On Tue, 18 Jun 2002, Sam Ewalt wrote: > On Tue, 18 Jun 2002 10:28:24 -0400 (EDT), Steve <[EMAIL PROTECTED]> wrote: > > (regarding Linux browsers) > > > I use Netscape 3.04 as my default browser. I use Arachne > > as my "launch from e-mail" browser. > > > I have both Netscape 4.79 and Mozilla 0.9.9 installed, > > but I use them both as "last resorts." I want small, fast, > > and efficient, no matter what hardware I'm using. > > > Isn't there a point at which the advantage of small, fast and > efficient is outdone by the speed and capacity of the hardware?
No. Twice as fast is twice as fast no matter what the base number is. > For instance, I just bought and soon will take home a refurbished > computer with a 500 Mhz CPU, 20 gig hd and 128 megs of RAM. This > "obsolete" system was dirt cheap. My main machine is an AMD K/2-500 with 64MB. Here are the "load, run & quit" times (without the browsers being already cached): Mozilla 0.9.9 - 0:19.56 elapsed 39%CPU Netscape 4.79 - 0:07.88 elapsed 27%CPU Netscape 3.04 - 0:05.83 elapsed 31%CPU Amaya 4.3 - 0:04.39 elapsed 31%CPU Arachne 1.66 - 0:02.89 elapsed 24%CPU Here's how I conducted the test. At the command line, type $ time <browser> Then as soon as possible, click File -> Quit (or equivalent). Granted, there's a bit of reflex time built in here, and Arachne has a one click advantage since it only requires a single click to exit instead of two... but running through the list three times and using the shortest elapsed time for each browser should yield fairly meaningful results. > In terms of practical performance would I even see an observable > differance between Netscape 3.04 and 4.79 on this new system? "Practical" performance? Hmmm... I generally have NS 3.04 on the screen at all times, so perhaps the above test isn't really "practical." If I left Mozilla up all the time, its inherent bloat wouldn't show up in load times, but of course, mozilla takes some 19MB to run, so would tend to slow everything else I'm running. Mozilla takes 30% of my RAM, which would be only 15% of yours. It obviously won't bog your machine down as much as it does mine. Also keep in mind that all the above pertains to a machine running setiathome, which is using 100% of the CPU cycles. Switching from that to a browser would likely be slower than launching the browser from "idle." How about page load/render time? Well, since only Arachne and Mozilla show us those times, let's pull this page off the hard drive... to ensure internetwork traffic is eliminated as a variable (empty both caches first). I give the online address below, but use that page from the hard drive for the test. Fairly short HTML, but with several different graphics included, many iterations of one of them. Seems like a pretty good test page. $ arachne http://twoloonscoffee.com/order.html (load time=0:02) | After cached =0:01 $ mozilla http://twoloonscoffee.com/order.html Document Done: (4.79 secs) | After cached 3.305 secs So, yes, even "practical performance" is vastly different between Arachne and Mozilla, even on a 500Mhz machine... wouldn't you say? -- Steve Ackman http://twoloonscoffee.com (Need green beans?) http://twovoyagers.com (glass, linux & other stuff)
