performance-modeling

Open Source Architecture Code Analyzer

A highly-flexible GPU simulator for AMD GPUs.

Extra-P, automated performance modeling for HPC applications

A framework to find good combinations of optimizations for computational kernels on GPUs.

This python package provides various tools to predict energy expenditure and recovery dynamics of an athlete. The name pypermod stands for Python Performance Modeling.

Nebula: a microarchitecture simulator built from loosely coupled microservices

Performance Modeling and Sandboxing for Microservices Applications

Dnnamo: Tools for Analysis, Modeling, and Optimization of Deep Neural Networks

Taint-based program analysis framework for empirical performance modeling.

Modelling parallel processing with GPU

CM-DARE is a measurement infrastructure for monitoring distributed training in Google Cloud (ICDCS'20)

Cycle-Accurate Simulator of the Atalla-0x01

When does network, reliability, or recovery investment beat buying more GPUs? Accounting framework, substitution metric, and decision maps for AI training infrastructure. Paper + reproducible model.

Open-source NAND flash controller architecture lab with an executable FTL and channel scheduler simulator, Linux PCI/blk-mq driver scaffolding, fio workloads, and a roadmap toward QEMU, firmware, and FPGA validation.

A more abstract interpretation and formalization of the established three component hydraulic model by Morton. We remove its ties to concrete metabolic measures and use evolutionary computation to fit its parameters to an athlete.

Source code for the work "dSpark: Deadline-Based Resource Allocation for Big Data Applications in Apache Spark" published in IEEE e-Science 2017

Silicon workbench for chip blueprints, roofline analysis, tile-memory flow, and AI accelerator co-design.

Trace-driven research harness for KV-cache hierarchy policy evaluation in long-context LLM inference.

Discrete-event simulator and capacity planner for vLLM-style LLM serving (continuous batching, paged KV cache, chunked prefill, prefix caching) with a roofline timing model. Compares against Erlang-C sizing.

Research framework for cost-aware Apache Spark self-tuning, combining Bayesian Optimization, Safe Transfer Learning, and hybrid metaheuristics for adaptive performance optimization.