[ANN] Pharo bootstrap

Le 20 janv. 2016 à 20:09, Craig Latta a écrit :

Hi all--

    Phil writes:

The point is not to clean but to start with an empty object engine and

fill it in with the minimum required core. Currently, the image is

the result of an evolutionary process and rebuilding it from zero was

not possible.

    Another approach is to modify the virtual machine so that it marks
methods as they are run, and modify the garbage collector to reclaim
methods that haven't been run. Then you can create systems that consist
of only what is necessary to run unit tests, effectively imprinting the
unit tests. You can interact with the target system from completely
independent one over a remote messaging network connection, so your unit
tests need not include graphics support, etc.[1] This seems much simpler
to me than making a virtual machine that can run multiple object
memories, and distributed object memories have several other important
uses too.

Guillermo also implemented this kind of approach with Tornado. 

"Tornado, a technique to dynamically tailor applications to only embed code (classes and methods) they use. Tornado uses a run-fail-grow approach to prepare an application for deployment. It launches minimal version of an application and installs a minimal set of statements that will start the user's application. This ap- plication is run and these statements are executed. When the application fails because there are classes or methods missing, the necessary code is installed. 

Extended results of Tornado: A Run-Fail-Grow approach for Dynamic Application Tayloring (PDF Download Available). Available from: https://www.researchgate.net/publication/278629977_Extended_results_of_Tornado_A_Run-Fail-Grow_approach_for_Dynamic_Application_Tayloring [accessed Jan 21, 2016]."

But is not so easy ...

This makes it possible to audit the whole build process, create

different base image for various purposes (very small, headless with

no UI objects at all, using the object engine in a completely

different way that is used today).

That makes the image the result of a fully reproducible process, from

a file of zero bytes, to a file with a working image.

    You can do all of these things with imprinted unit tests as well.

Also, this allows to experiment with new low level mechanisms. This is

currently not possible since it makes us like performing surgery on

our own brain.

    With remote messaging, which has been around for many years, you
can perform this surgery from afar.

Also, the Oz-inspired system will help in exploring other image

memories (small or large) and manipulate them as entities. At this

point, this is also not common pratice.

    This has been common practice since the introduction of the virtual
machine simulator in 1996. The remote facilities I mentioned all work
under simulation, too.

    There's a lot of prior art here.

-C

[1] http://netjam.org/context

--
Craig Latta
netjam.org
+31   6 2757 7177 (SMS ok)
<a href="tel:%2B%201%20415%20%C2%A0287%203547" value="+14152873547" target="_blank">+ 1 415  287 3547 (no SMS)

On 20 ene 2016, at 9:08 a.m., Christophe Demarey <[hidden email]> wrote:

Hi Pavel,

Le 19 janv. 2016 à 19:53, Pavel Krivanek a écrit :

Hi,

amazing! Do you have any idea how to speed up it? The bootstrap process is now running on my machine about one and half hour and it is still far from finish. 

The process is indeed far too long (takes around 10 hours) on the CI server.

Reasons are:

- we use a stack interpreter VM

- we modified it and is a bit slower than a classical stack VM

Not quite so. My measurements did not show that… But I can understand the feeling as the StackVM is in general half as fast as the CogVM, which is the VM we are used to use since years.

- we use AST interpretation to build the bootstrap and it is very slow. Guille already did some speed improvements by avoiding to interpret loops by example.

I did not yet spend time on this point because I first wanted to have something working.

Definitely, this problem will be tackled when we will put in production the bootstrap.

We should really measure because I see three main aspects:

1) class creation (executes the class builder using AST interpretation)

2) method compilation

3) method installation  (using AST interpretation to respect trait copying)

My feeling (from what I observe only) is that most of the time is consumed in 2) and 3), and that specially it is in 3).

And I particularly want to see how the growing of collections perform.

Cheers,

Christophe

Cheers,

-- Pavel

2016-01-19 15:29 GMT+01:00 Christophe Demarey <[hidden email]>:

Hi all,

In case you do not know, we work on bootstrapping Pharo, i.e. create a Pharo image from sources, not based on a previous image (well, we use a pharo image to produce it but no code / state from it).

This process will allow to define a minimal Pharo kernel (currently 52 packages but we could have it far smaller) and to modularize the whole image (currently packages have too much dependencies on packages already loaded into the image).

