Triton
Fewer than 5 people recommend it so far, so there's no written verdict. Here is what commenters said.
2recommend
1warn against
Nov 7, 2025first mention
Aug 12, 2026latest mention
Why people recommend it
“For more generic GPU targets there's TRITON”
Caveats and dissent
“Focusing exclusively on Triton and making HIP a second class citizen is non-sensical.”
Often compared with
Who recommends it in their own words
From the makers not counted in the ranking
“And Triton object storage will have S3 support in the next release!”
Threads where it came up
- Helion: A high-level DSL for performant and portable ML kernels 3 mentions
- Mojo 1.0 Is Here 1 mention
- Oxide Computer raises $445M (SEC Form D) 1 mention
- Taking on CUDA with ROCm: 'One Step After Another' 1 mention
- Show HN: RBF-Attention – Trading dot-products for Euclidean distance 1 mention
- Show HN: Eyot, A programming language where the GPU is just another thread 1 mention
Quotes are verbatim; the verdict and theme labels are written by a model from these comments, and every clause links to its sources.
All comments 10
“consider Triton from openai”
“as a former Triton/SmartOS user”
“Focusing exclusively on Triton and making HIP a second class citizen is non-sensical.”
“wrote a custom Triton kernel”
“For more generic GPU targets there's TRITON”
“And there's also Triton[0] and vanilla SmartOS[1] it's based on”
“how well Triton works varies depending on what you want to do”
“Pallas is the Triton equivalent in JAX land.”
“Triton already has auto-tuning”
“Triton, but simpler and with automatic autotuning.”