!

At the same time I gotta say that the warmup time it takes has got me worried a bit. This is a very small application with one very hot loop (generating the valid moves). It doesn’t even use the standard library. The warmup times are rather huge exactly for Truffle and I made sure to call no other code in benchmark/avg as this might deoptimize everything again. However, it is still in an early stage and I know they are working on it 🙂

Second, “normal” JRuby is faster than CRuby which is not much of a surprise to me – in most benchmarks I do JRuby comes up ~twice as fast CRuby. So when it was only ~30% faster I was actually a bit disappointed, but then remembered the --server -Xcompile.invokedynamic=true switches and enabled them. BOOM! Almost 2.6 times faster than CRuby! Almost 90% faster than JRuby without those switches.

Now you might ask: “Why isn’t this the default?” Well, it was the default. Optimizing takes time and that slows down the startup time, for say rails, significantly which is why it was deactivated by default.

If I’m missing any of these magic switches for any of the other implementations please let me know and I’ll add them.

I’m also a bit sad to see rubinius somewhere between 1.9 and 2.2 performance wise, I had higher hopes for its performance with some appropriate warmup time.

Also opal is notably missing, I couldn’t get it to run but will try again in a next version to see what V8 can optimize here.

An important word of warning to conclude the high level look at the different implementations: These benchmarks are most likely not true for your application! Especially not for rails! Benchmark yourself 🙂

Now for another question that you probably have on your mind: “How fast is this compared to other languages/implementations?” See, that’s hard to answer. No serious Go engine does pure random playouts, they all use some heuristics slowing them down significantly. But, they are still faster. Here’s some data from this computer go thread, they all refer to the 19×19 board size:

  • it is suggested than one should be able to do at least 100 000 playouts per second without heuristics
  • With light playouts Aya did 25 000 playouts in 2008
  • well known C engine pachi does 2000 heavy playouts per thread per second

Which leads us to the question…

Is this the end of the line for Ruby?

No, there are still a couple of improvements that I have in mind that can make it much faster. How much faster? I don’t know. I have this goal of 1000 playouts on 19×19 per second per thread in mind. It’s still way behind other languages, but hey we’re talking about Ruby here 😉

Some possible improvements:

  • Move generation can still be improved a lot, instead of always looking for a new valid random moves a list of valid moves could be kept around, but it’s tricky
  • Scoring can also be done faster by leveraging neighbouring cells, but it’s not the bottleneck (yet)
  • a very clever but less accurate data structure can be used for liberty counting
  • also, of course, actually parallelize it and run on multiple threads
  • I could also use an up to date CPU for a change 😉

Other than that, I’m also looking over to the ruby implementations to get better, optimize more and make it even faster. I have especially high hopes for JRuby and JRuby + Truffle here.

So in the future I’ll try to find out how fast this can actually get, which is a fun ride and has taught me a lot so far already! You should try playing the benchmark game for yourselves 🙂

JRuby – Just Ruby

I wrote a blog post about my favorite Ruby implementation JRuby over in the eurucamp blog. It’s an article introducing people to JRuby highlighting the benefits of it as well as the misconceptions that exist. Why did I write that there and not here? Oh yeah, I’m organizing JRubyConf.EU this year and wanted to promote the conference a bit and the JRubyConf.EU is basically a sub set of the eurucamp team.

Go ahead and read the whole article.

Shoes Presentation from JRubyConf

So today i gave my first full time presentation at a conference – JRubyConf that is. It went well! Thanks for having me! 🙂

My presentation itself was written and presented in shoes (yes a presentation about shoes in shoes!) and you can grab it on my github repository (there are instructions there how to install/run it)  – but I thought providing a PDF with the screenshots of the presentation might be nice.  But I really encourage you to try the shoes version, you get way nicer effects there 🙂 And yes my little presentation tool doesn’t have PDF export – yet 😉

So here you can get the presentation:

Have a great week, try out shoes and most importantly Shoes on!


Shoes 4 – a progress report

My Google Summer of Code has been going on for one month now. The first commit on the shoes4 repository is almost one year and two months old. I think this is a good time to introduce shoes4 to more people and take a look and see what we have accomplished so far.


Shoes is a multi-platform gui toolkit for Ruby aimed at simplicity. The probably most well-known Shoes program is Hacketyhack, a tool to teach programming to beginners. Shoes truly is one of a kind, for instance with its very own layout mechanisms “stack” and “flow”. Personally I always wanted to write little GUI applications but always found it to be too cumbersome and frustrating. Until I met shoes. See how simple it is:

Shoes. app title: 'Hello Shoes' do
  background gradient
  para 'This is just a very basic app'
  button 'Click me' do alert 'Hello there!' end
  image ''

Screenshot from 2013-07-17 21:14:29

But this is not the only great feature of shoes, although it’s certainly one of the reasons why I have fallen in love with shoes. Another feature is packaging: You can package your shoes applications as standalone applications for the different operating systems.

Shoes 4?

