Javascript:SpiderMonkey:ProjectGenerationGarbageCollection: Difference between revisions

From MozillaWiki
Jump to navigation Jump to search
m (Undo revision 504696 by Wei Wu (talk))
No edit summary
Line 1: Line 1:
Objective: Implement Generational Garbage collection in Spider Monkey. <Some goal based off V8 performance delta % on Earley Boyer benchmark>
== Objective ==
Implement Generational Garbage collection in Spider Monkey. <Some goal based off V8 performance delta % on Earley Boyer benchmark>


Accountable: Naveed
* Accountable: Naveed
Responsible: Terrence, Jon, and Steve, evilpie Ms2ger
 
Consulted: bhackett billm
* Responsible: Terrence, Jon, and Steve, evilpie Ms2ger
Informed: Product Marketing
 
* Consulted: bhackett billm
 
* Informed: Product Marketing
 
Generational GC - [https://bugzilla.mozilla.org/show_bug.cgi?id=619558 Bug 619558]
 
== Steps ==
 
Exact Rooting done in JS Shell [https://bugzilla.mozilla.org/show_bug.cgi?id=753203 Bug 753203] - 12 weeks started


Generational GC - Bug 619558
Steps
Exact Rooting done in JS Shell [Bug 753203] - 12 weeks started
Rooting Analysis on TinderBox (sfink) - 3 weeks
Rooting Analysis on TinderBox (sfink) - 3 weeks
patch is ready for review, would fix the problem with some pushes not getting builds
patch is ready for review, would fix the problem with some pushes not getting builds
fix up other stuff broken by above change
fix up other stuff broken by above change
patch would fix last remaining problem with getting it on Try server
patch would fix last remaining problem with getting it on Try server
Unhide it for inbound - keep it green
Unhide it for inbound - keep it green
Get the root analysis build green - two weeks?
Get the root analysis build green - two weeks?
Get jit-tests green [Bug 745742] - 2 days
 
Get jit-tests green [https://bugzilla.mozilla.org/show_bug.cgi?id=745742 Bug 745742] - 2 days
 
Get js ref tests green  
Get js ref tests green  
Get jsapi-tests green [Bug 831376]
 
Fuzzer coverage of code [Bug 773746]
Get jsapi-tests green [https://bugzilla.mozilla.org/show_bug.cgi?id=831376 Bug 831376]
 
Fuzzer coverage of code [https://bugzilla.mozilla.org/show_bug.cgi?id=773746 Bug 773746]
 
rely on static analysis to make this not be a whack-a-mole game
rely on static analysis to make this not be a whack-a-mole game
Remove E4X  
Remove E4X  
Do something about JSD
Do something about JSD
Add exact roots to stack structures - 2 weeks
Add exact roots to stack structures - 2 weeks
AddRoot/RemoveRoot for Heap structures - 4 weeks
AddRoot/RemoveRoot for Heap structures - 4 weeks
Static code analysis [Bug 831409]
 
Static code analysis [https://bugzilla.mozilla.org/show_bug.cgi?id=831409 Bug 831409]
 
Fix all discovered rooting hazards (~800) (sfink,jonco) - 4 weeks
Fix all discovered rooting hazards (~800) (sfink,jonco) - 4 weeks
Optimize all discovered over-rooting (~100)
Optimize all discovered over-rooting (~100)
Automate static analysis (sfink)
Automate static analysis (sfink)
Need a server (dm-sixgill01?)
Need a server (dm-sixgill01?)
http://people.mozilla.org/~bhackett/rootingHazards.html
http://people.mozilla.org/~bhackett/rootingHazards.html
http://people.mozilla.org/~bhackett/gcFunctions.html
http://people.mozilla.org/~bhackett/gcFunctions.html


JIT Integration with post barriers - 4 weeks, parallelized
JIT Integration with post barriers - 4 weeks, parallelized
IonMonkey [Bug 831506] - 1 week (bhackett)
 
JaegerMonkey [Bug 764876] - 2 days (bhackett)
IonMonkey [https://bugzilla.mozilla.org/show_bug.cgi?id=831506 Bug 831506] - 1 week (bhackett)
Baseline JIT [Bug 831507] - <ask jandem/djvj for an estimate>
 
JaegerMonkey [https://bugzilla.mozilla.org/show_bug.cgi?id=764876 Bug 764876] - 2 days (bhackett)
 
Baseline JIT [https://bugzilla.mozilla.org/show_bug.cgi?id=831507 Bug 831507] - <ask jandem/djvj for an estimate>


Generation Garbage Collection in the Shell
Generation Garbage Collection in the Shell
Implement prototype algorithm (terrence) - 1 week
Implement prototype algorithm (terrence) - 1 week
This is a potential milestone: we would need to build it and test it on TBPL similar to how |r| works now -- |GGCJS|.
This is a potential milestone: we would need to build it and test it on TBPL similar to how |r| works now -- |GGCJS|.


