Pharo VM performance Clement Bera Myself Clment Bra 2011-2013: - - PowerPoint PPT Presentation

▶

Jul 07, 2023 41 likes •294 views

Pharo VM performance Clement Bera Myself Clment Bra 2011-2013: Engineer on the Pharo VM 2013-2017: PhD student Optimisations of the Pharo VM JIT compiler Binary tree benchmark 16 14 12 10 8 6 4 2 0 Interpreter

SLIDE 1

Pharo VM performance

Clement Bera

SLIDE 2

Myself

Clément Béra
2011-2013: Engineer on the Pharo VM
2013-2017: PhD student
Optimisations of the Pharo VM JIT compiler

SLIDE 3

2 4 6 8 10 12 14 16 Interpreter 2005 Stack 2009 Cog V1 2010 Cog V2 2011 Spur 2014 Sista future

Binary tree benchmark

SLIDE 4

2 4 6 8 10 12 14 16 Interpreter 2005 Stack 2009 Cog V1 2010 Cog V2 2011 Spur 2014 Sista future

Binary tree benchmark

Pharo 5 2016

SLIDE 5

Plan

Pharo 5 (stable)
First time we out benched most competitors
Pharo 6 (released next week ???)
Pharo 7

SLIDE 6

Code execution GC

SLIDE 7

GC

Pharo 5
New memory manager Spur
Pharo 6
New compactor
Pharo 7
Incremental GC ???

SLIDE 8

Pharo 5: Spur

Efficient scavenges
In most applications, most GC time is now in

scavenges Code execution GC

SLIDE 9

Pharo 6: New compactor

Loading a 200 Mb Moose Model in 250 Mb image

February April Total time 2 min 1 min 2 sec Time in Full GC 1 min 2 sec Full GC avg pause 15 sec 0.5 sec Time in scavenge 15 sec 15 sec

SLIDE 10

Pharo 6: New compactor

Loading a 200 Mb Moose Model in 250 Mb image

February April Total time 2 min 1 min 2 sec Time in Full GC 1 min 2 sec Full GC avg pause 15 sec 0.5 sec Time in scavenge 15 sec 15 sec

<- GC tuning gets it down to 5 sec

SLIDE 11

Pharo 7: Incremental GC ??

Full GC pauses: ~500 ms at ~500Mb
Java default GC at 200ms soft real time
Solution
Incremental marking
Incremental compaction

SLIDE 12

Code execution

Pharo 5:
Spur got 1.8x
Pharo 6:
Polishing and micro-optimisations
Pharo 7:
Sista gets 1.5x-5x

SLIDE 13

Pharo 5: Spur 1.8x

Class table speeds-up look-up caches
New immediate objects
22 bits hash

SLIDE 14

Pharo 6

Register allocation improvements
Two path compilation
Frameless code for setter-like methods

SLIDE 15

Sista: Pharo 7 ?

Program introspection
Speculate on types based on previous runs
Optimize frequently used code
Deoptimize and reoptimize code incorrectly

speculated

SLIDE 16

Goals

Program readability
Performance

SLIDE 17

Program readability

1 to: array size do: [ :i | (array at: i) yourself ]. array do: [ :elem | elem yourself ]. array do: #yourself.

SLIDE 18

Program readability

1 to: array size do: [ :i | (array at: i) yourself ]. array do: [ :elem | elem yourself ]. array do: #yourself.

2 5 20 87M/ sec 28M/ sec 13M/ sec 3.7M /sec 15M/ sec 21M/ sec 10M/ sec 3.9M /sec 94M/ sec 40M/ sec 22M/ sec 6.5M /sec

SLIDE 19

Performance

1 2 3 4 5 6 A* ThreadRing SpectralNorm JSJSON BinaryTree DeltaBlue Richards TCAP Kmeans Sista Pharo

SLIDE 20

Getting stable

Support most development workflow
Support image recompilation
Integration has started

SLIDE 21

In-image design

Smalltalk image Virtual machine Cogit CompiledCode to native code Machine-specific optimisations Scorch CompiledCode to CompiledCode Smalltalk-specific optimisations CompiledCode (persisted across start-ups) native functions (discarded on shut-down) Baseline JIT Optimising JIT

SLIDE 22

Missing

IDE support
Debugger
Methods to show
Stability, testing

SLIDE 23

Are you interested ?

Incremental GC ?
VM performance ?
VM features ?
Come and talk to us !

SLIDE 24

We are looking for…

Use-cases showing what to improve
Large real-world benchmarks
Contributors
Investment

SLIDE 25

Conclusion

Pharo 5: Fastest VM
Pharo 6: Polishing
Pharo 7: Going further