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
Go言語でMac GPUプログラミング
Search
monochromegane
December 18, 2023
Programming
800
2
Share
Go言語でMac GPUプログラミング
2023.12.18 Fukuoka.go#19 Reboot
https://fukuokago.connpass.com/event/302717/
monochromegane
December 18, 2023
More Decks by monochromegane
See All by monochromegane
ベクトル検索のフィルタを用いた機械学習モデルとの統合 / python-meetup-fukuoka-06-vector-attr
monochromegane
2
710
Claude Codeセッション現状確認 2026福岡 / fukuoka-aicoding-00-beacon
monochromegane
4
520
Go言語での実装を通して学ぶLLMファインチューニングの仕組み / fukuokago22-llm-peft
monochromegane
0
230
不確実性下における目的と手段の統合的探索に向けた連続腕バンディットの応用 / iot70_gp_rff_mab
monochromegane
2
340
なめらかなシステムと運用維持の終わらぬ未来 / dicomo2025_coherently_fittable_system
monochromegane
0
10k
ベクトル検索システムの気持ち
monochromegane
38
13k
Go言語での実装を通して学ぶ、高速なベクトル検索を支えるクラスタリング技術/fukuokago-kmeans
monochromegane
1
290
Go言語でターミナルフレンドリーなAIコマンド、afaを作った/fukuokago20_afa
monochromegane
2
360
多様かつ継続的に変化する環境に適応する情報システム/thesis-defense-presentation
monochromegane
1
1.2k
Other Decks in Programming
See All in Programming
ついに来た!本格的なマルチクラウド時代の Google Cloud
maroon1st
0
340
Claude Codeをカスタムして自分だけのClaude Codeを作ろう
terisuke
0
160
ふりがな Deep Dive try! Swift Tokyo 2026
watura
0
260
GoogleCloudとterraform完全に理解した
terisuke
1
180
CDK Deployのための ”反響定位”
watany
5
920
GNU Makeの使い方 / How to use GNU Make
kaityo256
PRO
16
5.6k
GitHubCopilotCLIをはじめよう.pdf
htkym
0
310
検索設計から 推論設計への重心移動と Recall-First Retrieval
po3rin
5
1.4k
(Re)make Regexp in Ruby: Democratizing internals for the JIT
makenowjust
3
950
mruby on C#: From VM Implementation to Game Scripting (RubyKaigi 2026)
hadashia
2
1.4k
AI時代のエンジニアリングの原則 / Engineering Principles in the AI Era
haru860
0
960
決定論 vs 確率論:Gemini 3 FlashとTF-IDFを組み合わせた「法規判定エンジン」の構築
shukob
0
140
Featured
See All Featured
Building a Modern Day E-commerce SEO Strategy
aleyda
45
9k
Exploring the relationship between traditional SERPs and Gen AI search
raygrieselhuber
PRO
2
3.9k
30 Presentation Tips
portentint
PRO
1
280
Building Adaptive Systems
keathley
44
3k
GraphQLの誤解/rethinking-graphql
sonatard
75
12k
Exploring anti-patterns in Rails
aemeredith
3
340
Accessibility Awareness
sabderemane
1
110
What Being in a Rock Band Can Teach Us About Real World SEO
427marketing
0
220
Evolving SEO for Evolving Search Engines
ryanjones
0
180
Noah Learner - AI + Me: how we built a GSC Bulk Export data pipeline
techseoconnect
PRO
0
170
Data-driven link building: lessons from a $708K investment (BrightonSEO talk)
szymonslowik
1
1k
Google's AI Overviews - The New Search
badams
0
990
Transcript
Lightning Talks ࡾ༔հ / Pepabo R&D Institute, GMO Pepabo, Inc.
