🚀 One of the proudest moments of my career happened today.
After the work of an incredible team, we officially released CUDA Python 1.0 🎉
- Faster Numba Backend
- CCCL from host callable Python
- Graphs, Green Context, IPC and more
https://t.co/Q4C60KiO4w
@Rexetdeus@anacondainc We were using the ACE Step project: https://t.co/OAlYyZrC4Y
It generated a new track each time on the fly.
Basically you would generate music while listening so there was a few minutes before the first track played.
Come join me at @anacondainc booth today for our CUDA Python 1.0 release. I’ll be hosting our Local LLM generated DJ party on a Spark DGX from 1:15-3 and presenting on the stack at 4pm!
#PyConUS
Join experts from NVIDIA and probabl and learn how to accelerate @scikit_learn on GPUs, and how OSS codes can use CUDA directly with supported APIs and languages from day one.
📅 May 5
📍 GOSIM 2026 | Station F, Paris https://t.co/FBukOkl57d
New Drop. CUDA Tile team has optimizations of Flash Attention algorithms. The computational bottleneck of every transformer.
1.66x speedup on Blackwell.
https://t.co/Ef8FJ4bE0c
🎉 We’re thrilled to announce the release of CuPy v14.0.0 — packed with new features, major performance boosts, important bug fixes, and more.
✅ Full compliance and support for NumPy v2 and Python Array API standard
✅ Support for CUDA pip wheels - much better installation footprint and interoperability with CUDA Python and PyTorch
✅ Support for structured dtypes & ML dtypes (starting with bfloat16)
✅ Better coverage over new NumPy & SciPy APIs
✅ Support for new cuFFT JIT callbacks - works on both Linux and Windows!
✅ New %gpu_timeit magic for Jupyter notebook users to properly benchmark GPU code
This release marks the CuPy 10th anniversary, highlighting our close collaboration between @PreferredNet, and the wider open source community. @CuPy_Team 🥳
Accelerate your Python program on our GPUs.👇
Anyone who has spent an hour with a two toddlers year olds trying to avoid a nap, has no problem believing AI will be super intelligent in the next decade.
CUDA Tile is the future of GPU programming, and it's in Python. 🐍
cuTile Python in CUDA 13.1 lets you focus on your algorithm, not Tensor Core specifics. Simplify your HPC/AI code now. ✨
🔗 https://t.co/5MTRNYGCRp
Proud to support #PyConUS
➡️ https://t.co/d9EWvkCkpO
Meet team in open spaces, summits, sprints & sessions:
🔹 Building transparent systems in Python
🔹 Scale smarter not harder with cuPyNumeric
🔹 Community-driven roadmap for Python packaging
🔹 GPU Programming in pure Python
🚀 Calling all CUDA developers! Whether you're a GPU programming newbie or a seasoned CUDA ninja, join our Meet Up!
📅 Jan 30, 5:30 PM PST
📍 Q Bay Center, 160 E Tasman Dr, San Jose, CA
🤝 Connect with the community and shape the future of CUDA.
https://t.co/9KXTj9Fbjs
Excited to announce my new role on the Admin Board of NumFOCUS! Recently, I posted about our initial meeting on my personal blog. Check it out here: https://t.co/dyrB7DMzYO