I don't know how much global appeal this would hold, but I've been thinking
about doing a presentation on how Scala interacts with the JVM. Mostly to
cover common misconceptions, optimisations the JVM can make and where this
all fits into the memory model.
It might be a *little* more advanced that you planned though, as some of the
topics I had in mind were
- escape analysis
- scalar replacement
- TLAS vs stack allocation
- mutable vs mutable objects and optimisations the JVM can perform
- why object creation is cheap and the eden space
- garbage collection strategies
It might be worth trying to tie some of this stuff into the upcoming event
(June) that the LJC have planned to cover the Java memory model, assuming I
can somehow make enough time to arrange enough material!
> I don't know how much global appeal this would hold, but I've been thinking
> about doing a presentation on how Scala interacts with the JVM. Mostly to
> cover common misconceptions, optimisations the JVM can make and where this
> all fits into the memory model.
> It might be a *little* more advanced that you planned though, as some of
> the topics I had in mind were
> - escape analysis
> - scalar replacement
> - TLAS vs stack allocation
> - mutable vs mutable objects and optimisations the JVM can perform
> - why object creation is cheap and the eden space
> - garbage collection strategies
> It might be worth trying to tie some of this stuff into the upcoming event
> (June) that the LJC have planned to cover the Java memory model, assuming I
> can somehow make enough time to arrange enough material!
I'd be interested in that. It might shed light on such questions as
"do small case classes get stack allocated, and should they?" "what
simple things should I be doing or avoiding to make it easier for the
JVM to optimise?" "how can we expect the performance of our code to
change between JVM versions?"
I don't live in London. Is there any chance of recording these talks
and publishing them somewhere? Is that already happening and I missed
it?
> I don't know how much global appeal this would hold, but I've been thinking
> about doing a presentation on how Scala interacts with the JVM. Mostly to
> cover common misconceptions, optimisations the JVM can make and where this
> all fits into the memory model.
> It might be a *little* more advanced that you planned though, as some of the
> topics I had in mind were
> - escape analysis
> - scalar replacement
> - TLAS vs stack allocation
> - mutable vs mutable objects and optimisations the JVM can perform
> - why object creation is cheap and the eden space
> - garbage collection strategies
> It might be worth trying to tie some of this stuff into the upcoming event
> (June) that the LJC have planned to cover the Java memory model, assuming I
> can somehow make enough time to arrange enough material!
> Whaddya all think?
> I don't know how much global appeal this would hold, but I've been thinking
> about doing a presentation on how Scala interacts with the JVM. Mostly to
> cover common misconceptions, optimisations the JVM can make and where this
> all fits into the memory model.
> It might be a *little* more advanced that you planned though, as some of the
> topics I had in mind were
> - escape analysis
> - scalar replacement
> - TLAS vs stack allocation
> - mutable vs mutable objects and optimisations the JVM can perform
> - why object creation is cheap and the eden space
> - garbage collection strategies
> It might be worth trying to tie some of this stuff into the upcoming event
> (June) that the LJC have planned to cover the Java memory model, assuming I
> can somehow make enough time to arrange enough material!
> +1
> Sounds great!
> Is the interaction being changed in 2.8 in comparison to 2.7?
> Thanks
> Rustem
> On Apr 29, 10:07 am, Kevin Wright <kev.lee.wri...@googlemail.com>
> wrote:
>> I don't know how much global appeal this would hold, but I've been thinking
>> about doing a presentation on how Scala interacts with the JVM. Mostly to
>> cover common misconceptions, optimisations the JVM can make and where this
>> all fits into the memory model.
>> It might be a *little* more advanced that you planned though, as some of the
>> topics I had in mind were
>> - escape analysis
>> - scalar replacement
>> - TLAS vs stack allocation
>> - mutable vs mutable objects and optimisations the JVM can perform
>> - why object creation is cheap and the eden space
>> - garbage collection strategies
>> It might be worth trying to tie some of this stuff into the upcoming event
>> (June) that the LJC have planned to cover the Java memory model, assuming I
>> can somehow make enough time to arrange enough material!