2023.12.18 Fukuoka.go#19 Reboot GoݴޠͰMac GPUϓϩάϥϛϯά
ϓϦϯγύϧΤϯδχΞ ࡾ ༔հ / @monochromegane 2 https://blog.monochromegane.com Yusuke Miyake ϖύϘݚڀॴ
ݚڀһ
• ଟมྔਖ਼ن ʹै͏ཚੜʹ͕͔͔࣌ؒΔ • ͜ͷཚੜͷखॱʢͷҰͭʣҎԼͷ௨Γ 1. ֤ཁૉ͕ඪ४ਖ਼نʹै͏ཚ ΛಘΔ 2. ڞࢄߦྻ
ΛίϨεΩʔղʢ ʣͯ͠ࡾ֯ߦྻ ΛಘΔ 3. ΛٻΊΔ • ಛʹɺ֬ͷύϥϝʔλʢ ͱ ʣ͕ҟͳͬͨΓɺ࣍ݩ ͕େ͖͍ ߹ʹɺཚੜʹ͕͔͔࣌ؒͬͯ͠·͏ y ∼ 𝒩 (μ, Σ), μ ∈ ℝD, Σ ∈ ℝD×D z = {zi }1≤i≤D , zi ∼ 𝒩 (0,1) Σ Σ = LL⊤ L y = μ + Lz μ Σ D 3 ͡Ίʹ
• ߦྻܭࢉಠཱ͔ͭฒߦͨ͠λεΫΛଟؚ͘ΉͨΊɺߴԽʹฒྻԽ͕༗ޮ • ͢ͳΘͪɺSIMDɺCPUͷϚϧνίΞɺGPUͳͲʹΑΔฒྻԽ • CPUόϯυͰλεΫཻখ͍͞ͷͰgoroutine͔ͳ͍ʢͱࢥ͏ʣ • GoݴޠͰͷߦྻܭࢉϥΠϒϥϦGonumCPUͷϚϧνίΞΛαϙʔτ͢Δ BLASͷόΠϯσΟϯάΛఏڙ͍ͯ͠Δ •
Apple silicon (M1) ʹGPU͕ࡌ͞Ε͍ͯΔͷͰɺͦͪΒ׆༻͍ͨ͠ 4 ͡Ίʹ
• GPUͷΞΫηεΛఏڙ͢ΔOSඪ४ࡌͷϑϨʔϜϫʔΫ • άϥϑΟοΫεॲཧҎ֎ʹɺGPU্Ͱͷฒྻܭࢉॲཧѻ͑Δ • Objective-C·ͨSwift͔ΒɺGPU্ͷॲཧΛهड़ͨ͠γΣʔμʔؔΛݺͿ • γΣʔμʔؔC++ϕʔεͷMetal Shader Language
(MSL) Ͱهड़ • Metal Performance Shaders (MPS) ͱ͍͏γΣʔμʔؔ܈ఏڙ͞ΕΔ 5 Metal: MacͰGPUϓϩάϥϛϯά
6 Metal: MacͰGPUϓϩάϥϛϯά • جຊతͳྲྀΕɺσόΠεʢGPUʣͷίϚϯυΩϡʔʹର͠ɺίϚϯυόο ϑΝͱ͍͏୯ҐͰγΣʔμʔؔΛొ͠ɺ݁ՌΛड͚औΔͱ͍͏ͷ • ͳ͓ɺCPUͱGPUͷͷΓऔΓʹઐ༻ͷόοϑΝ͕༻ҙ͞Ε͍ͯΔ ίϚϯυΩϡʔ ͷ४උ
ΓऔΓ༻ͷ όοϑΝͷ४උ όοϑΝͷσʔλ͔Β ߦྻΠϯελϯεੜ .14ͷγΣʔμʔؔ ΛॳظԽɺίϚϯυ όοϑΝͱͯ͠Τϯ ίʔυɺΩϡʔʹొ όοϑΝ͔Β݁Ռͷड ͚औΓ 0CKFDUJW$Ͱͷ ࣮ྫ
• Goݴޠ͔ΒcgoΛ͑͜ΕΒͷObjective-CͷίʔυΛݺΔ༷ࢠ • https://github.com/a-h/gpu ϥΠϒϥϦͱͯ͠ར༻Ͱ͖Δ͕MPSʹରԠ͍ͯ͠ͳ͍ • https://github.com/mikecvet/go-mm MPSͷݺͼग़͠Λ࣮͍ͯ͠Δ͕ϕϯνϚʔΫͷίʔυͷΈ • ্هΛࢀߟʹͭͭ͠ɺGoݴޠ্ͰͷGPUΛ༻͍ͨଟมྔਖ਼نʹै͏ཚ
ੜ͕Ͱ͖ͦ͏ 7 Cgo: GoݴޠͰMac GPUϓϩάϥϛϯά
1. Objective-CͷϔομϑΝΠϧΛinclude͠ɺLDFLAGSʹMetalϑϨʔϜϫʔΫΛࢦఆ͢Δ 2. ʢඞཁʹԠͯ͡ʣࣗલͷγΣʔμʔؔΛgo:embedͰΈࠐΜͰ͓͘ 3. C.xxͱͯ͠Objective-CͰهड़ͨ͠ॳظԽγΣʔμʔؔΛ࣮ߦ͢ΔؔΛݺͿɻ࣮ߦ࣌ ͷύϥϝʔλ݁ՌunsafeύοέʔδΛͬͯΞΫηεɻ 8 Cgo: GoݴޠͰMac
