PetaBricks and Julia Kathleen C. Alexander Massachusetts Institute - - PowerPoint PPT Presentation

PetaBricks and Julia

Kathleen C. Alexander Massachusetts Institute of Technology December 11th, 2013

Motivation

Motivation Background Approach Results Recommendations Index

The Programmer’s Dilemma

a personal example— energy landscapes

K.C. Alexander (MIT) PetaBricks and Julia 1 / 15

Motivation Background Approach Results Recommendations Index

The Programmer’s Dilemma

which algorithm is best?

K.C. Alexander (MIT) PetaBricks and Julia 1 / 15

Motivation Background Approach Results Recommendations Index

The Programmer’s Dilemma

which algorithm is best? Goal: determine the best algorithm for the application– which may be machine dependent

K.C. Alexander (MIT) PetaBricks and Julia 1 / 15

Motivation Background Approach Results Recommendations Index

Parallel Programming

• many parts of these al-

gorithms can be written in parallel

• often they can be paral-

lelized in many different ways

• optimizing these options

is a challenge

Determine the best way to parallelize the program– which will be machine dependent

K.C. Alexander (MIT) PetaBricks and Julia 2 / 15

Motivation Background Approach Results Recommendations Index

Parallel Programming

• many parts of these al-

gorithms can be written in parallel

• often they can be paral-

lelized in many different ways

• optimizing these options

is a challenge

Determine the best way to parallelize the program– which will be machine dependent

K.C. Alexander (MIT) PetaBricks and Julia 2 / 15

Background

Motivation Background Approach Results Recommendations Index

Petabricks – Algorithmic Choice

PetaBricks was developed to alleviate some of the optimiza- tion responsibility from the programmer the transform

K.C. Alexander (MIT) PetaBricks and Julia 3 / 15

Motivation Background Approach Results Recommendations Index

Petabricks – Algorithmic Choice

PetaBricks was developed to alleviate some of the optimiza- tion responsibility from the programmer the transform compiling framework

Ansel, et al. ACM SIGPLAN Conference (2009). K.C. Alexander (MIT) PetaBricks and Julia 3 / 15

Motivation Background Approach Results Recommendations Index

Petabricks – Autotuning

the autotuner determines the best configuration for the ma- chine under the tuning constraints

K.C. Alexander (MIT) PetaBricks and Julia 4 / 15

Motivation Background Approach Results Recommendations Index

Petabricks – Autotuning

Sort

Ansel, et al. ACM SIGPLAN Conference (2009). K.C. Alexander (MIT) PetaBricks and Julia 4 / 15

Motivation Background Approach Results Recommendations Index

Petabricks – Autotuning

Eigen Problem

Ansel, et al. ACM SIGPLAN Conference (2009). K.C. Alexander (MIT) PetaBricks and Julia 4 / 15

Motivation Background Approach Results Recommendations Index

Petabricks – Autotuning

Matrix Multiply

Ansel, et al. ACM SIGPLAN Conference (2009). K.C. Alexander (MIT) PetaBricks and Julia 4 / 15

Motivation Background Approach Results Recommendations Index

Julia

• Julia was developed to bridge the gap between interpreted

and compiled scientific computing

• streamlining parallelization techniques has been a priority

K.C. Alexander (MIT) PetaBricks and Julia 5 / 15

Motivation Background Approach Results Recommendations Index

Julia

http://forio.com/julia/julia K.C. Alexander (MIT) PetaBricks and Julia 5 / 15

Motivation Background Approach Results Recommendations Index

Julia

http://forio.com/julia/julia

Question: is there room for overlap between the PetaBricks and Julia approaches?

K.C. Alexander (MIT) PetaBricks and Julia 5 / 15

Approach

Motivation Background Approach Results Recommendations Index

Options for Implementation

Julia in PetaBricks

• can utilize PetaBricks autotuner and compiler
• PetaBricks compiler needs to interpret Julia

K.C. Alexander (MIT) PetaBricks and Julia 6 / 15

Motivation Background Approach Results Recommendations Index

Options for Implementation

Julia in PetaBricks

• can utilize PetaBricks autotuner and compiler
• PetaBricks compiler needs to interpret Julia

PetaBricks in Julia

• can run PetaBricks binaries inside Julia
• no PetaBricks shared object files, functions require disk i/o
• doesn’t take advantage of JuliaLang