Exactly Root the Browser [Bug 831379] - 10 weeks
Exactly Root the Browser [https://bugzilla.mozilla.org/show_bug.cgi?id=831379 Bug 831379] - 10 weeks
 
This is a potential milestone: we would turn on exact rooting for release FF at this point. We would not get a performance boost from this (necessarily), but it would lock in our work to this point.
This is a potential milestone: we would turn on exact rooting for release FF at this point. We would not get a performance boost from this (necessarily), but it would lock in our work to this point.


Performance Tuning - 4 weeks (This should probably start asap)
Performance Tuning - 4 weeks (This should probably start asap)
Test against V8 Earley-Boyer benchmark.
Test against V8 Earley-Boyer benchmark.
Also v8 deltablue, raytrace
Also v8 deltablue, raytrace
Implement a Nursery [Bug 706885] - 1 week
 
Implement a Nursery [https://bugzilla.mozilla.org/show_bug.cgi?id=706885 Bug 706885] - 1 week
 
We may need to Implement Bill’s Pools/Zones idea to get the nursery to the requisite perf - 6 weeks
We may need to Implement Bill’s Pools/Zones idea to get the nursery to the requisite perf - 6 weeks
Refactor code to avoid rooting on hot paths and keep rooter overhead acceptable - ??? weeks (start early) (bhackett)
Refactor code to avoid rooting on hot paths and keep rooter overhead acceptable - ??? weeks (start early) (bhackett)


Re-implement HashTable rekeying
Re-implement HashTable rekeying
We undid this code because it was a perf regression and was still a bit buggy. (terrence) - 3 weeks
We undid this code because it was a perf regression and was still a bit buggy. (terrence) - 3 weeks


Make the Post Barrier Verifier Green in the browser [Bug 764882]
Make the Post Barrier Verifier Green in the browser [https://bugzilla.mozilla.org/show_bug.cgi?id=764882 Bug 764882]
 
Investigate how long it will take to do generational barriers - 1 week
Investigate how long it will take to do generational barriers - 1 week


We may need to rewrite the maps in xpconnect and the browser in terms of HashTable: this could be a bunch of work.  
We may need to rewrite the maps in xpconnect and the browser in terms of HashTable: this could be a bunch of work.  
Make JS_IsAboutToBeFinalized indirect [Bug 765432] - 10 weeks
Make JS_IsAboutToBeFinalized indirect [https://bugzilla.mozilla.org/show_bug.cgi?id=765432 Bug 765432] - 10 weeks


OTHER
== Other ==


Should we establish a new  benchmark specifically for GGC. (sfink votes yes)
Should we establish a new  benchmark specifically for GGC. (sfink votes yes)
What would it measure? there are multiple goals, e.g. throughput/MMU/pause time
What would it measure? there are multiple goals, e.g. throughput/MMU/pause time
Should we make it a public benchmark?
Should we make it a public benchmark?
What workloads should we consider?
What workloads should we consider?
compartmental GC very important to us, not necessarily applicable to other implementations
compartmental GC very important to us, not necessarily applicable to other implementations
allocation rate (broken down into live vs garbage), steady behavior vs swapping between allocation + computation modes, etc.
allocation rate (broken down into live vs garbage), steady behavior vs swapping between allocation + computation modes, etc.
How can we keep our advantage on Splay benchmark
How can we keep our advantage on Splay benchmark
possible to do with TI --- look at types of objects promoted from nursery to major heap, eventually start allocating them directly in the major heap
possible to do with TI --- look at types of objects promoted from nursery to major heap, eventually start allocating them directly in the major heap
Compacting GC - Revisit in Febuary/March - 1 Month
Compacting GC - Revisit in Febuary/March - 1 Month
Delayed until exact rooting is fully done. We can add this in if we have time in the schedule that we cannot parallelize.
Delayed until exact rooting is fully done. We can add this in if we have time in the schedule that we cannot parallelize.