GPUϓϩάϥϛϯά (PͰͷ࣮ྫ
• Goݴޠ্ͰͷGPUΛ༻͍ͨଟมྔਖ਼نʹै͏ཚੜ • Goͷίʔυ͔Β ΛcgoΛܦ༝ͯ͠Objective-Cͷؔʹ͢ • MPSͷMPSMatrixDecompositionCholeskyΛ༻͍ͯίϨεΩʔղ • ࣗલγΣʔμʔؔΛ༻͍ͯԼࡾ֯ߦྻҎ֎Λ0ʹຒΊΔ •
MPSͷMPSMatrixVectorMultiplicationΛ༻͍ͯ Λܭࢉ • MPSͷMPSMatrixSumΛ༻͍ͯ Λܭࢉ • GoͷίʔυͰ݁ՌΛड͚औΔ z, μ, Σ Lz μ + Lz 9 Cgo: GoݴޠͰMac GPUϓϩάϥϛϯά
• GonumͱMetal࣮ͷ࣮ߦΛൺֱʢ1000࣍ݩʣ 10 ඪ४ਖ਼نཚͷมͷൺֱ BenchmarkTransformNormMetal-8 9 117668310 ns/op BenchmarkTransformNormGonumBLAS-8 55
21494668 ns/op BenchmarkTransformNormGonum-8 21 54288034 ns/op BenchmarkTransformNormCholMetal-8 1140 1070094 ns/op BenchmarkTransformNormCholGonumBLAS-8 15124 78960 ns/op BenchmarkTransformNormCholGonum-8 6712 177368 ns/op • ίϨεΩʔղͷ݁ՌΛผ్͢Α͏ʹͨ͠߹ͷൺֱ • MPSͷίϨεΩʔղগ͍͔͠͠Εͳ͍͕ɺͦͷଞͷࠩԿ͔
• ߦྻʢ1000x1000ʣͱߦྻʢ1000x1000ʣͷࢉΛൺֱ 11 ߦྻࢉͷൺֱ BenchmarkMatrixMultipicationMetal-8 494 2222134 ns/op BenchmarkMatrixMultipicationGonumBLAS-8 60
22063894 ns/op BenchmarkMatrixMultipicationGonum-8 19 59507228 ns/op BenchmarkMatrixVectorMultipicationMetal-8 1497 792244 ns/op BenchmarkMatrixVectorMultipicationGonumBLAS-8 10000 114843 ns/op BenchmarkMatrixVectorMultipicationGonum-8 972 1239177 ns/op • ߦྻʢ1000x1000ʣͱϕΫτϧʢ1000x1ʣͷࢉΛൺֱ • ߦྻಉ࢜ͷΑ͏ͳܭࢉྔͰGPUͷํ͕ߴɻ ͷΑ͏ͳߦྻͱϕΫτϧͷࢉͰ͜ ͷ࣍ݩʹ͓͍ͯGPUҠৡͷΦʔόʔϔουͷํ͕େ͖͔ͬͨͱߟ͑ΒΕΔ Lz
• GoݴޠͰMac GPUϓϩάϥϛϯά͢Δํ๏Λհͨ͠ • ؆қతͳͷൺֱධՁΛ௨ͯ͠ɺ͍ॴͷഽײΛಘΔ͜ͱ͕Ͱ͖ͨ • MPSͷϚχϡΞϧΛಡΉͱχϡʔϥϧωοτϫʔΫͷαϙʔτ͋ΓɺΞΠ σΟΞ࣍ୈͰ໘ന͍͜ͱ͕Ͱ͖ͦ͏ • MPSͷݺͼग़͠ՄೳͳϥΠϒϥϦΛ࡞ͬͯΈ͍ͨ
• Ͳ͏ϝϞϦϦʔΫͯͦ͠͏ͳͷͰͦͷลΓվળ͍ͨ͠ • ࡾGo 12 ·ͱΊ
None