POSIX Asynchronous I/O
- Sep 22, 2006
- Inside Modern X11 Programming
- Making JavaScript Fast, Part 2
- Security in Your Pocket: OpenBSD on ARM
- Making JavaScript Fast, Part 1
- The Failure of the GPL
- How Not To Optimize
- A Half-Way Step to Apple’s Source Code: An Interview with David Chisnall
- Advanced Flow Control for Objective-C
- Erica Sadun on the iPhone SDK, OS X, and the Computing Landscape
- From NeXTSTEP to Cocoa: Erik Buck on the Development of Cocoa and Objective-C
- Fun with the Objective-C Runtime
- Marcus Zarra and Matt Long on Core Animation
- Steve Kochan on the Evolution of Objective-C
- The Technology NeXT Gave the World
- Where the Web and the Desktop Meet: An Interview with Lee Barney
- Pandora: An Open Console
- The Future of Wireless Networking
- GNU or Linux?
- Debugging C-Family Languages
- How Small Is Your PC? The Rise of Netbooks and Other Small Form-Factor PCs
- David Chisnall's CPU Feature Wishlist
- The Dynamic Languages Renaissance
- Robert Seacord on the CERT C Secure Coding Standard
- Objective-C for C++ Programmers, Part 3
- Objective-C for C++ Programmers, Part 2
- Objective-C for C++ Programmers, Part 1
- Writing Insecure C, Part 3
- Writing Insecure C, Part 2
- Writing Insecure C, Part 1
- iRex iLiad e-Reader: Linux's Answer to the Kindle?
- How It Works: Filesystems
- How the LLVM Compiler Infrastructure Works
- How It Works: Virtual Memory
- What Is C For?
- The Future of eBooks
- Imagining an Open Network
- Understanding How Xen Approaches Device Drivers
- Examining the Legendary HURD Kernel
- Competition Among Open Source Compilers
- Inside Your OS: What is a Process Scheduler, and How Does it Work?
- Bad UI of the Week: Read This (OK/Cancel)
- The End of the Desktop Era
- The What and Why of Open IM
- A Look at the Modern X Server
- The Future of Digital Media
- The Future of Identity
- Bad UI of the Week: Ask Forgiveness, Not Permission
- Copyright Versus Free Software
- Is Computer Science Dying?
- A Brief History of Programming, Part 2
- A Brief History of Programming, Part 1
- The 700MHz Question: Will the Wireless Spectrum Auction Lead to Innovation or More of the Same?
- Bad UI of the Week: The Menu Bar
- The Dark Corners of x86
- Bad UI of the Week: The Cross-Platform User Interface
- Bad UI of the Week: The Mythical "is Like" Operator
- Bad UI of the Week: Don't Make Me Tell You Twice...
- Bad UI of the Week: Kettles and Washing Machines
- The BBC iPlayer Controversy Explained
- Bad UI of the Week: The Mitten Mouse
- Bad User Interface of the Week: File It Under “Bad”
- Bad User Interface of the Week: The DVD
- A Roundup of Free Operating Systems
- DragonFly BSD: UNIX for Clusters?
- CPU Wars, Part 3: Put Your Left ARM In
- CPU Wars, Part 2: POWER to the People
- CPU Wars, Part 1: When the Chips Are Down
- ZFS Uncovered
- Vector Programming with GCC
- Free Software Versus Open Source Software
- What Programming Languages Should You Know?
- Standardizing UNIX
- Prolog: Logic Programming for Rapid Development
- POSIX Parallel Programming, Part 3: Threads
- POSIX Parallel Programming, Part 2: Message Passing
- POSIX Parallel Programming, Part 1
- The Nokia 770 Revisited
- The Open Source Desktop Myth
- Separating Style and Content: LaTeX and Typesetting
- GNUstep: A Free Software alternative to OpenStep
- Behind the Scenes of Objective-C 2.0
- The Future of CPUs: What's After Multi-Core?
- What Makes a Good Programming Language?
- Emulation: Role-Playing for Computers
- NetBSD: Not Just for Toasters
- POSIX Asynchronous I/O
- Breaking Down GPL Version 3
- The Role of Binary Drivers in a Free OS
- Security Is a UI Problem
- Debunking the Myth of High-level Languages
- A Taste of Erlang, a Dynamic, Asynchronous Message-Passing Language
- Alternatives to LAMP
- BSD Packaging Systems
- DRM: Digital Rights or Digital Restrictions?
- Introducing OpenBSD 3.9
- The Need for Virtualization and Xen
- Making Effective Software TCO Calculations
- 10 Things I Hate About U(NIX) Revisited: Readers Speak
- Comparing Open Source Licenses: GPL vs. BSDL
- BSD: The Other Free UNIX Family
- Measuring the Effectiveness of Application Security Policies
- The Cost of Free Software
- Nokia 770 Internet Tablet Week-long Test Drive
- 10 Things I Hate About (U)NIX
- The Lure of Open Source Software: Why Consider It for Your Business?
Sorry, this author hasn't posted any blogs.
Like this article? We recommend
Advanced Programming in the UNIX® Environment, 2nd Edition
Since I first learned to program, a lot has changed in the way computers work. While this fact isn’t always apparent to the user, it means that optimizations I was taught now result in highly suboptimal code. One thing, however, has remained constant: Hard drives are slow.
In recent years, this situation has gotten worse rather than better. CPU speeds have grown faster exponentially, while hard drives have barely managed even a linear speed increase. DMA transfers, becoming common, have widened the gap even more; now CPUs don’t have to do nearly as much work when handling disk I/O.
With the traditional UNIX I/O mechanisms, you issue a read or write system call, and then wait until the call has completed. A typical hard drive takes about 9 ms to position the head and then a little longer to perform the operation—and that’s assuming a linear. A modern CPU runs at least 1 GHz. In the time it takes to position the disk head, a 1 GHz CPU runs for nine million clock cycles. You can get a lot done in that many cycles.
The obvious solution is to dispatch the read or write requests as soon as you know what data you’re going to need, and then do something else while the OS processes the request. The asynchronous I/O (AIO) APIs, provided as part of the POSIX real-time extensions, allow you to do exactly that.
Operating System Support
Commercial UNIX systems have supported POSIX asynchronous I/O. Some of the best documentation currently available comes from IRIX and HP/UX. More recently, Free UNIX variants have started to support it; Linux and FreeBSD both support most of the specification, for example.
Support is slightly different between operating systems. Code using AIO on some platforms, including Linux and Solaris, must be linked with -lrt to provide support for the POSIX real-time extensions. On FreeBSD, AIO support isn’t built into the default kernel and must be loaded as a module with the following command:
# kldload aio
On versions of Mac OS X prior to 10.3, AIO wasn’t supported at all, and until OS X 10.4 AIO operations were supported only on files, not other file descriptor types such as sockets. Using AIO can give significant performance benefits, but it can also increase the testing overhead when porting to new platforms.
Account Sign In
View your cart