Upgrade to Pro
— share decks privately, control downloads, hide ads and more …
Speaker Deck
Features
Speaker Deck
PRO
Sign in
Sign up for free
Search
Search
pypy - o interpretador mais rapido do velho oeste
Search
Andrews Medina
October 05, 2013
Technology
0
390
pypy - o interpretador mais rapido do velho oeste
Palestra sobre pypy apresentada na pythonbrasil[9].
Andrews Medina
October 05, 2013
Tweet
Share
More Decks by Andrews Medina
See All by Andrews Medina
Organizando dados juŕidicos em grafos
andrewsmedina
0
99
Clean Code - princípios e práticas para um código sustentável
andrewsmedina
0
630
Pytfalls
andrewsmedina
1
190
tsuru para quem sabe tsuru
andrewsmedina
0
77
globo.com s2 python
andrewsmedina
5
390
tsuru and docker
andrewsmedina
6
3.5k
fazendo deploys de forma simples e divertida com tsuru
andrewsmedina
3
140
let's go
andrewsmedina
2
310
TDD for Dummies
andrewsmedina
3
370
Other Decks in Technology
See All in Technology
プロダクトエンジニアリング組織への歩み、その現在地 / Our journey to becoming a product engineering organization
hiro_torii
0
130
セキュリティの民主化は何故必要なのか_AWS WAF 運用の 10 の苦悩から学ぶ
yoh
1
120
登壇ネタの見つけ方 / How to find talk topics
pinkumohikan
3
380
AIのAIによるAIのための出力評価と改善
chocoyama
2
550
Github Copilot エージェントモードで試してみた
ochtum
0
100
PHPでWebブラウザのレンダリングエンジンを実装する
dip_tech
PRO
0
200
VISITS_AIIoTビジネス共創ラボ登壇資料.pdf
iotcomjpadmin
0
160
AIの最新技術&テーマをつまんで紹介&フリートークするシリーズ #1 量子機械学習の入門
tkhresk
0
140
AIエージェント最前線! Amazon Bedrock、Amazon Q、そしてMCPを使いこなそう
minorun365
PRO
14
5k
Yamla: Rustでつくるリアルタイム性を追求した機械学習基盤 / Yamla: A Rust-Based Machine Learning Platform Pursuing Real-Time Capabilities
lycorptech_jp
PRO
2
110
Welcome to the LLM Club
koic
0
170
GeminiとNotebookLMによる金融実務の業務革新
abenben
0
230
Featured
See All Featured
Making Projects Easy
brettharned
116
6.3k
Done Done
chrislema
184
16k
Making the Leap to Tech Lead
cromwellryan
134
9.3k
Documentation Writing (for coders)
carmenintech
72
4.9k
How To Stay Up To Date on Web Technology
chriscoyier
790
250k
How to Think Like a Performance Engineer
csswizardry
24
1.7k
Improving Core Web Vitals using Speculation Rules API
sergeychernyshev
17
940
Producing Creativity
orderedlist
PRO
346
40k
Why You Should Never Use an ORM
jnunemaker
PRO
57
9.4k
"I'm Feeling Lucky" - Building Great Search Experiences for Today's Users (#IAC19)
danielanewman
229
22k
Raft: Consensus for Rubyists
vanstee
140
7k
Building Adaptive Systems
keathley
43
2.6k
Transcript
globo .com pypy o interpretador mais rápido do velho oeste
Saturday, October 5, 13
Saturday, October 5, 13
‣ djangobrasil ‣ globo.com desde 2008 ‣ splinter ‣ tsuru
@andrewsmedina Saturday, October 5, 13
‣ sabe o que é? ‣ sabe o que é
rpython? ‣ já usou pypy? ‣ já usou rpython? pypy? Saturday, October 5, 13
‣ rpython ‣ pypy ‣ mitos ‣ como contribuir cronograma
Saturday, October 5, 13
rpython Saturday, October 5, 13
‣ quem já teve vontade de escrever um compilador/ interpretador?
