ARM Community: Software Enablement - ARM Community

It was a full first day at the ESC Summit of Design West. I spent most of the time doing what I love to do best: talking with ARM Partners. Most of the conversations focused on how engineers can achieve better software on their ARM processor-based SoCs.

Andy Frame had a few meetings in the morning so he passed the baton (microphone) to me so I can share some of the insights with the ARMFlix followers. I’m looking forward to Day 2 and speaking with more of the ARM Partners at ESC (check out our handy map). Don’t miss the ARM Connected Community Theatre at ...

This is an article describing the steps how to set up your Android Phone to run ARM Streamline Performance Analyzer.

ARM Streamline Performance Analyzer is a system-wide visualizer and profiler for targets running ARM Linux or Android native applications and libraries. Combining an ARM Linux kernel module, target daemon, and a graphical user interface, it transforms system trace and sampling data into reports that present the data in both visual and statistical forms.

Streamline supports Cortex™-A8, Cortex-A9, ...

2011 was a busy year for developing software on ARM and the activity is reflected in the page views of top Software Enablement blogs. The topics included managing caches, Android (multiple), NEON (multiple), Memory Access Ordering, RISC vs CISC architectures (multiple), and optimizing assembly code (listed by popularity below). In addition, the Software Enablement Community pages (Linux, Solution Center for Android, RTOS, Microsoft, etc) were some of the highest referenced pages on the ARM site. Please let us know if you have more ideas for easing your software development on the ...

DirectFB (Direct Frame Buffer) is a graphics library that is widely used in embedded systems, especially home market. More and more applications or libraries choose DirectFB as backend, such as Cairo, GDK, Qt, V8, X11 and Webkit. ARM NEON technology could be well used in 2D acceleration. In this blog, I’ll describe how to optimize DirecFB using NEON.

1. Introduction
1.1 DirectFB Introduction
DirectFB (Direct Frame Buffer) is a thin library that provides hardware graphics acceleration, input device handling and abstraction, integrated windowing system with support for translucent windows and multiple display layers. It is free software licensed under the terms of the GNU Lesser General Public License (LGPL). Graphics features provided by DirectFB including Rectangle Filling/Drawing; Triangle Filling/Drawing; Line Drawing Blit; Alpha Blending (texture alpha, alpha modulation); Porter/Duff; Colorizing; Source Color Ke...

ARM Development Studio 5 (DS-5) is the software development tool that sets the standard for ARM with its optimising compiler, its extensible and easy to use debugger, and its unique analysis tool, Streamline. But DS-5 is not just for the ARM IP: if you're the designer of an ARM based SoC, an operating system that supports ARM or have productivity tools that support ARM, you can join the DS-5 ecosystem to make sure that your customers also get the DS-5 experience. Here are 10 ways you can do it.

1. Add DS-5 debug support to your ARM based SoC. The DS-5 debugger has a target database that is extensible by you,...

One of the most common requests from software engineers running software on a Virtual Platform is to be able to profile the executing software. In this document I will describe how to use the ARM Profiler included in RVDS Professional with the Fast Models from ARM that are commonly used to create SystemC Virtual Platforms with the Cadence Virtual System Platform (VSP). As you can see from the introduction there is a combination of products that all work together to enable non-intrusive profiling of embedded software, but at times the HOW TO details of profiling can be a bit of a mystery and knowing where to look to find the details may not be obvious. In fact, a recent VSP user tried to setup profiling by himself and was not successful and wasn’t really sure where to look. As a result I created the following information that I’m sure would be valuable to many other readers.

...

This article describes how to get Android running on your favourite ARM-based System on Chip (SoC) board. We run through the overall procedure and point out potential pitfalls and other things that you may encounter.

Since the Android software stack was primarily designed around the ARM Architecture, there are not many things that need amending to get it to work on another ARM platform.

We assume that your workstation has Ubuntu (10.10 or later) Operating System installed, and that you have already followed the instructions found at [1] to be ready to build Android sources. These instructions have been tested with Ubuntu 10.10, but they should be compatible with other GNU/Linux OSes.

Terminology

For the purposes of this document, we use the following terms.

Mainline kernel ...

Android作为优秀的开源软件解决方案, 它的作用域已经从手机市场,波及到了平板电脑,甚至以数字电视、机顶盒为典型应用的数字家庭领域。Android最初是为手机移动设备量身定做的,它默认支持的分辨率，色彩深度， 多媒体播放架构，用户交互方式，2D/3D图形的性能等都无法适应类似于数字电视，机顶盒这样的家庭应用。
因此，将Android移植到数字电视或机顶盒需要对Android进行大量的定制和修改。这些修改和定制涉及到Android软件架构中的各个层面，我将用四篇Blog来依次介绍如何将标准的Android移植到数字电视或机顶盒平台。

在开始我们的讨论之前， 我们先来简单探讨一下目前数字电视和机顶盒软件的现状和未来
数字电视/机顶盒软件的现状目前， 数字电视和机顶盒的软件架构中由于使用了不同的操作系统， �...

My previous post provided an introduction to the concept of memory access ordering. It did not however provide any solution to the problem, or necessarily specify where such ordering can be significant.

Now, not all software developers need to be deeply aware of memory access ordering or barriers. Unless your code interacts directly with hardware, interacts directly with code executing on other cores or directly loads or generates instructions to be executed, things will mostly Just Work. If your interaction with hardware is completely through a device driver (meaning: no device control registers mapped directly into your application), then it is the responsibility of the driver to enforce ordering. If your communication with software running on a different core makes use of a multithreading API, for example using Pthreads or Java threads, then it is the responsibility of that API to enforce ordering. If your program executes on an operating system that implements demand paging, then clearly it is the responsibility of the operating system to enforce ordering of such operations.