The bootstrap process also allows to write down the recipe to initialize a new image from scratch (some code is missing in the image or is wrong). In addition, I think we will clean a lot of historical objects that are not used anymore.

With the amazing work done by Guillermo Polito during his Phd (around Espell, Oz): https://guillep.github.io/files/publications/Poli15Thesis.pdf, we succeeded to get a first prototype of a bootstraped Pharo 5 image (from 5.392).

This prototype is able to run an eval command line handler and to log output / errors. Not all classes are yet initialized and you cannot yet save / restart this image but it is a big step forward.

It is a 4 mb image (could be half the size without unicode data). You can download it at: http://chercheurs.lille.inria.fr/~demarey/pmwiki/pub/pharo-bootstrap/pharo-bootstrap.zip.

Next steps are to have a bootstrapped image fully working, then to load packages on top of it (like network, sunit) to produce a minimal image.

Then, we need to implement an Oz VM on top of Spur.

After that, we need to work on a reliable way to build the bootstrap (not too sensitive to changes in the image).

Christophe.

-------

demarey@193-51-236-143:~/dev/rmod/bootstrap/bootstrap-2016-01-19$ ../pharo bootstrap.image --no-default-preferences eval "1 + 1"

2

demarey@193-51-236-143:~/dev/rmod/bootstrap/bootstrap-2016-01-19$ ../pharo bootstrap.image --no-default-preferences eval "'a' , 'b'"

'ab'

demarey@193-51-236-143:~/dev/rmod/bootstrap/bootstrap-2016-01-19$ ../pharo bootstrap.image --no-default-preferences eval "1 / 0"

ZeroDivide

SmallInteger>>/

UndefinedObject>>DoIt

OpalCompiler>>evaluate

OpalCompiler(AbstractCompiler)>>evaluate:

SmalltalkImage>>evaluate:

EvaluateCommandLineHandler>>no (source is Undeclared) 

no source in EvaluateCommandLineHandler>>evaluate: in Block: no source

BlockClosure>>on:do:

EvaluateCommandLineHandler>>evaluate:

EvaluateCommandLineHandler>>evaluateArguments

EvaluateCommandLineHandler>>activate

EvaluateCommandLineHandler class(CommandLineHandler class)>>activateWith:

BasicCommandLineHandler>>no (source is Undeclared) 

no source in PharoCommandLineHandler(BasicCommandLineHandler)>>activateSubCommand: in Block: no source

BlockClosure>>on:do:

PharoCommandLineHandler(BasicCommandLineHandler)>>activateSubCommand:

PharoCommandLineHandler(BasicCommandLineHandler)>>handleSubcommand

PharoCommandLineHandler(BasicCommandLineHandler)>>handleArgument:

BasicCommandLineHandler>>no (source is Undeclared) 

no source in PharoCommandLineHandler(BasicCommandLineHandler)>>activate in Block: no source

BlockClosure>>on:do:

PharoCommandLineHandler(BasicCommandLineHandler)>>activate

PharoCommandLineHandler>>activate

PharoCommandLineHandler class(CommandLineHandler class)>>activateWith:

PharoCommandLineHandler class>>no (source is Undeclared) 

no source in PharoCommandLineHandler class>>activateWith: in Block: no source

NonInteractiveUIManager(UIManager)>>defer:

PharoCommandLineHandler class>>activateWith:

no source in BasicCommandLineHandler>>activateSubCommand: in Block: no source

BlockClosure>>on:do:

BasicCommandLineHandler>>activateSubCommand:

BasicCommandLineHandler>>handleSubcommand

BasicCommandLineHandler>>handleArgument:

no source in BasicCommandLineHandler>>activate in Block: no source

SmallInteger>>no (source is Undeclared) 

UndefinedObject>>no (source is Undeclared) 

AbstractCompiler>>no (source is Undeclared) 

SmalltalkImage>>no (source is Undeclared) 

BlockClosure>>no (source is Undeclared) 

EvaluateCommandLineHandler>>no (source is Undeclared) 

EvaluateCommandLineHandler>>no (source is Undeclared) 

CommandLineHandler class>>no (source is Undeclared) 

