Bytecode for transfer

From Apertium
Revision as of 23:50, 27 February 2010 by Jacob Nordfalk (talk | contribs)
Jump to navigation Jump to search

Currently transfer is the bottleneck in Apertium, processing here takes 95% CPU. This is because the transfer file is being interpreted (tree walking of the XML in the transfer t1x file) instead of being compiled into machine code.

The Java transfer bytecode compiler converts arbitrarily complex transfer files into Java source code, which is then compiled into platform-indepent bytecode.

During transfer the Java Virtual Machine will convert the most used part (the 'hot spots') into machine code.

This enables

  • Faster transfer (currently factor 5) of a corpus
  • Debuggable transfer (using a Java development tool, for example Netbeans, you can step thru the transfer code to see exactly what is happening)
  • Validating transfer files

A concrete example: Esperanto-English

So http://apertium.svn.sourceforge.net/viewvc/apertium/trunk/lttoolbox-java/testdata/transfer/apertium-eo-en.eo-en.t1x?view=markup becomes http://apertium.svn.sourceforge.net/viewvc/apertium/trunk/lttoolbox-java/src/org/apertium/transfer/generated/apertium_eo_en_eo_en_t1x.java?view=markup

which is compiled into Java bytecode and executed with the Java JIT (Just-in-time) compiler.

Parsing /home/j/esperanto/apertium-svn/apertium/trunk/lttoolbox-java/testdata/transfer/apertium-eo-en.eo-en.t1x
// WARNING: Attribute a_np_acr is not defined. Valid attributes are: [a_nom, a_prp, a_adv, a_adj, a_vrb, a_vrb2, a_det, a_ord, a_prn, a_tns, a_nepersonaj_tempoj, a_gen, a_prs, a_nbr, a_cas, lem, lemq, lemh, whole, tags, chname, chcontent, content]
// Replacing with error_UNKNOWN_ATTR - for <transfer default="chunk">/<section-def-macros>/<def-macro n="firstWord" npar="1">/<choose>/<when>/<test>/<equal>/<clip part="a_np_acr" pos="1" side="sl">
Compiling: javac -cp dist/lttoolbox.jar transfertest/res/lttoolbox-java/testdata/transfer/apertium_eo_en_eo_en_t1x.java

Here is a speed comparison:

Interpreted transfer took 91.59 secs
bytecode compiled transfer took 15.88 secs
Speedup factor: 5.76

Further work

  • The Java code have not been optimized for speed, so perhaps the real potential speedup is 6-8, or even a higher factor, if using a mixed mode (mixing C and Java code instead of doing pure-Java).
  • Memory usage is also higher than really needed.
  • The underlying library, lttoolbox-java, is using 50% of the CPU, and there are some well known shortcomings which is easily fixable
  • The bytecode should be pulled thru an optimizer, like Soot
  • Considering that we have a full port lttoolbox, Apertium could be made to run purely on Java, enabling a wide range of platforms, i.a. Windows, phones (J2ME or Android), web pages, server systems. Only the tagger is missing for a full system.


Old contents

Adapt transfer to use bytecode instead of tree walking. This task would be write a compiler and interpreter for Apertium transfer rules into the format of an an off-the-shelf bytecode engine (e.g. Java, v8, kjs, ...).

This page is to list ideas and their pros and cons.


Java bytecode

  • Considering that lttoolbox is on its way to being ported to Java. If Java bytecode was chosen this might eventually make Apertium run on J2ME devices (only the tagger is missing for a full system).
<jacobEo> spectie: jimregan I don't know, but I suppose that Java byte would run fastest, as there have been extremely 
  much work on optimize its speed, on different platforms....
<jacobEo> spectie: jimregan Also think in terms of some day get Apertium on a mobile phone.... then transfer in Java 
  bytecode would be the easiest thing. But if we don't at least also do Java bytecode, then we would have to write 
  a (non-Java) bytecode executor in J2ME.... pheh....
<jacobEo> spectie: jimregan Actually, if we get lttoolbox-java to work AND have Java bytecode for transfer, then we 
  instantly HAVE apertium running on phones! And also on Windows, many Unix variants, web pages, whatever can run Java bytecode.

Javascript bytecode

A Javascript engine.

External links