If you attended my multicore webinar last month then you know that I’m a big fan of looking out the window at existing open source applications and seeing how they tackle performance issues. The webinar used the Apache HTTP server as an example. A recent Intel engineering study uses a modified version of snort as a test application for improving multicore performance.

The Intel study, commented extensively on by Lori Matassa, indicated performance scaling on a 4 core system of 6.2x. This is notable because the generally expected benefit from adding a core is typically slightly under 1. For example moving an application to a 4 core system and then getting a 3.8x performance boost would be expected and would likely require some work to attain.

The Intel engineering study is notable because it indicates something that experienced practitioners have always known: efficient cache usage is a critical factor in multicore performance. It simply makes no sense to play games elsewhere until you’ve got a grasp on cache efficiency and have maximized that aspect of your system performance.

The Intel engineering study did something interesting with “flow pinning”. Each TCP flow through the system was handled, for the lifetime of the flow, by a single assigned core. This improves cache efficiency by optimizing locality of reference.

The Intel paper also prompts some thoughts in my mind regarding those who are migrating RTOS applications from their dead-end platform to the new funky Linux world. A vogue thought these days is that a virtualization platform can be used to just run your RTOS side-by-side with the new Linux platform. My concern is that taking a non-multicore aware RTOS based application and just moving it to a new multicore processor implies either no cache efficiency or a decrease in cache efficiency. The RTOS based application never had multiple cores and hence has no awareness or ability to do flow-pinning as discussed in the paper.

Potentially a better approach is to migrate the RTOS application’s algorithms to a consistent Linux platform and then do cache optimization work now that you’ve eliminated other variables and have design flexibility. If you want to learn more about RTOS to Linux migration we’ve got an upcoming webinar on that, too.

Take a read… the paper is compelling. 6.2x performance boost on 4-cores is impressive.

ShareThis

I returned home to Atlanta on Saturday still pumped about the great learning and fun that we had that week at the Vision 2008 Embedded Linux Conference. I wanted to share with you a few of them.

I’m sure you saw the splash on the Dell Lattitude ON announcement. Many of my MV colleagues have been nose to the grindstone for months on this project. They were asked to be the systems software provider (using Mobilinux 5.0) and to do the software integration for the application stack. I’ve not seen much of it myself personally but I hear good reports. Hopefully this will be a great success for Dell.

Since I was working and presenting at the event I didn’t get to attend many of the sessions. I did attend one of the last sessions of the conference by Matt Porter of Embedded Alley. His topic was the new Unsorted Block Image (UBI) and the UBI Filesystem (UBIFS) that can be layered upon it. What was superb about Matt’s presentation was that it wasn’t just a summary of information you could have learned off of Google. Matt went beyond the basics and delved into how to model your application’s usage profile to predict how quickly your product will consume the flash lifecycle of modern multilayer NAND flash memories.

I’ll repost if we can get a video of Matt’s presentation up on the web.

The other big highlight of the event was the Thursday night “Streets of San Francisco” party. This was the brainchild of our events manager Pamela Sherwood and it turned out great. The hotel ballroom was converted into street scenes from San Francisco. Tucked in the nooks and crannies was the dinner for the evening and ample adult beverages. All of this led up to what I think was the highlight of the evening… the Rock Band 2 show. I’ve played Rock Band quite a lot but it is a different experience with a stage, lights, and a real sound system.

We’ve got all the videos up on YouTube.

That’s our European FAE team giving a spirited rendition of Girls on Film.

ShareThis

Bootcamp!

The last few weeks at MontaVista have been quite hectic. Our Vision 2008 Embedded Linux Developer’s conference is finally here and getting everything prepared has been a huge effort for many of the MontaVista family. Now that the conference is in progress it is fun to see colleagues from around the world and meet embedded Linux enthusiasts from around the world. The photo above is from the “Bootcamp” training session being held right now. Most everyone has brought their laptops, loaded Linux in a virtualized environment, and are up and learning.

We’ve got two more days of the Vision conference and you haven’t missed out yet. If you are in the San Francisco area the get yourself to the registration page and use the promo code “BRAD”. You’ll get a discount that is steep enough you won’t mind having to upend your week to learn something very useful to your day job.

I’m having a great time at Vision… hope you can join us!

Brad

ShareThis