BasicCommandLineHandler>>no (source is Undeclared) 

BasicCommandLineHandler>>no (source is Undeclared) 

PharoCommandLineHandler>>no (source is Undeclared) 

UIManager>>no (source is Undeclared) 

UndefinedObject>>no (source is Undeclared) 

CommandLineUIManager>>no (source is Undeclared) 

SmalltalkImage>>no (source is Undeclared) 

DelayMicrosecondScheduler>>no (source is Undeclared) 

BlockClosure>>no (source is Undeclared) 

SmalltalkImage>>no (source is Undeclared) 

WeakArray class>>no (source is Undeclared) 

ps: source cannot be displayed because there is no formatter available in the bootstrap

 

Well, there is no formal specification… But I could summarize the features as follows (and as Christophe points mostly out)

OzVM supports having two images in logically separated spaces. This separation is achieved by using the free bit in the object header. Thus we support only two spaces by now.

- The main addition is actually a new primitive: #resumeFromSpecialObjectsArray that receives an special objects array as argument and:

    - resumes the execution from the active process of such array

    - on context switch, if the VM is running from a second special objects array, it will return to the first one instead.

- fixes to primitives/bytecodes that assume a single special objects array. e.g.,

   - #someObject and #nextObject should only iterate objects from the correct “space”

   - #isNil comparisons should take the correct nil

   - the GC should set the correct nil object on Weak containers

The extensions I made are about 20 overrides to StackPrimitive and NewObjectMemory methods.

The rest is implemented completely on image side. We use mirror primitives to manipulate objects from other space (as each image/space has it’s own selector table).

I have in my long TODO list see if the "clever hack” I did can be supported in a nice manner on top of Spur’s segmented memory. However, I should stop commencing new projects and start finishing them :P.

On 20 ene 2016, at 9:03 a.m., Christophe Demarey <[hidden email]> wrote:

Hi Eliot,

Le 19 janv. 2016 à 19:29, Eliot Miranda a écrit :

Hi All,

    great news!  Where can I read a specification of the Oz VM facilities?

I do not know all the details of the Oz VM but basically, it is a standard stack interpreter vm specialized to be able to run 2 images at the same time on top of it. 

Guillermo made it possible by using one available bit in the object header of objects to mark the ownership of an object (e.g. is my object from image1 or image2?). Then, he mainly had to specialize VM primitives dealing with objects retrieval from memory. By example, he had to modify the garbage collector to set the right nil instance (yes, we have 2 images so 2 nil instances. we need to take the one of the active image) when an object is garbage collected; the primitive someObject has to return an object from the active image, etc.

There is also a way to switch execution from an image to the other just by switching the special objects array reference. 

You can find some information in: https://guillep.github.io/files/publications/Poli15Thesis.pdf.

The Oz code is in http://smalltalkhub.com/#!/~Guille/ObjectSpace and in https://github.com/guillep/OzVM

The current implementation uses Cog  6.6.1and OzVM-GuillermoPolito.22.

I do not know if Guille has a more formal specification of the Oz VM.

If you have advices on what is the best way to handle two distinct object (memory) spaces in Spur, they will be welcomed :)

Cheers,

Christophe

_,,,^..^,,,_ (phone)

On Jan 19, 2016, at 6:29 AM, Christophe Demarey <[hidden email]> wrote:

Hi all,

In case you do not know, we work on bootstrapping Pharo, i.e. create a Pharo image from sources, not based on a previous image (well, we use a pharo image to produce it but no code / state from it).

This process will allow to define a minimal Pharo kernel (currently 52 packages but we could have it far smaller) and to modularize the whole image (currently packages have too much dependencies on packages already loaded into the image).

The bootstrap process also allows to write down the recipe to initialize a new image from scratch (some code is missing in the image or is wrong). In addition, I think we will clean a lot of historical objects that are not used anymore.

With the amazing work done by Guillermo Polito during his Phd (around Espell, Oz): https://guillep.github.io/files/publications/Poli15Thesis.pdf, we succeeded to get a first prototype of a bootstraped Pharo 5 image (from 5.392).

This prototype is able to run an eval command line handler and to log output / errors. Not all classes are yet initialized and you cannot yet save / restart this image but it is a big step forward.