Shoes4 is a complete rewrite of its predecessor keeping its features and enhancing the DSL here ant there.  “Why rewrite?”, you might ask. Well Shoes3 (or red shoes) is more of a C-project than a Ruby project. It has separate backends for each one of the major three operating systems. It’s really hard to maintain. Unfortunately there are some bugs that make the install fail on some systems, a shot at a dependency upgrade was basically a frustrating story for everybody involved, packaging (sadly) is mostly broken etc. Also there always was the dream, that shoes could be a gem. That was just hard to accomplish with shoes3, so it is an executable embedding a ruby interpreter.

Over time people have also started writing their own Shoes versions. There are a lot of them out there, by our count it’s 9 versions of shoes right now. Shoes4 is an effort for all implementers to join forces and work together so shoes can be shiny and new again. To accommodate the apparent desire of people to build their own pair of shoes we now use a new architecture.

Basically there is a DSL layer with all the elements that the user will interact with. This layer already implements quite some of the logic. However an exchangeable backend does the real heavy lifting of drawing etc. . So almost every DSL class has a matching backend implementation. For now the default backend implementation uses JRuby + SWT, but there also is a proof of concept backend in Qt. One of the advantages of SWT is that it aims to have a native look and feel, which is in the spirit of the original shoes. And already using a cross-platform GUI library takes a lot of load from our shoulders.

Screenshot from 2013-07-17 21:23:27
A very basic and yet to be polished graphic illustrating the DSL + backend approach.

Where are we at?

Here are some numbers to give you a feeling of where we are at right now:

  • we are closing in on the 1000 commits mark
  • Code coverage is at 92%
  • 62 samples (little scripts to test things, some of them real shoes programs) are known to work
  • 30 samples are not yet working / for many of them only tiny bids of implementation are missing to make them work and some of them are not scheduled to ship with 4.0 (e.g. video support)
  • 21 people have contributed code and a lot more have helped with reporting issues, trying stuff out, being helpful etc. – THANK YOU ALL! 🙂

As mentioned in the beginning there is a Google Summer of Code going on: Faraaz and me (with the help of our mentor Davor) are working hard to push Shoes4 forward. Here are some numbers concerning the first month of Google Summer of Code, we…

  • closed 46 issues
  • opened 18 issues
  • pushed 193 commits…
  • ..changing 202 files.
  • Code Climate score has improved from 3.1 to 3.3 (and there is a continued effort to refactor the other offenders)

And since we are talking about graphical things, here are some screenshots of working samples with shoes4 as of now:

The shoes4 builtin manual
A little tank game

There are a lot more samples… you can play pong, watch some fancy animations, a simple to do list etc…

One of the apparent questions is: “Can I use shoes4 right now?” The answer is: “Not yet.” We are in pre-alpha stage. While a lot of things work, some don’t. We will let you know when we release an alpha release or a release candidate for you to check out. However if you’re adventurous you can check out the master on github no matter what, feedback and reports about bugs are highly appreciated!

So what is missing? Here are a couple of bigger things that are missing as of now:

  • The span element to style parts of a text (used in a lot of samples)
  • Reliable packaging for all platforms (there is basic .jar and Mac packaging though)
  • quite some styling options for the different elements
  • The shoes console and the methods that go with it


In general the shoes community is really nice and helpful, so if you want to take a swing at something we’re happy to help and even happier about your contribution! We also have a new “Newcomer Friendly” tag to show the way to potentially good issues to get started on. Otherwise just look at the issues and comment there if you’re interested in helping out. However just trying shoes4 out, fooling around or running a couple of samples is highly appreciated and helpful as well. Just refer to the README of the shoes4 repository to get started.

Lastly I want to thank the JRuby organization and Google for giving me the opportunity to work on an open source project I love full-time.

Screenshot from 2013-07-17 21:02:08Furthermore, if you want to know more about shoes you might consider going to JRuby Conf 2013 – I will be speaking there about shoes and there are lots of other great talks as well!

Shoes on!


Accepted for Google Summer of Code working on Shoes

Hi everyone,

this is just a quick note that I’m accepted as a student of Google Summer of Code. I will be working on Shoes, more specifically Shoes4, and the JRuby organization accepted me! So this will be a very interesting path on my journey 🙂

But what does this mean? Well I will get paid to work on an open source project, that I love. Which means getting paid for my hobby, as I already am an active contributor to this project. Which also is your periodic reminder that dreams DO come true. I’m really psyched about this.

I will be working on implementing as many features as possible so that we can publish an alpha release or (hopefully) even a release candidate this year. Not promising anything though – you know what they say about software projects and release dates! 😉

I’ll also make sure to focus on features that will make Hackety Hack run on Shoes4. Hackety is the biggest Shoes project out there and I would love to see it run on a stable platform again so many people can get into programming with the help of hackety again.

I just want to take this opportunity to thank Google for running such an awesome project supporting open source. And I want to thank the JRuby team for selecting me.

Enjoy your summer, I know I will enjoy mine,