However, if you are writing device drivers, implementing your own thread-communications or creating a JIT compiler, then not being aware of the proper use of barriers can lead to unexpected and difficult to diagnose problems. Where your program requires a ...

I recently gave a presentation at the Embedded Linux Conference Europe 2010 called Software implications of high-performance memory systems. This title was my sneaky (and fairly successful) way to get people to attend a presentation really about memory access (re)ordering and barriers. I would now like to follow that up with a few posts on the topic. In this post, I will be introducing a few concepts and explain the reasons behind them. In future posts, I will follow up with some practical examples.

The Sequential Execution Model

In the Good Old Days, computer programs behaved in practice pretty much the way you might instinctively expect them to from looking at the source code: Things happened in the way specified in the program.Things happened in the order specified in the program.Things happened the number of times specified in the program (no more, no less).Things happened one at a time.

In modern computer architecture, this nostalgic fantasy is sometimes referred to as the Sequential Execution Model. In order for existing programs and programming models to remain functional, even the most extreme modern processors will attempt to preserve the illusion of Sequential Execution from within the executing program. However, underneath your feet...

Previously I introduced UEFI and its history, now I will get into its benefits, especially for the usage on ARM-based systems. I will further explain the organization of the UEFI Forum.

Advantages
Even though existing ARM preboot firmware does not have the BIOS limitations, there are many advantages for ARM preboot firmware to standardize on UEFI. OEM/ODMs are always looking into reduced development cost. Code sharing among products is one way to achieve that.
With ARM and x86 both in the computing continuum, UEFI not only enables code sharing among ARM products or among x86 products, it also enables code-sharing across processor architectures. Products may share many of the peripheral devices (Network, SATA, USB controllers, etc.) and feature sets across the designs.
Figure 2 shows an ARM port where 99.42% lines of code does not need to change from an x86 port.
Figure 2: Lines Added/Change...

Version 3.6.0 of Valgrind was released a couple of weeks ago. Probably the largest change this release is the addition of support for Linux running on ARM.

Valgrind is a GPL'd framework for building simulation based debugging and profiling tools, plus a set of "standard" tools. The best known of these is Memcheck, a memory error detector, but in fact it is only one of eight tools in the standard distribution: two memory checkers, two thread checkers, two performance profilers and two space profilers.

You can download the sources from www.valgrind.org. Alternatively, you may be able to get pre-built packages via your Linux distro, or via Linaro, although note that the 3.6.0 upstream release post-dates pre-built packages. 3.6.0 is known to work on Ubuntu 10.04 and 10.10 on ARM, and on the Nokia N900 running Maemo 5.

Also available online is full documentation. For those impatient to get going, the ...

It was always about the code (and where it would be used!)

When I was a software developer I would often find that the project team I was in would try to guess how many devices the code would eventually run on. So at the launch of the Cortex-A15 last week one of the main points that hit home for me was just how wide the spectrum of power and performance points the Cortex-A family of processors could cover - from feature phone to superphone, tablet to DTV, home server to web server etc. This means that a developer could now find their software running across a huge range of devices in the future.

So is it the same software?

Absolutely. Cortex-A15 is based on the same ARMv7A architecture that the other Cortex-A processors use, therefore allowing the exact same application code to run on all of them, from a ...

In January 2009 Ed set about the task of rewriting the interpreter. Java byte codes are quite a compact code for representing programs but the Virtual Machine they target has a stack architecture rather than a register architecture. This makes the Java VM somewhat at odds with modern day processor architectures such as ARM which is register based. Ed’s approach was to use a Peephole Optimizer to spot common byte code sequences that loaded items onto the Java Virtual Machine stack, manipulated them and stored them back into memory. These complete sequences could then be executed in optimized ARM assembler. Having done this process for the first version of the optimized interpreter longhand it became clear that this repetitive task could be eased with the creation of a notation to describe the sequences and how they related to the ARM assembler. (Ironically I did the same thing some 20 years ago for a different VM). A tool could then be used to automatically generate the template interpreter from the notation. The tool proved to be extremely useful and was naturally processor agnostic so we contributed it back into open source along with the optimized interpreter it generated last year.

Ed’s optimized interpreter for OpenJDK increased the performance by around a factor of 4X for a couple of the classic Java benchmarks- Embedded Caffeine Mark and EEMBC Grinderbench ...

Sitting in the airport at the end of a week’s business trip to the US, I reflected back on the week. It turned out that my colleague on this trip has an even worse sense of direction than myself…Potentially disastrous, especially when you’re driving between airports, hotels and meetings in cities that you’ve never visited. This is where Google Maps becomes utterly indispensable. Installed on my Nokia E71 it makes use of the built in GPS and 3G and Edge networks to provide a running view of where we are, driving or walking. Without it we wouldn’t have found the wonderful Boulderado hotel or the Boulder Bookstore with its impressive converted ballroom. Actually, we’d probably still be driving around somewhere near Dallas.

Life changing and mind boggling as the online, always connected life of a sometime digital nomad is, w...

ARM is hiring OK, so that got some people’s attention and confused others – actually we are hiring, and in particular software developers. What can often come as a surprise to people is that as well as having a team of people that go plan and work alongside some our different software partners, ARM has a software engineering group that work on key bits of software, particularly on Cortex-A8 and Cortex-A9 projects. In fact it's very likely that some of the code running your mobile phone was developed by some of the ARM team.

The team cover a wide range of software projects that include:Web and web runtime optimization, for example JavaScript JIT optimization work on projects such as Tamarin, Webkit and Squirrelfish NitroExtreme), and OpenJDK optimization work.Operating System development – including Android, Linux kernel hacking and a ...