DeepSeek-V3, R1 671B on 8xH100 Throughput Benchmarks
-
Updated
Mar 13, 2025 - Python
#
h100
Here are 89 public repositories matching this topic...
From-scratch, heavily-annotated CUDA inference runtime for Qwen2.5-Coder-7B on H100 (sm_90). Custom INT4 packer, fused GEMV, paged KV, split-KV attention, CUDA graph decode — every hot path commented for the why. Educational, not a llama.cpp replacement.
-
Updated
Aug 11, 2026 - Cuda
Dashboard for AI Studio, Open Source Continuous Inference | Deepseek-R1, Qwen2.5, Llama3.1 | 4xRTX-5090 inside PRU2500, 2xH100 inside PRU2500, 8xMI210 in SuperMicro
amd jupyter-notebook nvidia p100 jupyter-lab a100 p40 h100 rtx-5090 gpu-benchmarking rtx5070 l40s mi210
-
Updated
Aug 7, 2026 - TypeScript
LLM benchmarking, GPU workload orchestration backend server | Deepseek-R1, Qwen2.5, Llama3.1 | 4xRTX-5090 inside PRU2500, 2xH100 inside PRU2500, 8xMI210 in SuperMicro
t4 amd jupyter-notebook nvidia p100 jupyter-lab a100 p40 h100 rtx-5090 gpu-benchmarking rtx-5070-ti mi210
-
Updated
Aug 18, 2026 - Python
Sub-real-time MiniMax H3 on Hopper: 13.506 s for a 14.375 s 768p video with audio on 8xH100. Four runtime patches removing 9.72 s of non-model overhead from FastVideo.
hopper video-generation inference-optimization diffusion-models fastvideo h100 video-diffusion minimax-h3
-
Updated
Aug 31, 2026 - Python
Multi-GPU tensor-parallel vLLM on AMD Radeon RX 7900 XT / XTX / GRE (7900XT, 7900XTX, RX7900XT, gfx1100, RDNA3, ROCm): root cause and fix for the RCCL hostcall / PCIe atomics (AtomicOps) crash "NCCL error: unhandled cuda error" / "operation cannot be performed in the present state", Proxmox VFIO passthrough; LLM inference benchmarks on 13 machines
benchmark amd hip proxmox rccl multi-gpu vfio rocm radeon gpu-passthrough tensor-parallelism vllm 7900xtx llm-inference gfx1100 h100 rdna3 pcie-atomics 7900xt rx7900xt
-
Updated
Oct 5, 2026 - Python
A Flexible and High-Performance Inference Serving Engine for Open Diffusion Language Models
cuda inference-engine b200 nvdia h200 h100 gh200 diffusion-language-models dllm-infra llada2 dllm-serving diffusion-gemma nemotron-diffusion
-
Updated
Oct 5, 2026 - Python
Rent ready-to-use cloud GPUs in seconds. Lium CLI makes it easy to launch, manage, and scale GPU compute directly from your terminal. Fast, cost-optimized, and built for AI & ML developers.
python cli machine-learning sdk deep-learning gpu devtools cloud-gpu h200 gpu-cloud cloud-gpus h100 ai-infrastructure gpu-rental cloud-compute
-
Updated
Oct 6, 2026 - Python
Faster attention kernels for serving TML's Inkling model on vLLM. 2.7x over the shipping path on H100, and the only implementation that runs on A100.
gpu cuda inference attention hopper cuda-kernels cutlass kernels inkling gpu-optimization mlsys llm-serving vllm flash-attention h100
-
Updated
Sep 18, 2026 - Python
CLI tool to check Oracle Cloud compute shape availability - find capacity across regions
-
Updated
May 11, 2026 - Python
Bunch of explanations and tutorials around confidential computing
-
Updated
Apr 11, 2025 - Rust
PyTorch DDP and multi-GPU LLM training — from a minimal distributed example to NanoGPT speedruns, Muon, H100 profiling, and modded-nanogpt. FBA LAB https://bubblnet.com
training gpu modal pytorch ddp pytorch-tutorial gpt-2 distributed-data-parallel marin nanogpt llm-training h100 qwen deepseek olmo llm-systems muon-optimizer modded-nanogpt
-
Updated
Aug 20, 2026 - Python
Demo for installing ComfyUI on Azure VM powered by Nvidia H100 to run Text to Image and to Video models like Z-Image, Qwen and Wan
-
Updated
Feb 2, 2026 - HCL
From-scratch 125M LLM: 300K steps, 57.8B-token corpus, 282K tok/s on 2x H100
-
Updated
Aug 13, 2026 - Python
In the recent competition, we were challenged to finetune a model that can convert a LaTeX expressions into Python code effectively. My team, which I led, secured 6th place overall.
-
Updated
Aug 30, 2024 - Jupyter Notebook
High-performance Triton kernels for NVIDIA H100. Implements fused FP8 LayerNorm, tiled FlashAttention, and SRAM-optimized memory primitives for Hopper architecture.
-
Updated
Oct 1, 2026 - Python
Collection of Flash attention 3 precompiled wheels for direct install on Hopper series NVIDIA GPU (H100, H200 sm_90a)
cuda pytorch pytorch-wheel flash-attention h100 flash-attention-3 flash-attn h200-ai-chips flash-attention-wheels hopper-sm90a
-
Updated
Sep 2, 2026
Add this topic to your repo
To associate your repository with the h100 topic, visit your repo's landing page and select "manage topics."