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
TT Korea OSS Program Overview
Search
Sponsored
·
Your Podcast. Everywhere. Effortlessly.
Share. Educate. Inspire. Entertain. You do you. We'll handle the rest.
→
Tenstorrent Japan
May 01, 2026
200
0
Share
Embed
Copy iframe code
Copy JS code
Copy link
Start on current slide
TT Korea OSS Program Overview
Tenstorrent Tech Talk #7, Session3
Tenstorrent Japan
May 01, 2026
More Decks by Tenstorrent Japan
See All by Tenstorrent Japan
TTデバイスで大規模計算
tenstorrent_japan
0
170
TTデバイス間通信を支えるTT-Fablic解説 Part1
tenstorrent_japan
0
250
RISC-V CPU「TT-Ascalon」
tenstorrent_japan
0
440
TT OSS Developer Program in Japan
tenstorrent_japan
0
690
TT-NN 概要 ~内部実装の概説
tenstorrent_japan
0
260
TTNNモデル実装 実践編 ~PytorchモデルのTTNN移植から最適化まで~
tenstorrent_japan
0
250
お外でBlackhole
tenstorrent_japan
1
400
Go Out with Blackhole
tenstorrent_japan
0
390
RVV1.0で遊ぼうと思ったら Blackholeを買っていた話
tenstorrent_japan
0
400
Featured
See All Featured
Imperfection Machines: The Place of Print at Facebook
scottboms
270
14k
The Cost Of JavaScript in 2023
addyosmani
55
10k
Refactoring Trust on Your Teams (GOTO; Chicago 2020)
rmw
35
3.8k
Being A Developer After 40
akosma
91
590k
Navigating the moral maze — ethical principles for Al-driven product design
skipperchong
2
500
What does AI have to do with Human Rights?
axbom
PRO
1
2.3k
The Impact of AI in SEO - AI Overviews June 2024 Edition
aleyda
6
1.2k
Odyssey Design
rkendrick25
PRO
2
770
More Than Pixels: Becoming A User Experience Designer
marktimemedia
3
500
The #1 spot is gone: here's how to win anyway
tamaranovitovic
3
1.1k
Neural Spatial Audio Processing for Sound Field Analysis and Control
skoyamalab
0
410
Highjacked: Video Game Concept Design
rkendrick25
PRO
1
440
Transcript
1 TT Korea OSS Program Overview 2026.04.20 Presentor: Hyunggi Chang
Location: TT-Japan
2 Korea OSS program Brief: • Provide p100a devices to
15 open-source developers / students • Developers work on a project 6-9 months • Developers get to keep the device after the program terminates Expected outcome: • Successful project showcase on social networks
3 Project proposals No Title 1 Implement autonomous driving models
(Sparse4D, SparseDrive) 2 Implement BitNet b1.58 inference model 3 Implement Keyword spotting model 4 Implement a full on-prem LLM Chatbot pipeline, including RAG/LLM/STT 5 RVV optimization on popular RISC-V libraries: llama.cpp, dav1d, ffmpeg, zstd 6 Enable support on Whisper + CTranslate2 7 Enable support on TorchAO 8 Implement 3D Gaussian Splatting 9 Enable IREE compiler 10 Implement Chatterbox TTS 11 Optimize DiT + LLM kernels 12 Implement OpenSora 1.2 13 Integrate TinyGrad
4 Noticeable outcomes
5 1. Improve RVV acceleration codes on open-source projects •
50x speed improvement on RISC-V algorithms on dav1d / ffmpeg
6 2. Discussing RISC-V CI integration for major open-source projects
• Working on CI runners for VideoLAN and LLVM under RISE RISC-V
7 3. Improve BH-Linux demo • Improved stability and capability
of Linux demo on Blackhole.
8 4. Many fixes • Multiple firmware bugs • Fixed
outdated dev docs • Mistakes in hardware designs
9 5. Autonomous driving models
10 6. Support Blackhole on Backend.ai
11 7. Qwen3.5-35B-a3B on p100a
12 7. Bun on RISC-V
13 7. Eggroll training on TT Evolution Strategies at the
Hyperscale
14 8. Robotics VLA models on Blackhole
15 8. Reinforcement learning library
16 Expected outcomes Expected to end in Sept 2026. •
Academic papers • Open-sourced projects • Deeper interests and expertise • Partnership with participants for further research & development