NaiveRT: a self-contained TP8 inference runtime for the Naive N series on eight Blackwell GPUs with UMMA support. One persistent CUDA graph per rank runs draft, eight-row verify, top-p sampling and cache commit with no host in the loop: 1,700 tokens/s at batch size 1, one-second peaks above 2,000.
☆30Sep 27, 2026Updated 3 weeks ago
Alternatives and similar repositories for NaiveRT
Users that are interested in NaiveRT are comparing it to the libraries listed below. We may earn a commission when you buy through links labeled 'Ad' on this page.
Sorting:
- ☆38Aug 4, 2026Updated 2 months ago
- A linker for partial GPU kernels.☆27Oct 12, 2026Updated last week
- Ulysses sequence-parallel all-to-all as a torch custom op, moved by the GPU copy engines into torch symmetric memory. Zero SM usage; 1.66…☆54Aug 25, 2026Updated last month
- TensorIR is a lightweight NVIDIA-owned MLIR compiler frontend for expressing tensor computations and lowering them to NVIDIA CUDA Tile IR…☆139Updated this week
- ☆57Sep 27, 2026Updated 3 weeks ago
- Managed hosting for WordPress and PHP on Cloudways • AdManaged hosting for WordPress, Magento, Laravel, or PHP apps, on multiple cloud providers. Deploy in minutes on Cloudways by DigitalOcean.
- A collection of megakernels for TPU☆164Updated this week
- Reference implementation and examples of the CuTe Layout representation and algebra.☆379Sep 30, 2026Updated 2 weeks ago
- Efficient and unified implementations for TopK-based sparse attention☆41Jul 10, 2026Updated 3 months ago
- ☆32Jul 2, 2025Updated last year
- DeepStack: Facilitating Co-Design Exploration of 3D DRAM-Stacked Accelerators for Distributed LLM Inference. Includes the MICRO 2026 AE a…☆65Oct 5, 2026Updated 2 weeks ago
- ☆116Aug 5, 2026Updated 2 months ago
- Communication patterns for AI, built on top of NCCL device and host APIs☆69Updated this week
- A Triton-only attention backend for vLLM☆28Jul 14, 2026Updated 3 months ago
- Curated high-performance TIRx GPU kernels☆32Updated this week
- Managed Kubernetes at scale on DigitalOcean • AdDigitalOcean Kubernetes includes the control plane, bandwidth allowance, container registry, automatic updates, and more for free.
- ☆129Updated this week
- An Open Compiler Harness for Agentic GPU Programming☆160Updated this week
- Muon in Int8 Precision Made Possible☆20Jun 18, 2026Updated 4 months ago
- An external memory allocator example for PyTorch.☆16Aug 10, 2025Updated last year
- ☆54May 19, 2025Updated last year
- Make FP4 on 5090 Great Again☆24Jul 20, 2026Updated 2 months ago
- NVSHMEM‑Tutorial: Build a DeepEP‑like GPU Buffer☆207Feb 11, 2026Updated 8 months ago
- FA4-based Relative Attention Kernel developed by TML and Colfax☆18Sep 11, 2026Updated last month
- A NCCL extension library, designed to efficiently offload GPU memory allocated by the NCCL communication library.☆121Dec 17, 2025Updated 10 months ago
- Open source password manager - Proton Pass • AdSecurely store, share, and autofill your credentials with Proton Pass, the end-to-end encrypted password manager trusted by millions.
- FLA but cuTile☆27Apr 17, 2026Updated 6 months ago
- A lightweight library for xPU kernel JIT compilation☆392Updated this week
- ☆46Aug 3, 2026Updated 2 months ago
- Production-ready MoE load balancing via real-time expert replication☆257Updated this week
- An experimental communicating attention kernel based on DeepEP.☆34Jul 29, 2025Updated last year
- Automatic formal verification for warp-specialized GPU kernels.☆51Updated this week
- High performance RMSNorm Implement by using SM Core Storage(Registers and Shared Memory)☆31Jan 22, 2026Updated 8 months ago
- TileFusion is an experimental C++ macro kernel template library that elevates the abstraction level in CUDA C for tile processing.☆119Aug 4, 2026Updated 2 months ago
- mKernel: fast multi-node, multi-GPU fused kernels☆296Updated this week
- Deploy on Railway without the complexity - Free Credits Offer • AdConnect your repo and Railway handles the rest with instant previews. Quickly provision container image services, databases, and storage volumes.
- Expert Specialization MoE Solution based on CUTLASS☆27Apr 14, 2026Updated 6 months ago
- Clangd for Ascend for Code Lint & Code Completion☆26Sep 29, 2026Updated 2 weeks ago
- Open source version of DOCA GPUNetIO and DOCA Verbs libraries (limited features) to enable GDAKI technology on RDMA (IB and RoCE)☆76Oct 5, 2026Updated 2 weeks ago
- MoonEP: A Perfectly Balanced Expert Parallelism Library via Dynamic Redundant Experts☆1,186Oct 12, 2026Updated last week
- Mixture-of-experts (MoE) training megakernel for NVL72s☆621Oct 1, 2026Updated 2 weeks ago
- Open deep learning compiler stack for cpu, gpu and specialized accelerators☆21Oct 7, 2026Updated last week
- LLM Inference via Triton (Flexible & Modular): Focused on Kernel Optimization using CUBIN binaries, Starting from gpt-oss Model☆119Apr 28, 2026Updated 5 months ago