omega tau science & engineering podcast

[We have episodes in German and English] How do scientists uncover phenomena and explain their connections? How do engineers design machines, methods and infrastructure? At omega tau, experts give detailed answers. Over the last ten years, we have produced 300 episodes in which we dug deeper, until we ran out of questions. Join us on our journey through the world of science and engineering: the closer you look and listen, the more interesting things get.

http://omegataupodcast.net

subscribe
share






episode 300: How Processors Got So Fast


Guest: Lex Augusteijn    Host: Markus Voelter    Shownoter: Stefaan Rillaert

Have you ever wondered how the processor in your phone or computer got so much more faster than what the increase in megahertz suggests? In this episode we talk with Lex Augusteijn about superscalar processors, pipelining, speculative execution, register renaming and the like. We also discuss concerns other than speed, in particular, energy efficiency.

Introduction 00:06:00

Lex Augusteijn | KIM-1 computer | 6502 processor | Functional programming | Philips Research Eindhoven | Compiler | Media processor | DSP | Neuromorphic computing

Basics 00:10:53

Moore's law | Richard Feynmans' 'There's Plenty of Room at the Bottom' | Richard Feynman | "Surely You're Joking, Mr. Feynman!" | Gordon Moore | Clock speed | Transistor | Episode of 'The Freak Show' podcast about the 'monster 6502' | NAND gate | Fab | ASML | Episode about ASML | Wafer stepper | Processor | Computer memory | Hard drive | Graphics processing unit | Computer bus | Processor register | Branching | Instruction pipelining | Floating-point arithmetic | ALU | Instruction set | CISC | RISC | Compiler | Superscalar processor | Microcode | Von Neumann architecture | Harvard architecture

Speed optimizations in modern processors 00:42:11

Data width | SIMD | Pipeline stall | Intel 8080 | ARM architecture | Branch prediction | Two bit prediction scheme | Memory hierarchy | Clock cycle | CPU cache | Cache line | Write-Through cache line | Operating system | Cache coherence | Voatile keyword in C | Simultaneous multithreading (SMT) | Hyper-threading | Context switch | Single instruction, multiple data (SIMD) | Speculative execution | Register renaming | Very long instruction word (VLIW) | Intel Atom | Abstract interpretation | Static program analysis | Domain-specific language (DSL) | (German) episode about DSLs | C pragma | TriMedia mediaprocessor | Out-of-order execution | Instruction scheduling | Multi-core processor | The Free Lunch Is Over | MMX instruction set

Additional concerns 01:44:44

Application-Specific Integrated Circuit (ASIC) | Image processor | Dynamic voltage scaling | ARM architecture | Memory management unit (MMU) | Spectre bug | Side-channel attack | CUDA | Convolutional neural network | NXP Semiconductors | Processor design | Intel Tick-tock strategy


fyyd: Podcast Search Engine
share








 December 17, 2018  2h25m
 
 
curated by stiefkind in Dem stiefkind seine Lieblingspodcastfolgen | November 1, 2020