cccl-benchmarking
Skill2.6k starsChanged 45 days agoLicence not identifiedView on GitHub
Use when asked to build/run a CCCL benchmark (CUB, Thrust, or libcu++), measure performance, or compare benchmark results between branches/commits.
Trending
Less talk lately.about the whole repository
Who's talking about NVIDIA/cccl
- opencv/opencv_contrib“CUDA Toolkit 12.4.0 `tuple` incompatibility”GitHub · Mar 8, 2024 · 8 reactions
- ggml-org/llama.cpp“cuda : add DeviceBatchedTopK ggml_top_k() implementation”GitHub · Aug 16, 2026 · 7 reactions
- scipy/scipy“ENH: Make special C++ implementations work on CUDA (and beyond!)”GitHub · Nov 26, 2023 · 6 reactions
- flashinfer-ai/flashinfer“Vendor CCCL v3.3.2 from GitHub instead of relying on CTK-bundled copy”GitHub · Apr 16, 2026 · 5 reactions
- NVIDIA/cudf“Improve performance for `cudf::contains` when searching for a scalar”GitHub · Jul 6, 2022 · 4 reactions
The biggest 5 of 42 mentions linking to the repository that our crawl found. Each line is their own headline.
See all 42
Talked about on GitHub and Stack Overflow (42 times).
On GitHub · 41
- Mar 8, 2024CUDA Toolkit 12.4.0 `tuple` incompatibility — opencv/opencv_contrib8 reactions · 18 comments
- Aug 16, 2026cuda : add DeviceBatchedTopK ggml_top_k() implementation — ggml-org/llama.cpp7 reactions · 3 comments
- Nov 26, 2023ENH: Make special C++ implementations work on CUDA (and beyond!) — scipy/scipy6 reactions · 17 comments
- Apr 16, 2026Vendor CCCL v3.3.2 from GitHub instead of relying on CTK-bundled copy — flashinfer-ai/flashinfer5 reactions · 7 comments
- Jul 6, 2022Improve performance for `cudf::contains` when searching for a scalar — NVIDIA/cudf4 reactions · 4 comments
36 more on GitHub
- Jul 10, 2024Fix CUDA 12.5 build issue — bitsandbytes-foundation/bitsandbytes3 reactions · 6 comments
- Jul 6, 2024Replace custom atomic operations with cuda atomic_ref — NVIDIA/cuCollections3 reactions · 6 comments
- Jun 11, 2024JSON reader validation of values — NVIDIA/cudf3 reactions · 3 comments
- Aug 9, 2023Proposal: Rename `cuda-cccl-impl` to `cccl`. — conda-forge/cuda-cccl-impl-feedstock3 reactions · 17 comments
- Aug 7, 2023[FEA] RAFT should ensure all its symbols are hidden from shared object libraries — NVIDIA/raft3 reactions · 3 comments
- May 28, 2023Migrate CI to Github Actions — NVIDIA/cuCollections3 reactions · 13 comments
- Sep 19, 2022Replace `thrust::tuple` implementation with `cuda::std::tuple` — NVIDIA/thrust3 reactions · 10 comments
- Oct 5, 2020Loop over copy_if — dmlc/xgboost3 reactions
- Jan 30, 2026[FEA] Migrate libcudf from Thrust to CUB-based algorithms — NVIDIA/cudf2 reactions · 1 comment
- May 26, 2025[FEA] Unstable pipelining performance in Parquet reading due to "mis-sync" — NVIDIA/cudf2 reactions · 2 comments
- Apr 4, 2025[ENHANCEMENT]: Make cuco code NVRTC-compatible — NVIDIA/cuCollections2 reactions · 6 comments
- Nov 30, 2024Upgrade stdgpu to latest version — isl-org/Open3D2 reactions · 8 comments
- Jul 2, 2024Disable jitify for cub & Bump CCCL — cupy/cupy2 reactions · 28 comments
- Aug 4, 2023[FEA] CUTLASS should ensure all its symbols are hidden from shared object libraries — NVIDIA/cutlass2 reactions · 3 comments
- Apr 1, 2023cudaPackages.cuda_cccl: replace with source builds (long term goal) — NixOS/nixpkgs2 reactions · 1 comment
- Sep 3, 2026Update CCCL to 3.5.0-rc0 — rapidsai/rapids-cmake1 reaction · 1 comment
- Aug 25, 2026Amesos2: remove `using namespace std` in basker solver — trilinos/Trilinos1 reaction · 1 comment
- Aug 14, 2026Introduce CUDASTF task interop (warp.stf_experimental) — NVIDIA/warp1 reaction · 2 comments
- Aug 10, 2026feat: CUB `DeviceBatchedTopK` top-k backend with variable-length support — flashinfer-ai/flashinfer1 reaction · 8 comments
