This free RAM cleaner uses native Windows features to optimize memory areas. It's a compact, portable, and smart application.

Run larger LLMs with longer contexts on Apple Silicon by using differentiated precision for KV cache quantization. KVSplit enables 8-bit keys & 4-bit values, reducing memory by 59% with <1% quality loss. Includes benchmarking, visualization, and one-command setup. Optimized for M1/M2/M3 Macs with Metal support.

VS Code extension: Go struct layout, padding, reorder

A fast and memory-optimized string library for heavy-text manipulation in Python

Keep ChatGPT fast — Firefox & Chrome extension that trims DOM in long conversations

Materials about memory optimization and zero-allocation samples.

Flash weight streaming for MLX: run massive models larger than your RAM on Apple Silicon.

NVIDIA Sol-Attn for ComfyUI / Triton kernel on SM89 - SM121, with zero-copy MiniMax H3 nodes: memory-efficient attention, scheduled tau with graph preview, and feed-forward chunking. Measured 1.14–1.44× vs SageAttention and −37% MLP peak VRAM on H3

Free up your memory for free!

This code repository contains the code used for my "Optimizing Memory Usage for Training LLMs and Vision Transformers in PyTorch" blog post.

First open-source implementation of Google TurboQuant (ICLR 2026) -- near-optimal KV cache compression for LLM inference. 5x compression with near-zero quality loss.

Lightweight Python lazy imports that defer module loading to reduce startup time and initial memory use.

A performant and memory efficient storage for immutable strings with C++17. Supports all standard char types: char, wchar_t, char16_t, char32_t and C++20's char8_t.

Simple RAM Cleaner

🚀 Zero-config automatic memory allocator for Rust - just add one line and get up to 1.6x faster allocation performance across all platforms

A curated list of awesome optimizations that you can do to improve your redis deployment (both client side and server side).

PRISM: O(1) Photonic Block Selection for Long-Context LLM Inference — eliminates the O(N) KV cache scan via photonic broadcast-and-weight similarity engine on TFLN

An extension of micro mouse on WEBOTS using the flood filled algorithm, A star, Dijkstra’s and Breadth first search algorithm for moving the E-puck robot from start to goal in an NxM sized maze whose map was unknown to the robot (mapping and path planning). Further, leveraged Error Correction for accurate turning and recursive Backtracking algor…

Memory List Manager

Unified KV-cache compression for LLM inference: 12 Python-native methods, Debian-tested isolated add-ons, Godzilla KVarN/TriAttention, exact Godzilla/Gigatoken profiles, CUDA weight sharing, and multi-GPU planning.