‣ que já implementou um? ? Saturday, October 5, 13
‣ python subset ‣ traduz rpython para c ‣ ferramenta
para escrever interpretadores rpython Saturday, October 5, 13
‣ sintaxe x implementação x backend ‣ js x jit
x c ‣ ruby x gc mark sweep x c ‣ ruby x jit x jvm rpython Saturday, October 5, 13
‣ pypy ‣ topaz ‣ lang-js ‣ lang-scheme ‣ lang-smaltalk
‣ pygirl (gameboy) ‣ lang-io ‣ hippy interpretadores Saturday, October 5, 13
pypy Saturday, October 5, 13
‣ performance ‣ eficiência ‣ compatibilidade com cPython pypy Saturday,
October 5, 13
performance Saturday, October 5, 13
‣ python 2.7.3 ‣ algumas coisas em rpython compatibilidade Saturday,
October 5, 13
‣ efeito colateral por escrever python em python interpretador level
vs app level Saturday, October 5, 13
app level Saturday, October 5, 13
interpretador level Saturday, October 5, 13
compatibilidade Saturday, October 5, 13
‣ navite code > bytecode ‣ compilação do bytecode ‣
“esquentar o interpretador” jit Saturday, October 5, 13
def f2(x, y): x = hint(x, promote=True) z = x
* 2 + 1 return z + y jit Saturday, October 5, 13
guard(x == 4) v1 = x * 2 z =
v1 + 1 v2 = z + y return(v2) jit Saturday, October 5, 13
guard(x == 4) v2 = 9 + y return(v2) jit
Saturday, October 5, 13
‣ cPython: reference counting ‣ pypy: minimark garbage collector Saturday,
October 5, 13
class B(object): def __del__(self): print "oie" garbage collector Saturday, October
5, 13
>>>> b = B() >>>> b = 1 >>>> garbage
collector - pypy Saturday, October 5, 13
>>> b = B() >>> b = 1 oie garbage
collector - cPython Saturday, October 5, 13
b = B() a = B() b.x = a a.x
= b garbage collector Saturday, October 5, 13
‣ executar c no python ‣ pip install cffi cffi
Saturday, October 5, 13
cffi Saturday, October 5, 13
‣ concorrencia ‣ inspirado pelo http://www.stackless.com/ ‣ http://greenlet.readthedocs.org/en/latest/ stackless Saturday,
October 5, 13
‣ paralelismo ‣ baseado em banco de dados stm Saturday,
October 5, 13
‣ escrito em rpython ‣ já funciona :) ‣ http://buildbot.pypy.org/numpy-status/latest.html
numpy + pypy Saturday, October 5, 13
>>>> import numpypy >>>> numpypy.array([1,2,3]) array([1, 2, 3]) >>>> numpy
+ pypy Saturday, October 5, 13
numpy + pypy ‣ 231/559 names ‣ 131/161 ndarray attributes
‣ 34/48 dtype attributes ‣ 63/134 generic attributes ‣ 28/32 flatiter attribute Saturday, October 5, 13
mitos Saturday, October 5, 13
integração com c Saturday, October 5, 13
‣ funciona com ctypes (mysqldb) ‣ cffi integração com c
Saturday, October 5, 13
‣ funciona com ctypes (mysqldb) ‣ cffi integração com c
Saturday, October 5, 13
python3 Saturday, October 5, 13
python3 Saturday, October 5, 13
python3 Saturday, October 5, 13
6x mais rápido que cPython Saturday, October 5, 13
6x mais rápido que cPython Saturday, October 5, 13
6x mais rápido que cPython Saturday, October 5, 13
6x mais rápido que cPython Saturday, October 5, 13
compilar pypy demora Saturday, October 5, 13
compilar pypy demora Saturday, October 5, 13
compilar pypy demora Saturday, October 5, 13
compilar pypy demora Saturday, October 5, 13
compilar pypy demora Saturday, October 5, 13
‣ pypy.org ‣ virtualenv ‣ testes ‣ benchmarks como começar
Saturday, October 5, 13
‣ bugs.pypy.org ‣ https://bitbucket.org/pypy/pypy ‣ http://doc.pypy.org ‣ doações (stm, numpy,
python3) como contribuir Saturday, October 5, 13
obrigado! @andrewsmedina Saturday, October 5, 13