Profine automatically profiles and optimizes PyTorch training jobs on real GPUs, delivering measurable speedups and lower GPU costs before teams waste days tuning configs by hand.

NAV extracts and analyzes GPU performance traces from NVIDIA Nsight™ Systems (NSYS), enabling comparative analysis and visualization for efficient performance profiling and regression testing.

Agent Skill + Claude Code Plugin for debugging Chrome WebGPU on macOS — Chrome DevTools MCP + Xcode Metal tracing

Hands-on ML accelerator profiling labs using PyTorch Profiler, mixed precision benchmarking, Google Colab, and Perfetto.

Collection of examples and links that uses different profiling tools to show memory usage and timings.

NAV extracts and analyzes GPU performance traces from NVIDIA Nsight™ Systems (NSYS), enabling comparative analysis and visualization for efficient performance profiling and regression testing.

Communication cost modeling for tensor parallel LLM inference with TP vs PP vs hybrid comparison, VRAM analysis, pipeline bubble modeling, regime detection, and cost-efficiency. Shows TP dominates on NVLink, PP has 47% bubble at 8 GPUs, and LLaMA-70B needs 8× A100 or 2× H100 for VRAM.

Capture and analyze Metal GPU frames from Unity on macOS — an Editor window + AI Assistant skill driving macOS 27 gpucapture/gpudebug (no Xcode): real GPU frame/pass timing, frame-budget gauge, GPU bottleneck classification, and deterministic Top-3 URP optimization insights.

Automated GPU profiling analysis for Adreno — turns Snapdragon Profiler captures into actionable insights with LLM

Unified benchmarking and profiling framework for the JAX scientific ML ecosystem. Timing, GPU/energy monitoring, FLOPS counting, roofline analysis, statistical testing, regression detection, and CI integration.

Complexity Assessment of LC methods on CPU and GPU

Real implementation of speculative decoding showing that speedup is strongly prompt- and target-dependent: GPT-2 → GPT-2-medium achieves mean best speedup 1.013×, while GPT-2 → GPT-2-large reaches 1.253× and up to 1.846× on high-agreement prompts.

Empirical profiling of chunked prefill showing two regimes: for isolated requests, full prefill always wins TTFT; under mixed workloads, chunked prefill improves fairness by reducing short-request TTFT, with chunk256 offering the best tradeoff.

Profiling and Triton-based KV-cache optimization for protein language model inference on consumer GPUs.

Attention backend benchmark on Turing GPUs comparing Vanilla, SDPA Math, SDPA Efficient, and a custom Triton FlashAttention implementation. SDPA efficient achieves 130× memory reduction and 10× speedup; Triton FA achieves O(n) memory but is 64× slower than SDPA efficient on RTX 2070.

"The GPU Watchers swore upon their shared memory hierarchy, from L1 to global memory, which also served as their mandate as lords of parallel computation."

Long-context benchmark pushing Qwen2-0.5B from 4K to 32K tokens on RTX 2070 using SDPA + chunked prefill. Shows 40x speedup at 8K, FP16 beating INT4 at long context, and that quantization is NOT a long-context solution — KV-cache is the real bottleneck.

C++23 Vulkan renderer for glTF/BIM/USD scenes with PBR materials, render graph, GPU culling, telemetry, and debug visualization.

Empirical profiling of FP32, FP16, INT8, and INT4 quantization on GPT-2 and GPT-2-medium: throughput, decode latency, model memory, and perplexity across batch sizes 1–16. FP16 wins on throughput; INT4 wins on memory with negligible quality loss.

This project demonstrates the integration of a CUDA kernel within an NVIDIA Holoscan application. It consists of two custom operators: one for memory allocation and data initialization, and another for executing the CUDA kernel. The application was profiled using Nsight systems and the kernel with Nsight compute