K.C. Alexander (MIT) PetaBricks and Julia 6 / 15

Motivation Background Approach Results Recommendations Index

Options for Implementation

Julia in PetaBricks

• can utilize PetaBricks autotuner and compiler
• PetaBricks compiler needs to interpret Julia

PetaBricks in Julia

• can run PetaBricks binaries inside Julia
• no PetaBricks shared object files, functions require disk i/o
• doesn’t take advantage of JuliaLang

Julia + OpenTuner

• apply PetaBricks framework to Julia
• utilize OpenTuner to optimize Julia

K.C. Alexander (MIT) PetaBricks and Julia 6 / 15

Motivation Background Approach Results Recommendations Index

Approach Used Here

PetaBricks in Julia

• can run PetaBricks binaries inside Julia
• no PetaBricks shared object files, functions require disk i/o
• doesn’t take advantage of JuliaLang

K.C. Alexander (MIT) PetaBricks and Julia 7 / 15

Motivation Background Approach Results Recommendations Index

Approach Used Here

PetaBricks in Julia

• can run PetaBricks binaries inside Julia
• no PetaBricks shared object files, functions require disk i/o
• doesn’t take advantage of JuliaLang

⇒ most naive approach possible: → compile PetaBricks executable, exe → julia ¿ run(‘$exe $in $out‘)

K.C. Alexander (MIT) PetaBricks and Julia 7 / 15

Motivation Background Approach Results Recommendations Index

Approach Used Here

PetaBricks in Julia

• can run PetaBricks binaries inside Julia
• no PetaBricks shared object files, functions require disk i/o
• doesn’t take advantage of JuliaLang

⇒ most naive approach possible: → compile PetaBricks executable, exe → julia ¿ run(‘$exe $in $out‘) ⇒ compare with PetaBricks and Julia alone → lower bound of performance improvement → is there proof of benefit?

K.C. Alexander (MIT) PetaBricks and Julia 7 / 15

Results

Motivation Background Approach Results Recommendations Index

PetaBricks- Tuning Improvements

performance improvement— tuned and untuned PetaBricks Matrix Multiply

1000 2000 3000 4000 Size 50 100 150 200 250 Wall-Clock Time [s] tuned untuned

K.C. Alexander (MIT) PetaBricks and Julia 8 / 15

Motivation Background Approach Results Recommendations Index

Comparing PetaBricks with Julia - Apples to Apples

PetaBricks → functions read in ASCII files and output same → determines parallelization during autotuning → autotuning can take days Julia → JIT for each independent execution → can addprocs(n), but may not parallelize → can be used interactively

K.C. Alexander (MIT) PetaBricks and Julia 9 / 15

Motivation Background Approach Results Recommendations Index

Comparing PetaBricks with Julia - Apples to Apples

PetaBricks → functions read in ASCII files and output same → determines parallelization during autotuning → autotuning can take days PetaBricks → JIT for each independent executable → can addprocs(n), but may not parallelize → can be used interactively → make both programs do i/o → run both programs from shell → try addprocs(n) in Julia, with no other instructions → subtract ’hello world’ start-up time from Julia wall-clock

K.C. Alexander (MIT) PetaBricks and Julia 9 / 15

Motivation Background Approach Results Recommendations Index

Comparing PetaBricks to Julia - EigenSolve

EigenSolve

200 400 600 800 1000 Size 2 4 6 8 10 Wall-Clock Time [s] Julia Julia-Scaled PetaBricks

→ Julia seems to do the best for large matrices → however, the results were not comparable → this test was not a good apples-to-apples perfor- mance test

K.C. Alexander (MIT) PetaBricks and Julia 10 / 15

Motivation Background Approach Results Recommendations Index

Comparing PetaBricks with Julia - Sort

Sort

20 40 60 80 100 Size [104 ] 2 4 6 8 Wall-Clock Time [s] Julia Julia-Scaled PetaBricks

→ Julia and PetBricks con- verge for large vectors → PetaBricks is better with shorter vectors → effect of i/o not consid- ered wrt performance

K.C. Alexander (MIT) PetaBricks and Julia 11 / 15

Motivation Background Approach Results Recommendations Index

Comparing PetaBricks with Julia Matrix Multiply

i5-3339 (4 CPU)

200 400 600 800 1000 Size 2 4 6 8 Wall-Clock Time [s] Julia Julia-Scaled PetaBricks

i7-3770 (8 CPU)

