⚡ Cache Optimization - Kaushik · Scour

Surviving the Performance Cliff of Disk-Bound Data 🚀Performance Engineering

tigerdata.com·6d

A Full-Stack Performance Evaluation Infrastructure for 3D-DRAM-based LLM Accelerators 🚀Performance Engineering

arxiv.org·6h

Runahead Execution vs. Conventional Data Prefetching in the IBM POWER6 Microprocessor (2010) 🔀Parallel Computing

pages.cs.wisc.edu·1d·Lobsters

Data Oriented Design by example (2017) ⚙️C++

nikitablack.github.io·3h·Hacker News

Cloudflare and ETH Zurich Outline Approaches for AI-Driven Cache Optimization 🚀Performance Engineering

infoq.com·1d

Low-Rank Key Value Attention: Reducing KV Cache Memory and Maintaining Head Diversity 🤖ML Systems

fin.ai·17h·Hacker News

CppCon 2025 Cache Me Maybe: The Performance Secret Every C++ Developer Needs -- Michelle D'Souza ⚙️C++

isocpp.org·6d

Compression technique makes AI models leaner and faster while they're still learning 🤖ML Systems

techxplore.com·15h

Principles of Mechanical Sympathy 🔀Parallel Computing

martinfowler.com·2d·Hacker News, Hacker News

AMD Ryzen 7 9850X3D Review: Zen 5's minor X3D refresh reigns as best gaming CPU while keeping Core Ultra 200S Plus at bay 🚀Performance Engineering

notebookcheck.net·3h

Advancing AI performance with HBM4, SPHBM4 DRAM solutions 🚀Performance Engineering

edn.com·19h

The Hidden Value of CPU-Intensive Compression on Modern Hardware 🎯Low Latency

klarasystems.com·1d

Optimising a Pipelined RISC-V Core: From Naive Pipeline to Near-Superscalar Performance 🚀Performance Engineering

mummanajagadeesh.github.io·3d·Lobsters, Hacker News

Bespoke OLAP: Using AI to Synthesize Workload-Specific Database Engines from Scratch 📈Trading Systems

ucbskyadrs.github.io·11h·Hacker News

SOTA Normalization Performance with Torch.compile 🚀Performance Engineering

pytorch.org·2d·Hacker News

Sorting performance rabbit hole 🚀Performance Engineering

nibblestew.blogspot.com·3d·Hacker News, nibblestew.blogspot.com

MysticCodingCat/CUDA-Native-HUBO: A GPU-native solver for 3-way combinatorial optimization (HUBO). Achieving digital-annealer-level performance on a single RTX 3060 Ti 🔀Parallel Computing

github.com·1d·Hacker News

The Insert Benchmark vs MariaDB 10.2 to 13.0 on a 32-core server 🎯Low Latency

smalldatum.blogspot.com·1d·smalldatum.blogspot.com

3x faster project loads with the origin private file system 🚀Performance Engineering

barndoors.lumafield.com·3d·Hacker News

Garbage Collection in the Era of Virtual Threads (Project Loom) 🔀Parallel Computing

blog.gceasy.io·1d

Loading more...