Archiv der Kategorie performance

Linux kernel report

Jonathan Corbet presented a informative presentation about the recent linux kernel work. It was held at the 11′th realtime Linux workshop, and offers some insides and statements not always presented in the normal press, like from 2.6.27 -> 2.6.31++ rough timeframe (October 9, 2008 to September 18, 2009) 48,000 changesets was merged by 2,500 developers and 400 employers. This result is that the kernel grew by 2.5 million lines, or better 140 changesets merged per day and 7267 lines of code added every day. But he also depicts some major functionality added by each version and why it’s needed. paired with some funny pictures a nice walk through. The other papers from the conference are also a good read, take a look at them by yourself

60 GHZ ECMA-387 demonstrated

TUD presented EASY-A and achieved a record in transmission. They claim its very energy efficient in that the emitter and receiver only need one bit resolution for transmission. The signals use 4GHz bandwith and it took 6.4 seconds for 8 Gigabyte.
one have to admit, that they (still) used cable for transmission, but they’ve designed/created the demonstrator using the concept of their 60 GHz wireless transmission. They estimate about 5 years before this could be used in commercial products.

File system

File System and Storage Lab provides a wealth of informations about various issues in that space

Code Splitting for Network-Bound Web 2.0 Applications

Ajax apps provides us with rich and responsive user interfaces. And aside all code reductions modern frameworks provide, if you use firebug or fiddler or tamperdata one easily thinks about further optimization. Microsoft research has researched this topic and created Doloto. Take a look at their paper and of course at the tool itself and yes, you need windows for it, but i still like the paper :-)

hoard

Emery Berger has created the Hoard fast memory allocator. The whole site is a fantastic memory managment resource. Hoard is easily incorporated (drop in replacement), so if you’re using C/C++ and threading, take a look. I mentioned this lib so often, that I needed to mention it.

Latency

Latency is Everywhere and it Costs You Sales - How to Crush it. Todd Hoff has created a good collection about various topics and links around this issue. By reading the next entry by the author, I got also aware that the Computer Communication Review is also online. How could we live without the web ;-)

10 Million Messages per Second

impressive numbers demonstrated by Tervela, especially if you look at the latency achieved. Wonder which companies have them on the watchlist.

Chalcogenide glass photonic chip

Australia’s Center for Ultrahigh-bandwidth Devices for Optical Systems explain in a research paper how they want to achive 640 Gb/s with all optical signal processing

Maybe this comes as a rescue to united states speed trap

Internet Measurement Conference

The entry page to the conference proceedings. If you’re interested in net measurement areas, it’s wortwhlie to have a look. Thorsten just send me the link to ufsdump. good papers, even I have some things to add about this. But unfortnunatly I have to postpone this. as a remark for my self, anonymous mem, dstat.

GWt extreme

seems I learned how to embed :-). I went through a couple of good presentations recorded at Google I/O recently. while I liked the Lombardi folks explaining how they switched from flash to dojo finally implementing GWT, as well as the one showing the recent advances in GWT 1.5 (Javascript and DOM programming in GWT) where its getting clear that they want to ease (or better shield you) from accessing low level DOM. But nevertheless the presentation embedded above gets into gwtquery, fast drawing and a very nice way to develop gadgets.