Much of it would be about JVM in general, but with a particular emphasis on
those aspects that would be of greater interest to scala developers.
For example, many idioms call for short-lived immutable objects whereas the
java equivalent would typically use just the one longer lived mutable
objects. Recursive functions and how they behave are also of special
interest.
On 29 April 2010 10:24, Andy Hicks <andy1...@gmail.com> wrote:
> Q: how much of this would be scala, this sound more like a Part2 to
> the LJC JMM talk and would be equally true for java?
> On 29 April 2010 10:16, Rustem.Suniev <vigo...@gmail.com> wrote:
> > +1
> > Sounds great!
> > Is the interaction being changed in 2.8 in comparison to 2.7?
> > Thanks
> > Rustem
> > On Apr 29, 10:07 am, Kevin Wright <kev.lee.wri...@googlemail.com>
> > wrote:
> >> I don't know how much global appeal this would hold, but I've been
> thinking
> >> about doing a presentation on how Scala interacts with the JVM. Mostly
> to
> >> cover common misconceptions, optimisations the JVM can make and where
> this
> >> all fits into the memory model.
> >> It might be a *little* more advanced that you planned though, as some of
> the
> >> topics I had in mind were
> >> - escape analysis
> >> - scalar replacement
> >> - TLAS vs stack allocation
> >> - mutable vs mutable objects and optimisations the JVM can perform
> >> - why object creation is cheap and the eden space
> >> - garbage collection strategies
> >> It might be worth trying to tie some of this stuff into the upcoming
> event
> >> (June) that the LJC have planned to cover the Java memory model,
> assuming I
> >> can somehow make enough time to arrange enough material!
(with my organizer hat on :-) Do you think it would be better as an
LJC event (with adverts on lsug), then it could be joined up with the
JMM talk more easily (and we can plug the LSUG to new people, at the
same time :-)
On 29 April 2010 10:52, Kevin Wright <kev.lee.wri...@googlemail.com> wrote:
> Much of it would be about JVM in general, but with a particular emphasis on
> those aspects that would be of greater interest to scala developers.
> For example, many idioms call for short-lived immutable objects whereas the
> java equivalent would typically use just the one longer lived mutable
> objects. Recursive functions and how they behave are also of special
> interest.
> On 29 April 2010 10:24, Andy Hicks <andy1...@gmail.com> wrote:
>> +1 from me
>> Q: how much of this would be scala, this sound more like a Part2 to
>> the LJC JMM talk and would be equally true for java?
>> On 29 April 2010 10:16, Rustem.Suniev <vigo...@gmail.com> wrote:
>> > +1
>> > Sounds great!
>> > Is the interaction being changed in 2.8 in comparison to 2.7?
>> > Thanks
>> > Rustem
>> > On Apr 29, 10:07 am, Kevin Wright <kev.lee.wri...@googlemail.com>
>> > wrote:
>> >> I don't know how much global appeal this would hold, but I've been
>> >> thinking
>> >> about doing a presentation on how Scala interacts with the JVM. Mostly
>> >> to
>> >> cover common misconceptions, optimisations the JVM can make and where
>> >> this
>> >> all fits into the memory model.
>> >> It might be a *little* more advanced that you planned though, as some
>> >> of the
>> >> topics I had in mind were
>> >> - escape analysis
>> >> - scalar replacement
>> >> - TLAS vs stack allocation
>> >> - mutable vs mutable objects and optimisations the JVM can perform
>> >> - why object creation is cheap and the eden space
>> >> - garbage collection strategies
>> >> It might be worth trying to tie some of this stuff into the upcoming
>> >> event
>> >> (June) that the LJC have planned to cover the Java memory model,
>> >> assuming I
>> >> can somehow make enough time to arrange enough material!
Kevin Wright wrote:
> I don't know how much global appeal this would hold, but I've been
> thinking about doing a presentation on how Scala interacts with the JVM.
> Mostly to cover common misconceptions, optimisations the JVM can make
> and where this all fits into the memory model.
That would be fascinating. I'd turn up for that. That's just the kind of
information that's difficult to get from the net or books. Exactly the
kind of fringe stuff and inventive angles a user group is for IMO.
