cache

Scott Meyers explains how CPU caches work and why they matter for performance, covering cache levels, access speeds, associativity, and effects in multithreaded code. The focus is C++ and cache-aware, data-oriented programming.

Bryan Cantrill presents the ARC (Adaptive Replacement Cache) paper by Megiddo and Modha at Papers We Love. He opens with a survey of caching strategies before digging into the self-tuning LRU/MFU design.

Matt Kulukundis explains how Google designed SwissTable, a fast, cache-friendly hash table, step by step from std::unordered_map. The talk covers open addressing, SIMD probing and the tradeoffs of changing a fundamental data structure.

Cliff Click walks through a small performance example on a modern out-of-order CPU. He covers pipelining, ILP, CISC vs RISC, multicore, and why cache misses and the memory subsystem dominate performance.

Mike Acton of Insomniac Games argues that a program's purpose is transforming data, and shows how common C++ approaches work against that in performance-critical game development. He covers cache behavior, data layout and the limits of compilers.

Chandler Carruth's CppCon 2014 talk on writing fast C++, explaining why data structure choice and memory layout matter more than clever algorithms. It covers why standard containers like unordered_map perform poorly on modern CPUs.

Optiver's Carl Cook explains how to build high-frequency trading systems in C++, where the critical path runs rarely but must be extremely fast. He covers techniques for low and predictable latency, including avoiding main memory and cache misses.

Andrew Kelley, creator of Zig, presents practical data-oriented design techniques for shrinking memory footprint and making programs cache friendly. He ends with a case study of how these ideas sped up the Zig compiler.

Matt Godbolt builds a toy path tracer in C++ in three different styles and compares them on clarity, testability, compile time and run-time performance. The talk was given at CppCon 2019.

Chandler Carruth's CppCon 2016 talk on designing hybrid data structures for performance. He covers cache locality, working-set size and allocation overhead, and why standard C++ containers fall short.

Gor Nishanov's CppCon 2018 talk on using tiny C++ coroutines to hide memory latency, interleaving useful work while prefetches bring cache lines in from main memory.

Fedor Pikus's CppCon 2017 talk on C++ atomics, looking at what they let the programmer express and what the hardware actually does. It covers memory ordering and lock-free programming from basic to advanced.

David Schachter describes how he made a data-collection Python program run 114,000 times faster, drawing on decades of experience. The talk covers pipelining, cache effects and other optimization lessons.

Netflix engineer Drew Gallatin's EuroBSDcon 2022 talk quantifies the older FreeBSD optimizations that let Netflix serve web traffic at 800Gb/s, including TSO, LRO, VM page cache zones, asynchronous sendfile and unmapped mbufs.

Aaron Boodman of Replicache asks why sync engines have not yet fixed web app complexity and what is missing for them to go mainstream. The talk lays out the ideas that led to the Zero sync engine.

Armin Ronacher, creator of Flask, explains how CPython's internals and C API leak into the language's semantics. Using examples like the complexity of the add operation, he shows why Python is hard to optimize or reimplement.

Andrei Alexandrescu's CppCon 2015 talk argues that std::allocator is badly designed. He proposes building high-performance allocators from small composable, templated pieces.

A DevGAMM talk by Jonathan Blow arguing that civilizations can lose technological capability and that software quality is already deteriorating. He discusses how knowledge fails to transfer between generations.

R. Stanley Williams of HP Labs gives a keynote on the memristor, covering the physics, crossbar computation, and potential for memory and new computer architectures.

Emery Berger explains why performance is hard to measure and deliver on modern hardware, showing how code layout and alignment can swamp optimization effects. He presents Stabilizer for statistically sound evaluation and Coz, a causal profiler.

Casey Muratori's long talk arguing that modern software has ballooned to tens of millions of lines because of hardware and OS diversity. He proposes a stable ISA for entire system-on-a-chip packages and recalls when games shipped with their own boot disk.

Michael Steil's 33c3 talk explains essentially everything about the Game Boy: CPU, video hardware, peripherals, and the tricks used by games and demos. It revives the era of counting clock cycles around hardware limits.

Rick Reed of WhatsApp explains at Erlang Factory 2014 how the service scaled to billions of messages with a very small engineering team. He covers the FreeBSD and Erlang stack, Mnesia usage, server counts and bottleneck work.

Gil Tene of Azul Systems explains common pitfalls in measuring latency and response time, including why averages mislead and the problem of coordinated omission. He introduces open source tools like HdrHistogram for better measurement and reporting.

Timothy Roscoe of ETH Zurich delivers the joint USENIX ATC/OSDI 2021 keynote arguing that operating systems must reckon with modern hardware, which is a heterogeneous collection of processors, firmware and coprocessors rather than a simple machine. It shows how much runs below and beside Linux.

Michael Steil and Christian Hessmann's 34C3 talk explaining the Apollo Guidance Computer end to end: hardware design, operating system, and mission software.

Jim Keller gives a UC Berkeley EECS colloquium arguing that Moore's Law is not dead. He surveys upcoming technologies and the multiple innovation curves that can keep transistor density and performance scaling.

Jonathan Blow's DevGAMM 2019 talk argues that software knowledge is being lost between generations and that growing layers of abstraction could lead to a decline in technological civilization.

Chuck Moore, inventor of Forth, presents GreenArrays' 144-core asynchronous chip at Strange Loop 2013. He explains how idle cores draw almost no power and how tight code minimizes energy use.

Casey Muratori's two-hour talk tracing the history of object-oriented programming back to the 1960s and examining how its ideas, and later entity component systems, came about. It is followed by a Q&A.

Christopher Domas presents at Black Hat 2017 how page fault analysis and processor fuzzing can exhaustively search the x86 instruction set. The talk uncovers undocumented instructions and hardware bugs and shows limits of disassemblers.

Microsoft's Tony Chen explains how the Xbox One was designed to resist physical attacks, including secure boot, encrypted memory, a security processor and a hypervisor. The goal was to make any hardware attack cost more than the price of many games.

Chandler Carruth argues in a CppCon 2019 talk that C++ has no truly zero-cost abstractions. Using examples like unique_ptr, ABI constraints and compile-time costs, he shows that abstractions cost runtime, compile time or human effort.

Andrei Alexandrescu's CppCon 2019 talk revisits sorting, showing how branchless tricks, hybrid strategies and a cost metric can speed up std::sort-style algorithms on modern hardware.

Jake Archibald's JSConf.Asia 2018 talk "In the Loop" explains the browser event loop, covering tasks, microtasks, requestAnimationFrame, and how rendering interleaves with JavaScript. It uses visual walkthroughs to show why code runs in surprising orders and why setTimeout fixes are unreliable.

Herb Sutter's deep C++ and Beyond 2012 talk on atomics and the C++ memory model, covering why programmers chasing top performance need to understand hardware ordering. This is part one of two.

Bryan Cantrill of Oxide Computer Company gives a Stanford EE380 talk on rethinking the server. He argues that hyperscaler-style rack-scale machines with open firmware, RISC-V and Rust should be available to everyone, not just the giants.

Richard Feldman explains how pure functional languages can match or beat imperative ones on performance. Compilers can turn immutable code into in-place mutation, avoiding naive copying. The talk draws on his work on the Roc language.

Cliff Click of Azul Systems gives a Google Tech Talk on running Java business applications on a machine with about a thousand cores. He covers hardware and software codesign, including garbage collection hardware and the JIT.

A 2007 Google Tech Talk in which Cliff Click of Azul Systems presents a lock-free concurrent hash table that scales to 768 CPUs. He explains why it is correct by reasoning in a different way than usual happens-before and memory-order arguments.