git
ask
hub
Privacy
Terms
Sign in with GitHub
zihaomu/gpu_notebook
↗
Stars ·
1
Language ·
Python
Simplify and Visualize This Repo
Scan the safety of this repo
Self Host this repo
Check out more work of Developer: zihaomu
Ask anything about this repo to start.
Help me understand how to optimize matrix multiplication for GPUs. Walk me through the key concepts I need to know before attempting a CUDA implementation.
Explain how quantization works in AI models. Compare 8-bit integer and floating-point formats, and show me how compact number representation saves memory and speeds up inference.
What is FlashAttention and how does it optimize the attention mechanism? Help me implement an optimized version of attention in CUDA based on the basic algorithm.
Show me how to build a minimal AI inference engine using Python, CUDA, and C++. Explain how these three languages work together to run a model efficiently.
How can I improve a basic softmax implementation to be more memory-efficient on a GPU? Explain the optimization techniques step by step.
Full explanation on explaingit →
📎
Send
By chatting or signing in you agree to the
Terms
and chat-message logging (revocable in
History
).