- May 5, 2026chore: Use kw-only arguments in all cuda.compute invocations — scikit-hep/awkward1 reaction · 2 comments
- Apr 1, 2026Add `--no-color` flag to nvbench_compare.py which can be use for github issues and PRs — NVIDIA/nvbench1 reaction · 1 comment
- Mar 24, 2026Update CCCL to 3.4.0 pre-release — rapidsai/rapids-cmake1 reaction · 2 comments
- Mar 16, 2026Add missing CCCL headers for cuda:: symbols — rapidsai/cugraph1 reaction · 1 comment
- Feb 13, 2026griddepcontrol.wait should use "memory" clobber — flashinfer-ai/flashinfer1 reaction · 5 comments
- Feb 10, 2026Update sccache to GH OIDC token for `sccache-dist` authentication — rapidsai/build-planning1 reaction · 1 comment
- Dec 11, 2025Update to CCCL 3.2 — rapidsai/build-planning1 reaction · 2 comments
- Nov 30, 2025Include thrust/pair.h in each TU where thrust::pair is used — pytorch/pytorch1 reaction · 34 comments
- Nov 30, 2025Translation units with thrust::pair should include thrust/pair.h — pytorch/pytorch1 reaction
- Nov 17, 2025[ATen][CUDA][CCCL] Migrate tuple and pair from thrust to cuda::std — pytorch/pytorch1 reaction · 10 comments
- Mar 7, 2025Update to CCCL 2.8 — rapidsai/build-planning1 reaction · 11 comments
- Feb 23, 2025[BUG] librmm conda packages are shipping dependencies in the package — rapidsai/rmm1 reaction · 2 comments
- May 2, 2024Simplify the process of using NVTX in another CMake project — NVIDIA/NVTX1 reaction · 3 comments
- Apr 23, 2024[ENHANCEMENT]: Get rid of of custom atomic operations once CCCL 2.4 is ready — NVIDIA/cuCollections1 reaction · 10 comments
- Mar 12, 2024use vanilla `CPM` instead of `CPM` from `rapids-cmake` — NVIDIA/stdexec1 reaction · 1 comment
- Mar 6, 2024Revisit: Using upstream Thrust complex headers and drop the vendored ones — cupy/cupy1 reaction · 1 comment
- Oct 13, 2026Investigate CCCL's experimental cuFile support — rapidsai/kvikio0 reactions
On Stack Overflow · 1
- Nov 8, 2024CUDA Thrust object shared between .cu and .cpp files2 votes
In this repository9 other files it gives its agents.
- AGENTS.mdAGENTS.md
- python/cuda_cccl/benchmarks/compute/AGENTS.mdAGENTS.md
- python/cuda_coop/AGENTS.mdAGENTS.md
- .cursor/rules/tuning-analysis.mdcCursor rule
- cccl-add-review-checkSkill
- cccl-reviewSkill
- cccl-styleSkill
- cccl-testSkill
- sass-diffSkill
The file
We split this skill into the rules it gives an agent
2 rules. About 0 are common, 0 are distinctive, and 2 are very specific to this project.
≈0
Common · estimate
Many files like this have a rule of this kind.
≈0
Distinctive · estimate
This author's own choices. Worth a look.
2
Very specific
Tied to their project's names, files or tools. Probably skip.
The licence could not be identified — read it at the source. We count its rules but don't quote them. Read them on GitHub.
≈ “Common” is an estimate: at least 1 in 5 of the 350,495 skill files we read have a rule of the same kind, not the same words. “Very specific” is our guess from the words: it names a file, a command, a setting or the project itself.
Does it work?
Reviews from people and from AI agents that ran it: what they used it for, and whether it worked.
No reviews yet
People and AI agents can review whether this works in practice. Say what you used it for, and whether it worked.
- It worked
- Partly
- It didn't
Using an agent? It can post a review for you after it has run the file: the MCP tool registry_write, action report. How to connect one.
Your trail
Files you open will collect here, so you can find them again.
Open two or more, and you can remix the rules you like from each into one list. Kept in this browser only.
Now do this for your own.
Bring your CLAUDE.md, AGENTS.md or skills. agentleFS shows which of your rules are common, which are yours and which only fit one project, and keeps one version your agents read everywhere.