1000 2000 3000 4000 Size 20 40 60 80 Wall-Clock Time [s] Julia Julia-Scaled Julia-Scaled-8p PetaBricks

→ Julia and PetBricks converge moderate matrix sizes on fewer cores → PetaBricks is better with smaller lists and larger matrices → using addprocs(n) with no other instruction does not utilize parallel func- tionality in Julia

K.C. Alexander (MIT) PetaBricks and Julia 12 / 15

Motivation Background Approach Results Recommendations Index

Running PetaBricks from Julia

Matrix Multiply

1000 2000 3000 4000 Size 20 40 60 80 Wall-Clock Time [s] Julia-Scaled PB-Julia-Scaled PetaBricks

→ Can get PetaBricks im- provement by incorpo- rating PetaBricks exe- cutable in Julia → effect of i/o not consid- ered wrt performance

K.C. Alexander (MIT) PetaBricks and Julia 13 / 15

Recommendations

Motivation Background Approach Results Recommendations Index

Recommendations

Matrix Multiply

200 400 600 800 1000 Size 2 4 6 8 Wall-Clock Time [s] Julia Julia-Scaled PetaBricks → under many circum- stances, Julia performs as well as PetaBricks without days

• f

compilation

K.C. Alexander (MIT) PetaBricks and Julia 14 / 15

Motivation Background Approach Results Recommendations Index

Recommendations

Matrix Multiply

200 400 600 800 1000 Size 2 4 6 8 Wall-Clock Time [s] Julia Julia-Scaled PetaBricks → there is room for im- provement on the start- up time for Julia

K.C. Alexander (MIT) PetaBricks and Julia 14 / 15

Motivation Background Approach Results Recommendations Index

Recommendations

Matrix Multiply

1000 2000 3000 4000 Size 20 40 60 80 Wall-Clock Time [s] Julia-Scaled PB-Julia-Scaled PetaBricks → PetaBricks performance can be achieved by using a shell command in Julia

K.C. Alexander (MIT) PetaBricks and Julia 14 / 15

Motivation Background Approach Results Recommendations Index

Recommendations

Ansel, et. al. MIT CSAIL Technical Report MIT-CSAIL-TR-2013-026 (2013).

→ implementing Open- Tuner (when better documentation is avail- able) with Julia may be a reasonable long term goal for performance gains of this kind

K.C. Alexander (MIT) PetaBricks and Julia 14 / 15

Motivation Background Approach Results Recommendations Index

Index I

1

The Programmer’s Dilemma

a personal example— energy landscapes

2

Petabricks – Algorithmic Choice

PetaBricks was developed to alleviate some of the optimiza- tion responsibility from the programmer the transform compiling framework

3

Petabricks – Autotuning

Sort

4

Julia

Question: is there room for overlap between the PetaBricks and Julia approaches?

5

Approach Used Here

PetaBricks in Julia

• can run PetaBricks binaries inside Julia
• no PetaBricks shared object files, functions require disk i/o
• doesn’t take advantage of JuliaLang

⇒ most naive approach possible: → compile PetaBricks executable, exe → julia ¿ run(‘$exe $in $out‘) ⇒ compare with PetaBricks and Julia alone → lower bound of performance improvement → is there proof of benefit?

6

Comparing PetaBricks to Julia - EigenSolve

EigenSolve 200 400 600 800 1000 Size 2 4 6 8 10 Wall-Clock Time [s] Julia Julia-Scaled PetaBricks

→ Julia seems to do the best for large matrices → however, the results were not comparable → this test was not a good apples-to-apples perfor- mance test K.C. Alexander (MIT) PetaBricks and Julia 10 / 15

7

Comparing PetaBricks with Julia Matrix Multiply

i5-3339 (4 CPU)

i7-3770 (8 CPU)

8

Running PetaBricks from Julia

Matrix Multiply 1000 2000 3000 4000 Size 20 40 60 80 Wall-Clock Time [s] Julia-Scaled PB-Julia-Scaled PetaBricks

→ Can get PetaBricks im- provement by incorpo- rating PetaBricks exe- cutable in Julia → effect of i/o not consid- ered wrt performance K.C. Alexander (MIT) PetaBricks and Julia 13 / 15

9

Recommendations

K.C. Alexander (MIT) PetaBricks and Julia 15 / 15