Chrome, V8 and Strongtalk
There's lots to like about Google's new web browser, Chrome, which was released today. When I read the awesome comic strip introduction yesterday, however, the thing that stood out most for me was in very small type: the name Lars Bak attached to the V8 JavaScript engine. I know of Lars from his work on Self, Strongtalk, HotSpot and OOVM, and his involvement in V8 says a lot about the kind of language implementation it will be. David Griswold has posted some more information on the Strongtalk list:
The V8 development team has multiple members of the original Animorphic team; it is headed by Lars Bak, who was the technical lead for both Strongtalk and the HotSpot Java VM (as well as a huge contributor to the original Self VM). I think that you will find that V8 has a lot of the creamy goodness of the Strongtalk and Self VMs, with many big architectural improvements
I'll post more on this later, but things are getting interesting...
Update: the V8 code is already available, and builds and runs fine on Mac OS X. From the design docs, it's pretty clear that this is indeed what I was hoping for: a mainstream, open source dynamic language implementation that learned and applies the lessons from Smalltalk, Self and Strongtalk. Most telling are that the only two papers cited in that document are titled "An Efficient Implementation of Self" and "An Efficient Implementation of the Smalltalk-80 System".
The "classes as nodes in a state machine" trick for expando properties is especially neat.
The bad news: V8 is over 100,000 lines of C++.
Heh. I can see we've been spending the last 24 hours thinking many of the same thoughts. :-)
Posted by: Glenn Vanderburg | September 02, 2008 at 01:31 PM
I wonder of Urs Hölzle had any involvement. His wiki page has him as "senior vice president of operations and Google Fellow at Google." Probably too busy to get involved in VMs these days, but you never know. (wiki page: [http://en.wikipedia.org/wiki/Urs_Hoelzle], publications: [http://www.cs.ucsb.edu/~urs/oocsb/papers.shtml]).
Posted by: Steve Swerling | September 02, 2008 at 04:43 PM
More details:
* No type feedback, dynamic optimization or reoptimization. There are inline caches, but used only to speed method dispatch in the way Smalltalks have used them for decades. No PICs.
* No bytecodes or interpretation at all; it's a pure JIT, compiling straight from source the first time a function is executed. Based on history from Self, that would tend to penalize larger programs with a long startup time (although perhaps abandoning inlining speeds up code generation enough that it's not too bad). This also makes it harder to make v8 host a different language.
This is based on fairly quick perusal of the source, guided by some of Griswold's comments. It could mean that v8 will get much faster over time as that more advanced tech is added. It may also mean that v8 is targeted for Android as well and is trying to stay very small. (Or perhaps it will fork and go in both directions.)
Posted by: Glenn Vanderburg | September 02, 2008 at 08:32 PM
Meanwhile, where's Maglev already?!?!
Looking forward to some good news from the MagLev BOF at RailsConf Europe today.
Posted by: James Webster | September 03, 2008 at 01:54 AM
Hi Avi, a number of links are broken above.
Posted by: Peter William Lount | September 04, 2008 at 02:00 AM
In playing with V8 today in Chrome on web sites with heavy JavaScript the performance was stunningly faster than with Firefox. Just first impressions.
Often the pages take a while to come up but when they do they are fast.
I wonder would it take for a squeak vm to be ported?
Posted by: Peter William Lount | September 04, 2008 at 02:02 AM