Welcome to modern-gpu-programming-for-mlsys Discussions!

[tlopex]

We use Discussions as a place for readers, contributors, and maintainers to exchange ideas around the book, TIRx, and modern GPU kernel programming.

Here, you are welcome to:

• Ask questions about the book, examples, or concepts such as Blackwell GPU, Tensor Memory, TMA, GEMM, FlashAttention, and TIRx.
• Share your experience running the examples
• Suggest improvements to explanations, figures, examples, or chapter organization.
• Help other readers and contribute to a friendly, open-minded learning community.

For concrete bugs, typos, or code/documentation fixes, feel free to open an Issue or submit a Pull Request. For broader questions, ideas, and discussions, this is the right place.

We hope this space helps the community learn, build, and improve the book together.

[Liangliang-Ma]

hi. thanks for providing such amazing project!

I saw there is an Chinese version of website https://mlc.ai/modern-gpu-programming-for-mlsys/zh/index.html. But the Chinese version has fallen far behind the English version in terms of progress. Maybe you could use AI to help translate and fill in the remaining parts?

[tlopex]

@Liangliang-Ma Hi, thanks for the kind words and for following the project!

We are already working on the remaining Chinese chapters, and the update is almost complete. The process takes a little longer because we are not simply translating the English text—we are also polishing unclear descriptions in the original version and making sure the Chinese terminology and explanations are accurate.

Please stay tuned. The new content will be available very soon!

[Liangliang-Ma]

Nice, appreciate the focus on accuracy!

[tlopex]

@Liangliang-Ma The remaining Chinese chapters are now complete and available. Hope you find them helpful, and feel free to let us know if you spot anything that could be improved!

[Liangliang-Ma]

thanks! I am reading them :)