Home for "How To Scale Your Model", a short blog-style textbook about scaling LLMs on TPUs
-
Updated
Sep 19, 2026 - HTML
#
roofline
Here are 55 public repositories matching this topic...
Modular C++ Toolkit for Performance Analysis and Logging. Profiling API and Tools for C, C++, CUDA, Fortran, and Python. The C++ template API is essentially a framework to creating tools: it is designed to provide a unifying interface for recording various performance measurements alongside data logging and interfaces to other tools.
python c performance cplusplus cross-platform cpp analysis mpi cuda hardware-counters performance-measurement gotcha modular-design cross-language papi roofline cupti memory-measurements instrumentation-api
-
Updated
Jul 31, 2024 - C++
Analyze LLM inference: FLOPs, memory, Roofline model. Supports GQA, MoE, MLA, RoPE, SwiGLU. 19 models × 20+ hardware platforms.
-
Updated
Apr 16, 2026 - Python
Why is LLM inference slow — and how do you make it fast? A hands-on, first-principles course: roofline → KV cache → quantization → parallelism → vLLM/SGLang, with GPU labs on open models.
machine-learning deep-learning gpu cuda transformers self-hosting quantization roofline mlops inference-optimization kv-cache llm llmops llm-serving vllm llm-inference flash-attention speculative-decoding pagedattention sglang
-
Updated
Aug 11, 2026 - Python
Hand-written CUDA vs Mojo GPU kernels benchmarked on consumer Ampere (RTX 3090, sm_86), with roofline analysis
benchmark modular mojo gpu cuda nvidia high-performance-computing cuda-kernels performance-analysis ampere gpu-programming roofline kernel-optimization rtx-3090
-
Updated
Sep 1, 2026 - Mojo
Reproducible GPU performance workflow with Nsight CSV parsing, Roofline analysis and portable backend seams.
-
Updated
Aug 31, 2026 - Python
JAX benchmarking, profiling and evaluation metrics for Flax NNX: a registry of pure-function metrics (regression, classification, calibration, uncertainty, forecasting, generative, image, text, audio, graph, fairness), XLA FLOP counting, roofline analysis, GPU and energy monitoring, regression detection, publication exports, W&B and MLflow.
python benchmarking machine-learning metrics calibration statistical-analysis profiling uncertainty-quantification evaluation-metrics model-evaluation roofline jax energy-monitoring probabilistic-forecasting mlflow regression-detection flops wandb gpu-profiling flax-nnx
-
Updated
Sep 19, 2026 - Python
Interactive theoretical Kimi-K3 inference roofline calculator for H200, B300, and GB300
-
Updated
Aug 11, 2026 - Python
The physics before the engines: measured arithmetic intensity, prefill compute-bound vs decode memory-bandwidth-bound, with the roofline plot.
-
Updated
Aug 25, 2026 - Python
Benchmarked, verifiable CUDA exercises: correctness against a CPU reference, graded against your GPU's measured roofline.
-
Updated
Sep 17, 2026 - Python
Dense matrix multiply in four parallel models — hand-tiled OpenMP with roofline, Cannon's algorithm, hybrid MPI+OpenMP reaching 45 TFLOP/s on 16 nodes, and multi-GPU cuBLAS
-
Updated
Sep 16, 2026 - C++
Interactive 3D visualization of dense decoder-only LLM inference. Companion to the AI Inference Engineer 2026 course.
-
Updated
Jun 15, 2026 - TypeScript
Hand-written CUDA linear algebra library. A GEMM driven from a naive baseline to a compute-bound tensor-core kernel at ~95% of cuBLAS on an RTX 5070, profiled with Nsight Compute, plus GEMV, TRSM, CSR SpMV, cuSOLVER, a measured roofline, and NVML telemetry.
benchmarking hpc linear-algebra cuda cublas high-performance-computing cuda-kernels gpu-computing gemm ptx roofline tensor-cores blackwell nsight-compute wmma
-
Updated
Sep 4, 2026 - C++
Is your llama.cpp decode memory-bandwidth-bound? Find out in one command.
benchmark computer-architecture quantization memory-bandwidth roofline edge-ai llm llama-cpp local-llm gguf
-
Updated
Aug 27, 2026 - Python
Why code is slow, at four levels — an out-of-tree LLVM pass that predicts vectorisability from IR, cache-blocked GEMM against a roofline, and a CUDA FFT port
hpc llvm openmp cuda compilers performance-analysis vectorization llvm-pass roofline scalarevolution
-
Updated
Sep 16, 2026 - Cuda
Interactive macOS GPU benchmark suite for Apple Silicon - 25 Metal compute benchmarks (compute, memory, ML/tensor, ray tracing) with a full measurement harness and empirical roofline.
-
Updated
Jul 9, 2026 - Swift
Three hand-written Triton kernels for LLM inference (fused RMSNorm plus residual, online softmax, INT4 g128 GEMV) benchmarked on NVIDIA Blackwell against PyTorch eager and torch.compile, with every raw CUDA-event sample, measured device ceiling, and Nsight Compute report committed and CI-verified.
portfolio performance-engineering benchmarking gpu cuda pytorch triton gpu-kernels roofline nsight-compute gpu-profiling kernel-optimization int4-quantization
-
Updated
Aug 12, 2026 - Python
Benchmark workbench for MAX (Mojo) LLM decode kernels vs llama.cpp / cuBLAS / FlashInfer and the memory roofline on consumer NVIDIA GPUs (sm_86/sm_89). A public record, not a competing kernel library.
benchmark mojo cuda max nvidia quantization gpu-kernels roofline gemv llama-cpp llm-inference flash-attention gguf flashinfer q4-0
-
Updated
Sep 3, 2026 - Python
Add this topic to your repo
To associate your repository with the roofline topic, visit your repo's landing page and select "manage topics."