Risks
== Risks ==
Team is responsible for many top crashers
 
External rooting API has not been designed  
* Team is responsible for many top crashers
GGC algorithm has not been decided on - it may not be faster
 
JSD1 exact rooting is lurking
* External rooting API has not been designed  
 
* GGC algorithm has not been decided on - it may not be faster
 
* JSD1 exact rooting is lurking

Revision as of 01:57, 24 January 2013

Objective

Implement Generational Garbage collection in Spider Monkey. <Some goal based off V8 performance delta % on Earley Boyer benchmark>

  • Accountable: Naveed
  • Responsible: Terrence, Jon, and Steve, evilpie Ms2ger
  • Consulted: bhackett billm
  • Informed: Product Marketing

Generational GC - Bug 619558

Steps

Exact Rooting done in JS Shell Bug 753203 - 12 weeks started

Rooting Analysis on TinderBox (sfink) - 3 weeks

patch is ready for review, would fix the problem with some pushes not getting builds

fix up other stuff broken by above change

patch would fix last remaining problem with getting it on Try server

Unhide it for inbound - keep it green

Get the root analysis build green - two weeks?

Get jit-tests green Bug 745742 - 2 days

Get js ref tests green

Get jsapi-tests green Bug 831376

Fuzzer coverage of code Bug 773746

rely on static analysis to make this not be a whack-a-mole game

Remove E4X

Do something about JSD

Add exact roots to stack structures - 2 weeks

AddRoot/RemoveRoot for Heap structures - 4 weeks

Static code analysis Bug 831409

Fix all discovered rooting hazards (~800) (sfink,jonco) - 4 weeks

Optimize all discovered over-rooting (~100)

Automate static analysis (sfink)

Need a server (dm-sixgill01?)

http://people.mozilla.org/~bhackett/rootingHazards.html

http://people.mozilla.org/~bhackett/gcFunctions.html


JIT Integration with post barriers - 4 weeks, parallelized

IonMonkey Bug 831506 - 1 week (bhackett)

JaegerMonkey Bug 764876 - 2 days (bhackett)

Baseline JIT Bug 831507 - <ask jandem/djvj for an estimate>

Generation Garbage Collection in the Shell

Implement prototype algorithm (terrence) - 1 week

This is a potential milestone: we would need to build it and test it on TBPL similar to how |r| works now -- |GGCJS|.

Exactly Root the Browser Bug 831379 - 10 weeks

This is a potential milestone: we would turn on exact rooting for release FF at this point. We would not get a performance boost from this (necessarily), but it would lock in our work to this point.

Performance Tuning - 4 weeks (This should probably start asap)

Test against V8 Earley-Boyer benchmark.

Also v8 deltablue, raytrace

Implement a Nursery Bug 706885 - 1 week

We may need to Implement Bill’s Pools/Zones idea to get the nursery to the requisite perf - 6 weeks

Refactor code to avoid rooting on hot paths and keep rooter overhead acceptable - ??? weeks (start early) (bhackett)

Re-implement HashTable rekeying

We undid this code because it was a perf regression and was still a bit buggy. (terrence) - 3 weeks

Make the Post Barrier Verifier Green in the browser Bug 764882

Investigate how long it will take to do generational barriers - 1 week

We may need to rewrite the maps in xpconnect and the browser in terms of HashTable: this could be a bunch of work. Make JS_IsAboutToBeFinalized indirect Bug 765432 - 10 weeks

Other

Should we establish a new benchmark specifically for GGC. (sfink votes yes)

What would it measure? there are multiple goals, e.g. throughput/MMU/pause time

Should we make it a public benchmark?

What workloads should we consider?

compartmental GC very important to us, not necessarily applicable to other implementations allocation rate (broken down into live vs garbage), steady behavior vs swapping between allocation + computation modes, etc.

How can we keep our advantage on Splay benchmark possible to do with TI --- look at types of objects promoted from nursery to major heap, eventually start allocating them directly in the major heap

Compacting GC - Revisit in Febuary/March - 1 Month Delayed until exact rooting is fully done. We can add this in if we have time in the schedule that we cannot parallelize.

Risks

  • Team is responsible for many top crashers
  • External rooting API has not been designed
  • GGC algorithm has not been decided on - it may not be faster
  • JSD1 exact rooting is lurking