aToaD #3: PassMark AppTimer

Benchmarker. Measure launch time of an application.

With this tool of the day, we touch another of the main reasons I decided to start this series: have a note, somewhere, of a great application which I just found but that maybe I’ll forget in a week… and then will never be able to find again a year later when I ask myself “oh, there was this great application that could do that, but what the heck was its bloody name again”.

So, today here is PassMark AppTimer. (Note to self: set up a miror of it some time) As its full name suggests, it was published by PassMark, although I haven’t been able to find it on their site (hence the note to self). I had never heard of PassMark, or so I thought, because I then realized they are the guys behind cpubenchmark.net and videocardbenchmark.net, which I had already consulted a few times before. Not that I remembered the name, but I did remember the “unique” retro design ^^. AppTimer’s design is quite retro, too 😉

AppTimer is, as its name suggest, a program which lets you measure the time it takes for an application to launch. I discovered it after taking part in a discussion about, of course, browser launch speed (with again Firefox as the loser of the day). It has a fairly basic interface, its main limitation IMO is its inability to save your settings between 2 launches (that is, 2 launches of AppTimer, not 2 launches measured by it), which means that every time you’ll have to configure where to save the results (yeah, that’s another limitation: it doesn’t display the results, it only saves them into a log file, which you then have to open separately), how to detect that the program has finished opening, etc.

Well, what more can I say about this 1-window program? Why, it just does the job 🙂 For the sake of completion, just some results. Launch times on a Samsung 840 SSD, just after rebooting the PC:
– Firefox: 3.3 seconds
– Iron: 0.07 seconds
– Maxthon: 0.3 seconds
– Opera: 0.2 seconds
– Audacity: 0.02 seconds
Note that the results are given with a precision of 0.1 millisecond, but I don’t think rounding to 1 digit is inappropriate given the high variability of those measures (at least around 10-20%).