It is a 4 mb image (could be half the size without unicode data). You can download it at: http://chercheurs.lille.inria.fr/~demarey/pmwiki/pub/pharo-bootstrap/pharo-bootstrap.zip.

Next steps are to have a bootstrapped image fully working, then to load packages on top of it (like network, sunit) to produce a minimal image.

Then, we need to implement an Oz VM on top of Spur.

After that, we need to work on a reliable way to build the bootstrap (not too sensitive to changes in the image).

Christophe.

-------

demarey@193-51-236-143:~/dev/rmod/bootstrap/bootstrap-2016-01-19$ ../pharo bootstrap.image --no-default-preferences eval "1 + 1"

2

demarey@193-51-236-143:~/dev/rmod/bootstrap/bootstrap-2016-01-19$ ../pharo bootstrap.image --no-default-preferences eval "'a' , 'b'"

'ab'

demarey@193-51-236-143:~/dev/rmod/bootstrap/bootstrap-2016-01-19$ ../pharo bootstrap.image --no-default-preferences eval "1 / 0"

ZeroDivide

SmallInteger>>/

UndefinedObject>>DoIt

OpalCompiler>>evaluate

OpalCompiler(AbstractCompiler)>>evaluate:

SmalltalkImage>>evaluate:

EvaluateCommandLineHandler>>no (source is Undeclared) 

no source in EvaluateCommandLineHandler>>evaluate: in Block: no source

BlockClosure>>on:do:

EvaluateCommandLineHandler>>evaluate:

EvaluateCommandLineHandler>>evaluateArguments

EvaluateCommandLineHandler>>activate

EvaluateCommandLineHandler class(CommandLineHandler class)>>activateWith:

BasicCommandLineHandler>>no (source is Undeclared) 

no source in PharoCommandLineHandler(BasicCommandLineHandler)>>activateSubCommand: in Block: no source

BlockClosure>>on:do:

PharoCommandLineHandler(BasicCommandLineHandler)>>activateSubCommand:

PharoCommandLineHandler(BasicCommandLineHandler)>>handleSubcommand

PharoCommandLineHandler(BasicCommandLineHandler)>>handleArgument:

BasicCommandLineHandler>>no (source is Undeclared) 

no source in PharoCommandLineHandler(BasicCommandLineHandler)>>activate in Block: no source

BlockClosure>>on:do:

PharoCommandLineHandler(BasicCommandLineHandler)>>activate

PharoCommandLineHandler>>activate

PharoCommandLineHandler class(CommandLineHandler class)>>activateWith:

PharoCommandLineHandler class>>no (source is Undeclared) 

no source in PharoCommandLineHandler class>>activateWith: in Block: no source

NonInteractiveUIManager(UIManager)>>defer:

PharoCommandLineHandler class>>activateWith:

no source in BasicCommandLineHandler>>activateSubCommand: in Block: no source

BlockClosure>>on:do:

BasicCommandLineHandler>>activateSubCommand:

BasicCommandLineHandler>>handleSubcommand

BasicCommandLineHandler>>handleArgument:

no source in BasicCommandLineHandler>>activate in Block: no source

SmallInteger>>no (source is Undeclared) 

UndefinedObject>>no (source is Undeclared) 

AbstractCompiler>>no (source is Undeclared) 

SmalltalkImage>>no (source is Undeclared) 

BlockClosure>>no (source is Undeclared) 

EvaluateCommandLineHandler>>no (source is Undeclared) 

EvaluateCommandLineHandler>>no (source is Undeclared) 

CommandLineHandler class>>no (source is Undeclared) 

BasicCommandLineHandler>>no (source is Undeclared) 

BasicCommandLineHandler>>no (source is Undeclared) 

PharoCommandLineHandler>>no (source is Undeclared) 

UIManager>>no (source is Undeclared) 

UndefinedObject>>no (source is Undeclared) 

CommandLineUIManager>>no (source is Undeclared) 

SmalltalkImage>>no (source is Undeclared) 

DelayMicrosecondScheduler>>no (source is Undeclared) 

BlockClosure>>no (source is Undeclared) 

SmalltalkImage>>no (source is Undeclared) 

WeakArray class>>no (source is Undeclared) 

ps: source cannot be displayed because there is no formatter available in the bootstrap