James Lileks of the Star-Tribune in the twin cities writes:
I have a new program that displays weather information from a variety of cities â€“ such things are plentiful, I know, but this one suits my needs. Itâ€™s set up to tell me how things are in Mpls, Fargo, NYC, Scottsdale, and Glockamorra. Right now itâ€™s 96 degrees in Scottsdale. At eight PM.
Sometimes I miss Arizona weather… The warm nights were wonderful. Want to go for a walk or bike ride in the evening? No problem, just walk outside. Don’t bother with a jacket (or even long sleeves). Of course monsoon season is incredible as well.
With the lack of forecasts available in Seasonality 1.3.x, I decided to release a public beta of Seasonality 1.4. This way, international users can get a forecast back as soon as possible. There are a couple of other key features that are available with this update. One is the new moon phase/moonrise/moonset functionality that I talked about in an earlier post. The second is the ability to add custom locations in Seasonality. Since Seasonality no longer depends on the limited number of locations for which Environment Canada provided forecasts, the app can now display weather data for any location. The interface for this is still fairly basic in the beta, but it will be improved before the final release. Another feature in the pipeline for Seasonality 1.4 is the ability to edit locations. Right now you can check out the Location Info Panel (via the Window menu) to show the configuration of your current location. The plan is to make those fields editable in the near future.
This is the first public beta I’ve ever released. I was a little bit nervous putting out a beta release publicly, but so far it’s been going pretty well. There haven’t been any bug reports yet, which is great. I’m guessing there has to be at least one or two minor bugs in there somewhere, but it’s comforting that no-one has noticed anything yet.
After installing the new Athlon X2 4600+ processor in Rio, I have to say I’m very impressed with the performance gains. Not only was I able to re-install Kubuntu server and get all the necessary packages installed again quickly, but the overall responsiveness of the machine is greatly improved while multitasking. This is especially noticeable with running 2 virtual servers using VMware Server.
So what about the benchmarks? I decided to start with the svnmark that Luis introduced on his blog, and downloaded Subversion 1.3.0 (1.3.2 is available, but the last benchmarks I did previously used 1.3.0). After a couple of runs, I found make -j4 to be the quickest. Here are the numbers:
|Mac Pro Quad 3Ghz:
|Dual Core Athlon 64 2.4Ghz:
|Quad 2.5Ghz G5:
|MacBook Pro Dual 1.83Ghz:
| Dual 2.5Ghz G5:
|Single Core Athlon 64 2Ghz (same server before upgrade):
After running this test and seeing the Athlon X2 compile faster than even a Quad G5, I’m pretty happy. Granted, the OS is Linux and not Mac OS X, but I doubt Linux would be that much more efficient when compiling software using gcc.
So how does it stack up with the forecast processing I need the server to do? Well in this case, I don’t have any solid benchmarks, I’m just running off memory here. The old processor was able to generate and serve a forecast in 0.4 seconds, where the new one can do the same request just under 0.3 seconds–a pretty solid 25% performance gain. This is all sequential code, so this doesn’t take into account the availability of a second processor. The forecast update is also a lot quicker, and with a second core the machine should still be able to handle connections from Seasonality to generate forecasts while updating the forecast database back-end.