Tech News
← Home  ·  All topics

Cpu Architecture

5 GoKawiil briefs on this topic · RSS for cpu architecture

AI system designs new 'Frostbite' CPU architecture from scratch

A company's AI systems autonomously developed Frostbite, a general-purpose CPU with a new instruction set based on interaction nets, covering everything from compiler and ISA design through to final chip layout (GDS). The team reports a 4.59x performance-per-watt improvement over a silicon-proven dual-issue in-order superscalar CPU across 25 test workloads, 15 of which were held out until final evaluation.

Explainer breaks down how modern CPU components work together

A new explainer piece examines the internal architecture of modern CPUs, describing how components like the core, cache, and memory controllers interact to execute code and process data. The piece notes that today's processors are increasingly system-on-chip designs, bundling GPUs, AI accelerators and memory controllers alongside traditional CPU cores from makers like Intel, AMD, Apple and Qualcomm.

Startup Tendrils raises $30M to build massively parallel CPU architecture

Tendrils announced it closed a $30 million funding round led by Benchmark, with participation from Visionaries Club, UK Sovereign AI Fund, FR8, Antimatter, Cloudberry VC, Conviction, Embassy Ventures, SV Angels and angel investors. The company is designing a processor built around thousands of parallel cores rather than the sequential, single-core lineage of existing CPUs, paired with an open-source programming language called Vine meant for parallel-first chips. Jack Altman joins Tendrils' board as part of the announcement.

Technical blog argues database CPU bottlenecks often trace to memory latency, not raw compute

A technical blog post explains that a CPU showing 100% utilization may not actually be fully working, because much of that time can be spent waiting on memory accesses rather than executing instructions. It breaks down the latency gap between CPU cores and memory tiers, noting RAM access can take 50-100ns versus roughly 1ns for L1 cache, meaning a CPU could execute 400-800 instructions in the time it takes to fetch data from RAM once.

Fearless SIMD gains traction as author becomes maintainer, surveys Rust SIMD landscape

The author, who previously surveyed Rust's SIMD ecosystem, has since become a maintainer of the Fearless SIMD library after contributing to it. For this year's follow-up survey, they invited maintainers of competing libraries—std::simd, wide, pulp, and macerator—to review a draft, while keeping final editorial control. The piece revisits SIMD fundamentals and surveys progress across the Rust SIMD tooling landscape over the past year.