This POVRay raytracing took 9 days and 15 hours for my laptop (Dell Inspiron 5150, 3.2 GHz) to render. It’s not mine (if I could do stuff like this, I’d probably be in a different line of work), just a very cool image that caught my eye while I was browsing the POVRay Hall of Fame. I thought it would be fun to see how long it would take a scene like this to render. I figured a few hours, couple of days at most. Boy, was I wrong.
I love the things POVRay is capable of. If I had more time on my hands I’d probably spend some of it playing with POVRay some more.
Whither Palm
PDAs have a well deserved reputation for suddenly going belly up (bits up?) when you least expect it. Although mine isn’t showing any signs of dying anytime soon (aside from wear and tear) it’ll be three years old in a few months, so it’s probably time to start thinking about a new one soon.
Problem is, unless you want an all-in-one combo smartphone type device (which I most definitely do not want), Palm‘s offerings have become rather thin. Palm’s only new announcement over the normally busy holiday season was the Treo 680. No new ‘plain’ PDAs in sight, nor have I caught any rumours of any coming down the pipe either. There’s just the E2 and TX left (the Z22 is too entry-level for me) now that the LifeDrive has been EoL‘d.
I guess if I needed to get one today, it would probably be the TX. Hopefully my T3 will live long enough for something new to come out of Palm. Otherwise I might have to go back to actually using my brain to remember stuff.
Thumbdrive guts
This is what the inside of a HAL9000 Memory Unit. For something that’s only 43x19x3mm, it sure packs a lot of storage in there (2 GB).
.
The x-ray was acquired on an 8″x10″ cassette at 50 kVp, 1.6mAs using the small focal spot. The HAL9000 memory unit was placed directly on the cassette.
Down CPU memory lane
Anybody remember these?
This is a pair of 200 MHz Pentium Pro CPUs that I found lying around in our PACS storage area. It was probably pulled out of an old server several years ago. I thought they would make a good decoration piece for my desk.
It’s a beast of a chip with the entire package measuring 6.8×6.3 cm. The Pentium Pro’s were introduced in 1996, with this particular variant, 200MHz and 512 kB cache, coming out in mid 1997. This makes it a veritable fossil in computer time.
Haven’t quite decided yet what I’ll do with them or how I want to display them. Maybe I’ll see if I can pop the heat spreader off one of them to expose the actual CPU. Tried to take an x-ray of the die, but despite using some pretty hefty techniques, I wasn’t able to get any significant penetration through the chip. Managed to get a faint outline of some of the CPU at 90kVp and 50mAs, but nothing interesting enough to post. It’s a pretty dense package. Might have to try a different tactic and shield the surrounding area with some lead so I can really blast through the chip.
Hang on, I’m not done waiting yet
The server that hosts radinfo.musc.edu is hardly a speed demon (it’s a dual 450 MHz Pentium II), so for most things I’ve got running on this server, I’m used to waiting a little bit for things to get processed.
Lately I’ve been running the wheeljack dev build of Movable Type (I like to be on the bleeding edge of software) which has incorporated support for FastCGI (recently rolled out with MT 3.34). I finally got around to building FastCGI and incorporating it into the Apache config to see how it would help with the back end of MT. While it’s not much of an improvement on this pokey slow machine, it’s definitely faster than pre-FastCGI. MT’s user interface has become noticeably faster although I’ve had to give up an old hit counter CGI program that just didn’t want to run under FastCGI. It’s definitely enough of a speed improvement to make me happy.
Still need to test things to make sure everything works. If anybody happens to notice something got busted, let me know.
Update 1: Well, publishing an entry took a curiously long time. Normally once I hit the publish button I usually see the next screen that say things are getting published and sites are getting pinged. This time it stayed on the entry screen for several minutes before switching. I wonder if this is normal…
Update 2: Looks like commenting is really slow too. Seeing a lot of error messages in the Apache error log about the mt-comment.cgi and mt.cgi processes getting killed and restarted during comment processing. Seems like anything involving rebuilding takes a long long time. Might have to back out the FastCGI stuff.
Update 3: There were too many unexplained errors showing up in the Apache error log so I decided to back out the FastCGI stuff for now, at least until after the weekend when I might have some more time to investigate.